Page 1 of 1

ITT We Test 3rd Seat Hands

Posted: Sun Mar 28, 2021 3:39 am
by Wes (aka the legend)
This thread is a continuation of what Irishwolf started. Irishwolf showed us that calling with this hand from 3S-R1 is the best play:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

Vs (Card_K-H) upcard

Here's my sample results of his hand:

Sample size: 70
1,1,1,1,1,0,1,1,0,1,-4,0,0,-4,1,-4,-4,1,1,0,-2,-3,-1,1,-4,1,-1,1,-3,1,1,1,-1,1,1,0,0,0,-3,-3,1,1,1,1,1,0,-4,0,-1,1,0,0,1,-1,-1,-3,1,-3,-4,-2,1,1,0,1,1,-3,1,0,0,-3
Mean of calling: .4286
Standard deviation: 1.7740
P value: .0471

Calling gets euchred: 22/70 = 31.43%
Calling gets 1 pt: 34/70 = 48.57%
Calling gets sweep: 14/17 = 20%
EO of calling: .2571
EO of passing: -.1714
EV of calling: .4286

In my next post I will be testing this hand:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

Vs (Card_A-H) upcard

This is the worst possible 3 trump + green suited ace hand which is intentional as I want to firmly establish that all these hand types are calls from 3S-R1 vs a non-jack upcard. After I accomplish this, then I will move on and test a hand like this:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

My hypothesis is it will be better to pass that hand and thus bag the dealer with an expert partner becuz you hit a Next call. But who really knows until it's put to the test.

Re: ITT We Test 3rd Seat Hands

Posted: Sun Mar 28, 2021 3:50 am
by Wes (aka the legend)
Here's the data I have on this hand so far:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

Vs (Card_A-H) upcard.

Sample size: 40
-1,3,0,1,0,-1,0,-3,1,-1,0,0,-1,3,-1,2,3,4,3,0,-1,-1,3,3,1,3,3,-1,-1,-4,0,0,3,-3,1,-1,1,0,0,0
Mean of calling: .45
Standard deviation: 1.9075
P value: .1437

Calling gets euchred: 14/40 = 35%
Calling gets 1 pt: 19/40 = 47.50%
Calling gets sweep: 7/40 = 17.50%
EO of calling: .1250
EO of passing: -.3250
EV of calling: .45

I still haven't reached a 95% confidence interval yet so I'll keep going. As far as the conditions of my test, I'm playing every seat the way I would play it.

Re: ITT We Test 3rd Seat Hands

Posted: Sun Mar 28, 2021 4:22 am
by Wes (aka the legend)
In case others are interested, here's how to test a hand:

Step 1) Have a computer + Excel + deck of cards + kitchen table.

Step 2) Create three column headings in your excel spreadsheet. First column being "EO Of Calling", 2nd column "EO Of Passing", 3rd column "EV of Calling". Then it's a matter of dealing out the cards and entering numbers in boxes. For example, if calling got your team 1 pt and passing led to your team euchring the dealer and getting 2 points, the EO of calling would be 1, the EO of passing 2, and the EV of calling (1 - 2) would be -1. You can create a subtraction formula for the "EV of Calling" column if you want. Just type in =(A2-B2) in the first box and drag that down the column. I'm sure there's probably a smarter way to do it but that gets the job done. I'm no excel expert.

Step 3) Take the numbers from the "EV of Calling" column and enter them in this box from this website: https://www.calculator.net/standard-dev ... &x=74&y=18

This will give us our standard deviation and mean. Make sure the mean matches the mean in your excel spreadsheet (just highlight the entire "EV of Calling" column, the mean will be at the bottom). If the mean from your excel spread sheet doesn't match the mean from this site then you've made a data entry error somewhere.

Step 4) Then go to this site: https://www.medcalc.org/calc/test_one_mean.php

This is where we get our P value. Enter the mean, standard deviation and sample size. Set the null hypothesis value to zero. Click on "test" and then out comes your P value. Once your P value is below .05 you have reached a 95% confidence interval.

You now have the power to solve every first round situation. Use it wisely! You no longer have to guess and wonder or take anyone's word for it.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Mar 29, 2021 4:18 am
by Wes (aka the legend)
Well this is a crazy development. Feels like the deck turned possessed against the calling strategy.

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

Vs (Card_A-H) upcard.

Sample size: 70
-1,3,0,1,0,-1,0,-3,1,-1,0,0,-1,3,-1,2,3,4,3,0,-1,-1,3,3,1,3,3,-1,-1,-4,0,0,3,-3,1,-1,1,0,0,0,-1,3,-1,0,-1,0,-4,0,3,0,-3,3,-1,2,0,-3,-4,-4,1,-1,-1,-1,0,-3,2,0,-1,-1,-4,-1
Mean of calling: -.0429
Standard deviation: 2.039
P value: .8608

Calling gets euchred: 27/70 = 38.57%
Calling gets 1 pt: 33/70 = 47.14%
Calling gets sweep: 10/70 = 14.29%
EO of calling: -.0143
EO of passing: .0286
EV of calling: -.0429

Ok so now calling has a slight negative EV and I'm not even close to reaching a 95% confidence interval. This one could take awhile.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Mar 29, 2021 9:53 am
by irishwolf
Good job!

You can now really see the value of AS/KS vs AS/9S for increasing the euchre rate and decreasing Sweeps. But no surprise as opponents will have KS, QS or JS doubleton to catch that 9S that changes the dynamics of the two hands.

Because of the means this one might require over 400 hands to reach 95% CF (is it worth it?). With only 70 hands the the mean might change.

Irishwolf

Re: ITT We Test 3rd Seat Hands

Posted: Tue Mar 30, 2021 12:30 pm
by Wes (aka the legend)
irishwolf wrote:
Mon Mar 29, 2021 9:53 am
Good job!

You can now really see the value of AS/KS vs AS/9S for increasing the euchre rate and decreasing Sweeps. But no surprise as opponents will have KS, QS or JS doubleton to catch that 9S that changes the dynamics of the two hands.
True. There is also a significant difference between a dealer upcard King and an Ace. When the dealer has an Ace upcard everytime he has R+1 he is guaranteed 2 tricks when 3S calls. That is not the case with the King upcard as S1 can always have Left+Ace in trump. It's especially demoralizing when the dealer has AhJh and S1 has JdKh becuz when 3S calls an expert dealer is going to render both of S1's trump completely useless as S1 will lead the Left, the dealer will play the Right and then send the Ace of trump to clean S1 out and take two more enemy trump with one lead. And then it often comes down to whether S3's 9S from his outside AS9S can take a trick. If not, then it's usually a euchre. Whereas if S3 passes, then the dealer is calling with his R+1 (AhJh) unless he has a stopper/euchre hand and now S1-S3's team is prime to euchre him with 5 trump between them and the Left behind the maker. So this is where -4s happen to the calling strategy. S3 calls and gets euchred (-2), but if he passes S4 calls, S1 no loner leads trump, and S4 gets euchred (+2), hence -4 total for the S3 call strategy.

Notice when S3 has an outside AsKs, this dealer line with AhJh can never euchre him. Another reason AsKs is so valuable. So instead of creating -4s in this spot, a S3 call will create -1s (A call gets +1 pts, but a pass gets +2). That's a big difference obviously. This is why it's easy for me to conclude that

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

Vs a (Card_A-H)

Would still show that calling is +EV, so we don't have to test that configuration.
irishwolf wrote:
Mon Mar 29, 2021 9:53 am
Because of the means this one might require over 400 hands to reach 95% CF (is it worth it?). With only 70 hands the the mean might change.

Irishwolf
I don't know if it's worth it or not lol. So far it looks like

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

Vs a (Card_A-H)

is right at that break-even point. So close that I will now confidently hypothesize that calling with the following hand is a losing play:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-C)

That's another hand I will test out soon.

Re: ITT We Test 3rd Seat Hands

Posted: Tue Mar 30, 2021 4:50 pm
by Wes (aka the legend)
Wes (aka the legend) wrote:
Tue Mar 30, 2021 12:30 pm
True. There is also a significant difference between a dealer upcard King and an Ace. When the dealer has an Ace upcard everytime he has R+1 he is guaranteed 2 tricks when 3S calls. That is not the case with the King upcard as S1 can always have Left+Ace in trump. It's especially demoralizing when the dealer has AhJh and S1 has JdKh becuz when 3S calls an expert dealer is going to render both of S1's trump completely useless as S1 will lead the Left, the dealer will play the Right and then send the Ace of trump to clean S1 out and take two more enemy trump with one lead. And then it often comes down to whether S3's 9S from his outside AS9S can take a trick. If not, then it's usually a euchre. Whereas if S3 passes, then the dealer is calling with his R+1 (AhJh) unless he has a stopper/euchre hand and now S1-S3's team is prime to euchre him with 5 trump between them and the Left behind the maker. So this is where -4s happen to the calling strategy. S3 calls and gets euchred (-2), but if he passes S4 calls, S1 no loner leads trump, and S4 gets euchred (+2), hence -4 total for the S3 call strategy.

Notice when S3 has an outside AsKs, this dealer line with AhJh can never euchre him. Another reason AsKs is so valuable. So instead of creating -4s in this spot, a S3 call will create -1s (A call gets +1 pts, but a pass gets +2). That's a big difference obviously. This is why it's easy for me to conclude that

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

Vs a (Card_A-H)

Would still show that calling is +EV, so we don't have to test that configuration.
I would also add that IF we have reached a breakeven point with this hand:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

vs a (Card_A-H)

I would then deduce that this hand is a +EV call:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

vs a (Card_K-H)

The difference between a KH and AH upcard is significant enough to make that deduction imo.

Re: ITT We Test 3rd Seat Hands

Posted: Tue Mar 30, 2021 9:38 pm
by irishwolf
I am not sure it will make that much difference because King and the Ace are both bigger than anything S3 has. I think less that 5%, still making either significant and positive EV for the hand. But, for sure, needs to be proven.


Upcard (Card_A-H)

I would then deduce that this hand is a +EV call:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

vs a (Card_K-H)

The difference between a KH and AH upcard is significant enough to make that deduction imo.

Re: ITT We Test 3rd Seat Hands

Posted: Wed Mar 31, 2021 2:11 am
by irishwolf
I log and record all my tests so it is easy to go back and switch the KH up with the AH as the up card. I am very comfortable ordering the AH from the 3rd seat with the hand QH 10H 9H AS KS and the EV being positive and statistically significant - best strategy to order. Of course I would not order the JH as he the dealer should only turn this down less than 35% of the time.

Funny thing is an actual tournament I ordered and the dealer had JH JD AH (AH WAS THE UP CARD) and my partner even led an ace. Shit happens, that would be 6.5% both bowers. I would do it again, and again.

But do your thing!

Re: ITT We Test 3rd Seat Hands

Posted: Wed Mar 31, 2021 3:22 am
by Wes (aka the legend)
irishwolf wrote:
Wed Mar 31, 2021 2:11 am
I log and record all my tests so it is easy to go back and switch the KH up with the AH as the up card. I am very comfortable ordering the AH from the 3rd seat with the hand QH 10H 9H AS KS and the EV being positive and statistically significant - best strategy to order. Of course I would not order the JH as he the dealer should only turn this down less than 35% of the time.

Funny thing is an actual tournament I ordered and the dealer had JH JD AH (AH WAS THE UP CARD) and my partner even led an ace. Shit happens, that would be 6.5% both bowers. I would do it again, and again.

But do your thing!
Yep, I now have no doubts that ordering up:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

is the right play no matter what the upcard (excluding the Jack).

That said, down 9-6 we pass hoping to either euchre the dealer or hoping the dealer passes and our P is sitting on a game winning loner, correct? IOW just like there are theoretical spots where we should make a -EV donate due to the fact that we are only playing up to 10, sometimes we gotta make slight -EV passes going for the big play. Assuming you agree with the down 9-6 scenario, then the question becomes how much do we have to be down in other spots to justify this -EV pass. Down 3+ in general seems like a good rule of thumb to me.

This opens up another off-topic debate. In the other thread we basically established that ordering any 3 trump-no 2nd rd hand is +EV from S1-R1 vs a non-jack, but should we make the -EV pass hoping for a big play (euchring the dealer) when down a certain score. I'm actually not convinced we should ever deviate from a +EV call in that spot no matter what the score becuz the potential payoff isn't that great. Whereas when we make a -EV pass from the 3S-R1 it's not just a dealer euchre we can get, but a S1-R2 4 pt loner also. Rare? Sure, but an enticing potential payoff that a -EV S1-R1 pass can't give us.

Re: ITT We Test 3rd Seat Hands

Posted: Thu Apr 01, 2021 4:09 am
by Wes (aka the legend)
(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_9-S)

Vs (Card_A-H) upcard.

Sample size: 100
-1,3,0,1,0,-1,0,-3,1,-1,0,0,-1,3,-1,2,3,4,3,0,-1,-1,3,3,1,3,3,-1,-1,-4,0,0,3,-3,1,-1,1,0,0,0,-1,3,-1,0,-1,0,-4,0,3,0,-3,3,-1,2,0,-3,-4,-4,1,-1,-1,-1,0,-3,2,0,-1,-1,-4,-1,2,-1,0,3,3,0,0,0,-1,-1,0,-1,3,-1,-1,0,0,-1,0,-4,3,0,2,-1,4,0,4,-4,0,3
Mean of calling: .08
Standard deviation: 2.038
P value: .6956

Calling gets euchred: 37/100 = 37%
Calling gets 1 pt: 50/100 = 50%
Calling gets sweep: 13/100 = 13%
EO of calling: .02
EO of passing: -.06
EV of calling: +.08

It feels like this is gonna take forever to reach a 95% confidence interval. I'm just gonna call it now becuz I got other hands to work on. It's a statistical tie between calling and passing with this hand. For the sake of simplicity, if I were writing a euchre book I would just recommend to always call from the 3rd seat with QhTh9h + an outside suited green ace vs a non-jack. The exception being when you're down a lot and you wanna hunt for that S4 euchre/S1-2Rd loner.

Re: ITT We Test 3rd Seat Hands

Posted: Thu Apr 01, 2021 5:39 am
by Wes (aka the legend)
Now I'm testing this hand:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

Vs (Card_K-H)

Sample size: 20
1,-1,0,-1,0,-1,0,-1,-1,1,-1,-1,0,1,-4,-2,-2,-1,-1,0
Mean of calling: -.7
Standard deviation: 1.1743
P value: .0153

Calling gets euchred: 7/20 = 35%
Calling gets 1 pt: 8/20 = 40%
Calling gets sweep: 5/20 = 25%
EO of calling: .2
EO of passing: .9
EV of calling: -.7

Well this is crazy. It actually only took 17 hands to cross the 95% confidence interval. Wow. So apparently calling with this hand is a clear loser. Passing is best. The parameters of this sample are pretty simple. Each seat plays like me, so S1 is always calling something in the 2nd rd veering towards Next if he doesn't block reverse Next. S3 passing in the first round works out well becuz his hand hits a Next call. I would have predicted passing beats out calling with this hand, but no way in hell would I have thought we could reach a 95% confidence interval that quickly. I'll probably do at least 20 more sample hands and see what happens. It just doesn't feel right to stop at 20 even tho the data suggests we can.

Re: ITT We Test 3rd Seat Hands

Posted: Thu Apr 01, 2021 9:56 am
by irishwolf
Shocking! I am surprised not very close to xxxASKS even though one less off suit card. I don't get it?

But I might pass on that hand if I have my regular partner. So what about (EV) if passing and Next is called every hand?

Irishwolf

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 6:38 am
by Wes (aka the legend)
irishwolf wrote:
Thu Apr 01, 2021 9:56 am
Shocking! I am surprised not very close to xxxASKS even though one less off suit card. I don't get it?

But I might pass on that hand if I have my regular partner. So what about (EV) if passing and Next is called every hand?

Irishwolf
I assume you're talking about:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

If you have a good partner this hand is a pass. With a bad partner or unknown who knows tho. I'm not really interested in the EV of passing this hand if Next is called every time. That's not realistic to me. I mean S1 plays like me so he's already calling Next a lot. As you know when I'm in S1 I always call something veering towards Next if I don't block reverse Next, but I'm still crossing the river with dubious hands like this (assume KH was turned down):

Calling spades:

(Card_Q-S) (Card_10-S) (Card_9-S) (Card_J-H) (Card_10-H)

Calling clubs:

(Card_J-C) (Card_9-C) (Card_A-H) (Card_10-S) (Card_10-D)

Calling clubs:

(Card_A-C) (Card_Q-C) (Card_9-C) (Card_J-D) (Card_9-S)

So passing QhTh9hAdKd can still lead to some R2-S1 euchres in my sample. It's not a perfect strategy obviously but it still clearly outperforms calling hearts in R1 becuz the majority of S1's calling range in the 2nd rd is Next and it works well given that S3 has 2 trump in Next and 2 voids.

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 6:51 am
by Wes (aka the legend)
Given that I already reached a 95% confidence interval with QhTh9hAdKd vs Kh I decided to test a different hand:

(Card_K-H) (Card_Q-H) (Card_10-H) (Card_A-D) (Card_K-D)

Vs (Card_9-H)

The cool thing about this hand is I'm basically still testing QhTh9hAdKd becuz if passing beats out calling in the above hand then passing will beat out calling in QhTh9hAdKd. Why? Becuz KhQhThAdKd vs 9H is the strongest possible calling hand from this configuration. If calling can't beat out passing in that hand it wont beat out passing in any lesser hand either. So I just did 40 hands of KhQhThAdKd vs 9h:

Sample size: 40
-1,1,-3,0,-3,-1,-1,3,-1,-1,0,0,0,4,-3,-1,-4,-1,0,1,-1,0,0,-1,-4,1,2,1,-1,-1,-1,3,0,0,-1,-1,0,1,0,1
Mean of calling: -.325
Standard deviation: 1.7005
P value: .2340

Calling gets euchred: 10/40 = 25%
Calling gets 1 pt: 22/40 = 55%
Calling gets sweep: 8/40 = 20%
EO of calling: .45
EO of passing: .775
EV of calling: -.325

So far calling is the losing play even with the strongest possible calling hand from this configuration. Will keep going until I cross that 95% confidence interval.

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 7:54 am
by Wes (aka the legend)
Just did 20 more hands and things have swung in a different direction.

(Card_K-H) (Card_Q-H) (Card_10-H) (Card_A-D) (Card_K-D)

Vs (Card_9-H)

Sample size: 60
-1,1,-3,0,-3,-1,-1,3,-1,-1,0,0,0,4,-3,-1,-4,-1,0,1,-1,0,0,-1,-4,1,2,1,-1,-1,-1,3,0,0,-1,-1,0,1,0,1,1,-1,4,3,-2,4,0,3,-1,0,1,3,-1,-4,3,4,1,3,0,1
Mean of calling: .15
Standard deviation: 1.9985
P value: .5632

Calling gets euchred: 14/60 = 23.33%
Calling gets 1 pt: 27/60 = 45%
Calling gets sweep: 19/60 = 31.67%
EO of calling: .6167
EO of passing: .4667
EV of calling: +.15

I think calling is running hot here. No way that sweep rate is sustainable. I still suspect passing will be the winner here but who knows. Maybe this is another break-even spot. Will keep going.

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 9:50 am
by irishwolf
Just to confirm so we are ALL on the same page, the results is S3 calling R1 vs S1 calling Next Diamonds R2 every hand? I agree sweeps at 31.67 is lopsided - 18 to 22% is more like it for sweeps and euchre rate 22 to 26%. Just knowing how my similar hands came out (300 sample size). It would be very difficult for S2/S4 to bag against this hand. COOL! And with this holding if S1 were to Pass-Pass, watch out for the loners at S2.


Just did 20 more hands and things have swung in a different direction.

(Card_K-H) (Card_Q-H) (Card_10-H) (Card_A-D) (Card_K-D)

Vs (Card_9-H)

Sample size: 60

P.S. Just me, but I don't trust any results 50 or less sample size (you, me or anybody). The standard error is just too high with a small sample size. Always look at the range the t-test gives you for the Confidence Interval at 95%. Precision around the mean increases as the sample size increases. Especially when the means of the test sample is close to the that being tested against.

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 2:46 pm
by Wes (aka the legend)
irishwolf wrote:
Fri Apr 02, 2021 9:50 am
Just to confirm so we are ALL on the same page, the results is S3 calling R1 vs S1 calling Next Diamonds R2 every hand? I agree sweeps at 31.67 is lopsided - 18 to 22% is more like it for sweeps and euchre rate 22 to 26%. Just knowing how my similar hands came out (300 sample size). It would be very difficult for S2/S4 to bag against this hand. COOL! And with this holding if S1 were to Pass-Pass, watch out for the loners at S2.
No. The better way to phrase it: The result represents what's better between S3 calling or passing assuming I am your partner in S1. S1 plays just like me. I do not call Next R2 every hand. This is especially informative to Edward given that I AM so often his P. He'll be able to precisely know which hands are -EV or not in this spot for our cash game. I believe my S1 play approximates expert play close enough to make these results very relevant. E.G. we can make conclusions like: If my partner is a very strong player passing with this hand from S3...

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

Vs (Card_K-H)

...is better than calling.

Or, EVEN IF my partner is a strong player, calling with this hand is better than passing:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

Vs (Card_K-H)

In S1-R2 I will jump the fence more often than most people.

For example after the dealer passes a heart, If I have this hand:

(Card_K-S) (Card_Q-S) (Card_9-S) (Card_Q-D) (Card_10-H)

I'm calling spades

But change that hand to this:

(Card_K-S) (Card_Q-S) (Card_Q-D) (Card_10-H) (Card_10-C)

I'm calling Next (diamonds)

Or this:

(Card_K-S) (Card_Q-S) (Card_9-S) (Card_10-D) (Card_9-D)

I'm calling Next (diamonds)

So I will jump the fence with any 3 trump depending on my other cards. I will also jump the fence with R+1+A depending on my other cards. And rarely, I will jump the fence with L+1+2 or 3 off aces. I'm capable of jumping the fence with R+1+0 too depending on my other cards but that usually only happens when I'm bagging 3 trump in R1 which is impossible in this hand cuz the only 3 trump I can bag is JhJdAh and If I bad that hand it will be cuz I have a better hand in Next.

So the bottom line is, S3 passing in R1 isn't always gonna work when I'm your P. It will cause some euchres when I jump the fence. Whenever you see a big number in the sample box (+3 or +4) thats probably what happened (or I called Next super weak). I.E. S3 passed a heart call that would've made 1 or 2 pts and then the dealer passed and I crossed the river and got euchred. Even if one disagrees with some of my jumping the fence hands they should appreciate the fact that that just makes my sample result even stronger when S3 passing beats out calling. The hope is that passing a 3S heart call with an outside AdKd will still beat out calling becuz I call Next often enough to make up for those times I cross the river and get destroyed.
irishwolf wrote:
Fri Apr 02, 2021 9:50 am
P.S. Just me, but I don't trust any results 50 or less sample size (you, me or anybody). The standard error is just too high with a small sample size. Always look at the range the t-test gives you for the Confidence Interval at 95%. Precision around the mean increases as the sample size increases. Especially when the means of the test sample is close to the that being tested against.
That sounds good. If it turns out that passing doesn't beat out calling with KhQhThAdKd vs 9H then I'll go back and add to my 20 sample of QhTh9hAdKd

PS: if you're wondering what I would do with any hand configuration from S1-R2 just post the hand here and I'll tell you exactly what I would do. That might help us get on the same page.

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 3:53 pm
by irishwolf
The comment I made, was only in reference to your results as posted, just the one hand results of QH 10H 9H AD KD , AH UP, S3 ordering (not what you would actually do) vs S1 calling Next, Diamonds, R2 every hand? That is important as it only confirms Passing with that hand and S1 calling Next is a sound strategy.

Other than that, I pretty much know what and how you play. That's my job!

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 6:16 pm
by Tbolt65
Wes (aka the legend) wrote:
Fri Apr 02, 2021 2:46 pm
irishwolf wrote:
Fri Apr 02, 2021 9:50 am
Just to confirm so we are ALL on the same page, the results is S3 calling R1 vs S1 calling Next Diamonds R2 every hand? I agree sweeps at 31.67 is lopsided - 18 to 22% is more like it for sweeps and euchre rate 22 to 26%. Just knowing how my similar hands came out (300 sample size). It would be very difficult for S2/S4 to bag against this hand. COOL! And with this holding if S1 were to Pass-Pass, watch out for the loners at S2.
No. The better way to phrase it: The result represents what's better between S3 calling or passing assuming I am your partner in S1. S1 plays just like me. I do not call Next R2 every hand. This is especially informative to Edward given that I AM so often his P. He'll be able to precisely know which hands are -EV or not in this spot for our cash game. I believe my S1 play approximates expert play close enough to make these results very relevant. E.G. we can make conclusions like: If my partner is a very strong player passing with this hand from S3...

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

Vs (Card_K-H)

...is better than calling.

Or, EVEN IF my partner is a strong player, calling with this hand is better than passing:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

Vs (Card_K-H)

In S1-R2 I will jump the fence more often than most people.

For example after the dealer passes a heart, If I have this hand:

(Card_K-S) (Card_Q-S) (Card_9-S) (Card_Q-D) (Card_10-H)

I'm calling spades

But change that hand to this:

(Card_K-S) (Card_Q-S) (Card_Q-D) (Card_10-H) (Card_10-C)

I'm calling Next (diamonds)

Or this:

(Card_K-S) (Card_Q-S) (Card_9-S) (Card_10-D) (Card_9-D)

I'm calling Next (diamonds)

So I will jump the fence with any 3 trump depending on my other cards. I will also jump the fence with R+1+A depending on my other cards. And rarely, I will jump the fence with L+1+2 or 3 off aces. I'm capable of jumping the fence with R+1+0 too depending on my other cards but that usually only happens when I'm bagging 3 trump in R1 which is impossible in this hand cuz the only 3 trump I can bag is JhJdAh and If I bad that hand it will be cuz I have a better hand in Next.

So the bottom line is, S3 passing in R1 isn't always gonna work when I'm your P. It will cause some euchres when I jump the fence. Whenever you see a big number in the sample box (+3 or +4) thats probably what happened (or I called Next super weak). I.E. S3 passed a heart call that would've made 1 or 2 pts and then the dealer passed and I crossed the river and got euchred. Even if one disagrees with some of my jumping the fence hands they should appreciate the fact that that just makes my sample result even stronger when S3 passing beats out calling. The hope is that passing a 3S heart call with an outside AdKd will still beat out calling becuz I call Next often enough to make up for those times I cross the river and get destroyed.
irishwolf wrote:
Fri Apr 02, 2021 9:50 am
P.S. Just me, but I don't trust any results 50 or less sample size (you, me or anybody). The standard error is just too high with a small sample size. Always look at the range the t-test gives you for the Confidence Interval at 95%. Precision around the mean increases as the sample size increases. Especially when the means of the test sample is close to the that being tested against.
That sounds good. If it turns out that passing doesn't beat out calling with KhQhThAdKd vs 9H then I'll go back and add to my 20 sample of QhTh9hAdKd

PS: if you're wondering what I would do with any hand configuration from S1-R2 just post the hand here and I'll tell you exactly what I would do. That might help us get on the same page.
You give simplistic views of not calling next. Every single one I would agree with you of what you shown. However the following I see from you too much from seat 1 and especially from seat 2.


We'll use hearts again here turned down.

You will have something like,

(Card_A-S) (Card_J-S) (Card_K-D) (Card_10-D) (Card_K-H) in the first seat and you'll call spades. It really frustrating when you do this. Especially when You know how much I'm bagging. Your mathematical mind does not allow you to play for your partner at times.


Or say in seat 2, and I turn down hearts and you have...

(Card_A-D) (Card_J-D) (Card_9-S) (Card_10-S) (Card_10-H)

You call diamonds.



Sometimes you make the point other times no. I dont fault you for calling something. I do fault you for not trying to play to my strenghths, especially when I turn down red. The great majority of the time Ill have something in black. Very rarely will I ever have red after turning red down, unless Im bagging or it just how the cards are dealt. This is very rare.


Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 02, 2021 6:23 pm
by LeftyK
I think that's why Edward and I play well together. "play for your partner" aka Hoyle

Re: ITT We Test 3rd Seat Hands

Posted: Sat Apr 03, 2021 2:45 am
by Wes (aka the legend)
Tbolt65 wrote:
Fri Apr 02, 2021 6:16 pm
You give simplistic views of not calling next. Every single one I would agree with you of what you shown. However the following I see from you too much from seat 1 and especially from seat 2.


We'll use hearts again here turned down.

You will have something like,

(Card_A-S) (Card_J-S) (Card_K-D) (Card_10-D) (Card_K-H) in the first seat and you'll call spades. It really frustrating when you do this. Especially when You know how much I'm bagging. Your mathematical mind does not allow you to play for your partner at times.
Yes I'll call spades in that spot (if the turned down card was the AH I would call Next for sure tho). There's no reason to get worked up over it tho. 1) No one really knows what does better between jumping the fence on a Spade call with AsJs and a Next call with two low trump + a doubleton off ace even assuming our bag strategy from 3S-R1. 2) This spot comes up very rarely. Like probably around 1-2 times per 1,000 games. That's just a guess but I feel like I'm barely ever in this spot and major grain of salt here but I don't really recall this call burning me much but who knows. 3) You're the one so worried about my S1-R2 strategy being exploitable. Surely an awesome way to help balance my range and make me less exploitable would be to jump the fence when the EV is close between competing lines.

Either way, there's no real reason to have strong thoughts in this spot, and sadly that may the case forever. Unfortunately 2nd rounds spots are too tedious for humans to test out at the kitchen table (or at least for this human). Like for this spot you'd have to basically keep dealing with a heart upcard in the first rd reshuffling every time someone had a legit heart call becuz the distribution would only be right in the 2nd rd if everyone passed hearts in the 1st. We actually need a computer simulation to test out 2nd rd strategies.
Tbolt65 wrote:
Fri Apr 02, 2021 6:16 pm
Or say in seat 2, and I turn down hearts and you have...

(Card_A-D) (Card_J-D) (Card_9-S) (Card_10-S) (Card_10-H)

You call diamonds.



Sometimes you make the point other times no. I dont fault you for calling something. I do fault you for not trying to play to my strenghths, especially when I turn down red. The great majority of the time Ill have something in black. Very rarely will I ever have red after turning red down, unless Im bagging or it just how the cards are dealt. This is very rare.


Tbolt65
Edward
Oh yes I'm definitely calling diamonds in that spot (that's a bad hand example tho cuz you have me passing a biddable hand in the first rd! No way I'm passing L+1+A in hearts when I only block 1 out of 3 2nd rd suits. That's an armature hour pass). Besides that derail, I'm not convinced at all that a spade call does better. It's important not to get too religiously tied to Hoyle and just assume something is wrong when we don't really know. Again, no reason to get worked up over this spot. It doesn't happen often, and no one really knows what's best. I'm betting a diamond call is best here obviously.

Kind've an irrelevant tidbit but If I'm in S1 and pass in that spot, I believe calling diamonds is unequivocally the best call for S2. Probably the case if any strong, aggressive player passes in S1. Something for you guys to chew on.

PS: Again it's worth pointing out, me jumping the fence with "dubious" holdings is actually a good thing for my sample. If S3 passing beats out calling, it makes that result THAT MUCH stronger.

Re: ITT We Test 3rd Seat Hands

Posted: Sat Apr 03, 2021 3:40 am
by Tbolt65
Wes (aka the legend) wrote:
Sat Apr 03, 2021 2:45 am
Tbolt65 wrote:
Fri Apr 02, 2021 6:16 pm
You give simplistic views of not calling next. Every single one I would agree with you of what you shown. However the following I see from you too much from seat 1 and especially from seat 2.


We'll use hearts again here turned down.

You will have something like,

(Card_A-S) (Card_J-S) (Card_K-D) (Card_10-D) (Card_K-H) in the first seat and you'll call spades. It really frustrating when you do this. Especially when You know how much I'm bagging. Your mathematical mind does not allow you to play for your partner at times.
Yes I'll call spades in that spot (if the turned down card was the AH I would call Next for sure tho). There's no reason to get worked up over it tho. 1) No one really knows what does better between jumping the fence on a Spade call with AsJs and a Next call with two low trump + a doubleton off ace even assuming our bag strategy from 3S-R1. 2) This spot comes up very rarely. Like probably around 1-2 times per 1,000 games. That's just a guess but I feel like I'm barely ever in this spot and major grain of salt here but I don't really recall this call burning me much but who knows. 3) You're the one so worried about my S1-R2 strategy being exploitable. Surely an awesome way to help balance my range and make me less exploitable would be to jump the fence when the EV is close between competing lines.

It comes up often enough for me to comment on. Plus your not thinking 4th Dimensionally(Back to the future reference) Meaning as I mentioned below your not taking into consideration my holdings. What I passed with and what I may/could be and should be holding. Your only considering yours. You see you have a good opening lead. You have A decent doubleton suited Ace, plus A void. The king of hearts is a possible winning hand as well. With my range of bagging. It's an easy next call. Not just for me but for anyone.

Either way, there's no real reason to have strong thoughts in this spot, and sadly that may the case forever. Unfortunately 2nd rounds spots are too tedious for humans to test out at the kitchen table (or at least for this human). Like for this spot you'd have to basically keep dealing with a heart upcard in the first rd reshuffling every time someone had a legit heart call becuz the distribution would only be right in the 2nd rd if everyone passed hearts in the 1st. We actually need a computer simulation to test out 2nd rd strategies.

I'm not just some guy that says, this might work. Let it go Luke(wes). Use the force(trust your p). Ahh I love when starwars analogies meshes with euchre. 8-)
Tbolt65 wrote:
Fri Apr 02, 2021 6:16 pm
Or say in seat 2, and I turn down hearts and you have...

(Card_A-D) (Card_J-D) (Card_9-S) (Card_10-S) (Card_10-H)

You call diamonds.



Sometimes you make the point other times no. I dont fault you for calling something. I do fault you for not trying to play to my strenghths, especially when I turn down red. The great majority of the time Ill have something in black. Very rarely will I ever have red after turning red down, unless Im bagging or it just how the cards are dealt. This is very rare.


Tbolt65
Edward
Oh yes I'm definitely calling diamonds in that spot (that's a bad hand example tho cuz you have me passing a biddable hand in the first rd! No way I'm passing L+1+A in hearts when I only block 1 out of 3 2nd rd suits. That's an armature hour pass).

I've passed that hand so many times to you. You go alone often enough for one. Secondly your picking up quite frequently and if not. If it get to me I'm 100% calling spades for you, for my partner. Even if you are stronger in something else I have enough to get the lead first and to pull trump and we are still have a decent shot at making point on the rare occasion you may be heavy something else.


Besides that derail, I'm not convinced at all that a spade call does better. It's important not to get too religiously tied to Hoyle and just assume something is wrong when we don't really know. Again, no reason to get worked up over this spot. It doesn't happen often, and no one really knows what's best. I'm betting a diamond call is best here obviously.

I understand not to get too tied to hoyle. I'm not saying to never call diamonds but you should be heavily considering what I may have to help your or not. If you eliminate that thought process from any seat with me as your partner or any thinking partner for that matter. Seriously you are no better in terms of thinking wise compared to people that need a certain handset to call or they won't call at all, Ie: 3 trump and Ace type mentality. Thinking is good, Going with your gut is good. Playing your hand based on some playing experience and how you can possibly make it work is good, but Never playing for/to your partner is bad.

I believe playing on Karman has given you some false sense. That app caters to weaker players and it's probably the worst euchre base of players I have EVER seen!!!! Now listen to me carefully here. I hope your mind is not so distracted from the last paragraph as to not to see what I'm saying in this one. In your break down here with-in, in your reply to me. If you were playing on Karman with those players. I would 10000000000000% agree with your playing and reasons. However I think this is where you have been lacking in terms of adjustments and playing for your partner since you have had to be on the more aggressive side and play and rely on your hand and your hand alone and some how make the best of it with the weaker players. Now I may not be the best player around but I am far from weak and learning to play off your partner and to their strengths and to rely a little more on hoyle at times. It's going make a lot of the guessing seem more seemless and true. Look when you call something and I'm expecting a particular call. I'm smart enough to know you called something for a reason and by not playing by hoyle or for your partner I'm going to help out the best I can and the interplay that comes with it. You see with me or a thinking partner that can follow your play. With our play you know we will play optimally in most situations baring a brainfart or a misclick. This will help your non-hoyle calls out when you miss us. With you expert playing your hands to give us chances to take tricks and what not there is a fair amount of times you'll be able to eeek out a point or maybe get two if the opponents misplay. When two competent, thinking players make a play or call that is different we have to know its for a reason. Whether that be me or you and we should then adjust accordingly and to the best we can. But I'm telling you. You have to play to/for your partner more often, plz, for me? It will be the difference between your Icecream and Marishia's. Both are great tasting Icecream but yours is soo much sweeter. And if you start to play to/for your partner more often than not the results will be that much sweeter too.


Kind've an irrelevant tidbit but If I'm in S1 and pass in that spot, I believe calling diamonds is unequivocally the best call for S2. Probably the case if any strong, aggressive player passes in S1. Something for you guys to chew on.

PS: Again it's worth pointing out, me jumping the fence with "dubious" holdings is actually a good thing for my sample. If S3 passing beats out calling, it makes that result THAT MUCH stronger.


Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Sat Apr 03, 2021 4:27 am
by Wes (aka the legend)
Tbolt65 wrote:
Sat Apr 03, 2021 3:40 am
It comes up often enough for me to comment on. Plus your not thinking 4th Dimensionally(Back to the future reference) Meaning as I mentioned below your not taking into consideration my holdings. What I passed with and what I may/could be and should be holding. Your only considering yours. You see you have a good opening lead. You have A decent doubleton suited Ace, plus A void. The king of hearts is a possible winning hand as well. With my range of bagging. It's an easy next call. Not just for me but for anyone.
I take into consideration ALL variables. You know this. I just so happen to disagree with you on this one.
Tbolt65 wrote:
Sat Apr 03, 2021 3:40 am
I'm not just some guy that says, this might work. Let it go Luke(wes). Use the force(trust your p). Ahh I love when starwars analogies meshes with euchre. 8-)
Your opinion is noted but I need hard data to move the needle on this one.
Tbolt65 wrote:
Sat Apr 03, 2021 3:40 am
I've passed that hand so many times to you. You go alone often enough for one. Secondly your picking up quite frequently and if not.

I hate that pass, but maaaybe if I'm your P it miiight be ok. The good news is this claim is actually testable. The bad news is this test is way down on my priority list. Although I will say in our tournament this is a clear call as my P. My alone range shrinks considerably when getting euchred costs my team 4 pts. Your idea could have merit when the opposing team has 8+ since getting euchred (-2 or -4) is the same result at that score and thus I can open my range back up as the dealer.
Tbolt65 wrote:
Sat Apr 03, 2021 3:40 am
I understand not to get too tied to hoyle. I'm not saying to never call diamonds but you should be heavily considering what I may have to help your or not. If you eliminate that thought process from any seat with me as your partner or any thinking partner for that matter. Seriously you are no better in terms of thinking wise compared to people that need a certain handset to call or they won't call at all, Ie: 3 trump and Ace type mentality. Thinking is good, Going with your gut is good. Playing your hand based on some playing experience and how you can possibly make it work is good, but Never playing for/to your partner is bad.

I consider all variables behind every decision I make. Do you know who you are talking to????
Tbolt65 wrote:
Sat Apr 03, 2021 3:40 am
I believe playing on Karman has given you some false sense. That app caters to weaker players and it's probably the worst euchre base of players I have EVER seen!!!!

Meh virtually all players suck on every app. For 99% of players, euchre is just a game to pass the time. Only really strange people take this game seriously (No offense guys :) )
Tbolt65 wrote:
Sat Apr 03, 2021 3:40 am
Now listen to me carefully here. I hope your mind is not so distracted from the last paragraph as to not to see what I'm saying in this one. In your break down here with-in, in your reply to me. If you were playing on Karman with those players. I would 10000000000000% agree with your playing and reasons. However I think this is where you have been lacking in terms of adjustments and playing for your partner since you have had to be on the more aggressive side and play and rely on your hand and your hand alone and some how make the best of it with the weaker players. Now I may not be the best player around but I am far from weak and learning to play off your partner and to their strengths and to rely a little more on hoyle at times. It's going make a lot of the guessing seem more seemless and true. Look when you call something and I'm expecting a particular call. I'm smart enough to know you called something for a reason and by not playing by hoyle or for your partner I'm going to help out the best I can and the interplay that comes with it. You see with me or a thinking partner that can follow your play. With our play you know we will play optimally in most situations baring a brainfart or a misclick. This will help your non-hoyle calls out when you miss us. With you expert playing your hands to give us chances to take tricks and what not there is a fair amount of times you'll be able to eeek out a point or maybe get two if the opponents misplay. When two competent, thinking players make a play or call that is different we have to know its for a reason. Whether that be me or you and we should then adjust accordingly and to the best we can. But I'm telling you. You have to play to/for your partner more often, plz, for me? It will be the difference between your Icecream and Marishia's. Both are great tasting Icecream but yours is soo much sweeter. And if you start to play to/for your partner more often than not the results will be that much sweeter too.

Are you f**king kidding me. Did you like forget who I am or something. No one has called Next or reverse Next more for you in your euchre life than me. I mean I'm not even sure about a lot of my BS hoyle calls, but I make them. You're literally nitpicking 1% of my range and just ignoring the 99%. And to top it off the difference in EV between the choices of that 1% are probably close enough to not even matter. And yes I do make the best plant-based ice cream in the state of Nevada!! The secret game changing ingredient behind awesome plant-based ice cream and shakes is this:



post edited by Dlan due to violation of forum rules.

Re: ITT We Test 3rd Seat Hands

Posted: Sat Apr 03, 2021 4:44 am
by Tbolt65
I do realize who you are and what you have been capable of and what you have done. What you call nitpicking may be just that but I'm trying to tune you up so that not only you can play a middle B but also a B#. In other words to make you a well oiled machine. Its the nitpicking in the end that will keep our lines taught and straight. So once the key is struck we will be a finely tuned well oiled machine. That will be a tough cookie to crack. Its the little things in the end that makes the real magic happen. We both come from two different approaches in euchre but I believe we both have learned from each other and come from the same cloth of perfection. The more we keep collaborating and coming to understandings. The better team and players we shall become.

Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 3:26 am
by Tbolt65
(Card_A-S) (Card_9-S) (Card_Q-H) (Card_10-H) (Card_9-H)




Here are my Kitchen table play that I did tonight. I played 50 hands and included the times 2nd seat would order as well. First seat didnt get a chance to order. However with all this being said. about 3/4 of the way through It occured to me I didn't account for passing in 3rd seat to see what happens when dealer either picked up and passed or if Seat 1 called. What I do know is I saw lots of loner hands in seat one for next and strong hands for black to in seat 1. Over all a lot of strong hands just on observation in seat 1. Anyways Just to let you all know, there is still a Looooooooooooooooooooooong way to go. That has to be accounted for.

Dealt 50 hands

Seat 2 orders up 6 times and makes it all 6pts. No Euchre's for seat 3s Team.

1 pt 26hands 59.09% out of 44 hands dealt that S3 got to make it.

2pt Marches 6hands 13.63% out of 44 hands dealt that S3 got to make it.

2pt got Euchred 12hands 27.27% out of the 44 hands dealt S3 got to make it.


+38points Marches and single points
-24 for Euchre's

0pts 6 times on Seat 2 order up.



Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 3:43 am
by Wes (aka the legend)
Tbolt65 wrote:
Mon Apr 05, 2021 3:26 am
(Card_A-S) (Card_9-S) (Card_Q-H) (Card_10-H) (Card_9-H)

Here are my Kitchen table play that I did tonight. I played 50 hands and included the times 2nd seat would order as well. First seat didnt get a chance to order. However with all this being said. about 3/4 of the way through It occured to me I didn't account for passing in 3rd seat to see what happens when dealer either picked up and passed or if Seat 1 called. What I do know is I saw lots of loner hands in seat one for next and strong hands for black to in seat 1. Over all a lot of strong hands just on observation in seat 1. Anyways Just to let you all know, there is still a Looooooooooooooooooooooong way to go. That has to be accounted for.

Dealt 50 hands

Seat 2 orders up 6 times and makes it all 6pts. No Euchre's for seat 3s Team.

1 pt 26hands 59.09% out of 44 hands dealt that S3 got to make it.

2pt Marches 6hands 13.63% out of 44 hands dealt that S3 got to make it.

2pt got Euchred 12hands 27.27% out of the 44 hands dealt S3 got to make it.


+38points Marches and single points
-24 for Euchre's

0pts 6 times on Seat 2 order up.



Tbolt65
Edward
When S2 or S1 has a calling hand you have to scratch that hand and reshuffle to make sure the distribution is right for S3. So you basically have a sample of 44 here which is fine.

The EO of calling here is 14/44 = +.3182 which doesn't tell us much cuz we have to compare that number to the EO of S3 passing to find out the EV. But based on the work I've done on QhTh9hAsKs and QhT9hAs9s passing shows a slight -EO. I believe Irishwolf's work also shows that. So your work backs up the claim that calling in S3 is better than passing with all 3 low heart + suited green ace hands.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 4:02 am
by Tbolt65
It'll be interesting to see what happens on passes, with dealer picking up or seat 1calls. My observation and intuition tells me their might be slightly more euchres but there will be also more loner opportunities. Will this be strong enough to offset the euchres? I'm not sure. I dont excpect a whole lot of marches but there should be atlreast some. Plenty of 1 pt hands. The difference maker is going to be how many loners get made. With how many euchres.


Side note: The more errors made by seat 1 and 3, the more times you will be euchred. So that must be a consideration when calling hands like these or any hand for that matter. Certain Leads, playing off/sloughing off are the difference maker.


Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 4:03 am
by Wes (aka the legend)
(Card_K-H) (Card_Q-H) (Card_10-H) (Card_A-D) (Card_K-D)

Vs (Card_9-H)

Sample size: 100
-1,1,-3,0,-3,-1,-1,3,-1,-1,0,0,0,4,-3,-1,-4,-1,0,1,-1,0,0,-1,-4,1,2,1,-1,-1,-1,3,0,0,-1,-1,0,1,0,1,1,-1,4,3,-2,4,0,3,-1,0,1,3,-1,-4,3,4,1,3,0,1,3,0,4,3,-1,-1,2,0,2,-1,-1,-1,-1,-1,-3,0,0,-4,0,-1,3,2,-1,-1,0,-4,2,0,3,0,-4,3,3,-1,-1,-1,1,-4,-1,0
Mean of calling: .07
Standard deviation: 2.0363
P value: .7318

Calling gets euchred: 30/100 = 30%
Calling gets 1 pt: 48/100 = 48%
Calling gets sweep: 22/100 = 22%
EO of calling: .32
EO of passing: .25
EV of calling: +.07

After a hundred samples, looks like we have another statistical tie. I have no clue which strategy is best in the long run with this hand. We're not even close to reaching a 95% confidence interval. If this hand is a break-even spot, then that would suggest that passing will beat out calling with all worse 3 trump + AdXd holdings. Remember, KhQhThAdKd vs 9h is the best possible hand for the calling strategy. If that hand is virtually break-even with passing then the rest follows.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 4:30 am
by Wes (aka the legend)
Tbolt65 wrote:
Mon Apr 05, 2021 4:02 am
It'll be interesting to see what happens on passes, with dealer picking up or seat 1calls. My observation and intuition tells me their might be slightly more euchres but there will be also more loner opportunities. Will this be strong enough to offset the euchres? I'm not sure. I dont excpect a whole lot of marches but there should be atlreast some. Plenty of 1 pt hands. The difference maker is going to be how many loners get made. With how many euchres.
One thing I've learned the hard way is not to trust my intuition in this spot. When you did your sample you should've played out both strategies. It's not too hard. Deal the cards and play out what happens when S3 calls in your mind and record the result, then play the cards out after S3 passes and record that result. If the hand is too complex to play out in your mind or your brain is just not working right (this happens to me sometimes) then write the hands down on a piece of paper so you can't forget who has what, then play the cards out for S3 calling and passing and record the results accordingly.
Tbolt65 wrote:
Mon Apr 05, 2021 4:02 am
Side note: The more errors made by seat 1 and 3, the more times you will be euchred. So that must be a consideration when calling hands like these or any hand for that matter. Certain Leads, playing off/sloughing off are the difference maker.


Tbolt65
Edward
I agree that the nature of our P could possibly change things but I'm not worried about that. What I care about is what is the best strategy in a tough game. Also, since you know I'm playing as myself in all seats these results should really speak to us specifically. Since your game is very similar to mine, these results tell us what how we should play when we are partners. Every hole we have in our game as partners can be plugged using this approach. And it is my goal to have every hole eliminated in less than a year. So many samples to do, so little time.

Also, even when I play in a standard weak game on the app, I don't adjust my strategy at all. I'm always making decisions as if I'm in a tough game? Why? becuz who cares if my results are slightly worse on the app. It's not like I'm playing for money. I'd rather always be preparing for that tough game. What about in our tournament where most people don't play well in the 2nd round? Well given that your typical S1 in the tournament is not gonna play sound defense in the 2nd rd, I suspect this should make us even more likely to call than pass with these marginal S3 hands. Perhaps even hands like QhTh9hAdKd is a call with a typical P? The thing is tho we don't have to wonder about that either. That's totally testable. It's not hard to assume S1 is your typical player who needs a real hand to call/just tries not to get euchred and then do the sample from there and see what happens. I think we all have a very good idea on how your typical player plays. The thing is tho, that's WAY down on my priority list. The first thing I want to establish is what is best in S3 assuming you're in that tough game or assuming I'm your P.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 7:08 am
by Wes (aka the legend)
Back to this hand. Following Wolf's advice to do at least 50 samples. Did 70 so far:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

Vs (Card_K-H)

Sample size: 70
1,-1,0,-1,0,-1,0,-1,-1,1,-1,-1,0,1,-4,-2,-2,-1,-1,0,-1,-1,-1,-3,0,1,0,-3,-1,-1,-1,2,-1,-1,-1,1,1,1,-1,-1,3,-1,0,2,2,3,0,2,-1,-2,-1,0,-2,0,4,-1,1,1,-1,3,0,0,-3,-1,0,-1,-1,-1,-1,-1
Mean of calling: -.3143
Standard deviation: 1.4797
P value: .0800

Calling gets euchred: 25/70 = 35.71%
Calling gets 1 pt: 30/70 = 42.86%
Calling gets sweep: 15/70 = 21.43%
EO of calling: .1429
EO of passing: .4571
EV of calling: -.3143

Passing is still better than calling but now I have to keep going until I reach a 95% confidence interval again.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 1:13 pm
by Tbolt65
Wes (aka the legend) wrote:
Mon Apr 05, 2021 4:30 am
Tbolt65 wrote:
Mon Apr 05, 2021 4:02 am
It'll be interesting to see what happens on passes, with dealer picking up or seat 1calls. My observation and intuition tells me their might be slightly more euchres but there will be also more loner opportunities. Will this be strong enough to offset the euchres? I'm not sure. I dont excpect a whole lot of marches but there should be atlreast some. Plenty of 1 pt hands. The difference maker is going to be how many loners get made. With how many euchres.
One thing I've learned the hard way is not to trust my intuition in this spot. When you did your sample you should've played out both strategies. It's not too hard. Deal the cards and play out what happens when S3 calls in your mind and record the result, then play the cards out after S3 passes and record that result. If the hand is too complex to play out in your mind or your brain is just not working right (this happens to me sometimes) then write the hands down on a piece of paper so you can't forget who has what, then play the cards out for S3 calling and passing and record the results accordingly.

:arrow: I told you in my data post that it occurred to me more than half way through. So I just continued with as normally. I will however do the seat 3 passing, the dealer order and seat 1 calling. :arrow:

Tbolt65 wrote:
Mon Apr 05, 2021 4:02 am
Side note: The more errors made by seat 1 and 3, the more times you will be euchred. So that must be a consideration when calling hands like these or any hand for that matter. Certain Leads, playing off/sloughing off are the difference maker.


Tbolt65
Edward
I agree that the nature of our P could possibly change things but I'm not worried about that. What I care about is what is the best strategy in a tough game. Also, since you know I'm playing as myself in all seats these results should really speak to us specifically. Since your game is very similar to mine, these results tell us what how we should play when we are partners. Every hole we have in our game as partners can be plugged using this approach. And it is my goal to have every hole eliminated in less than a year. So many samples to do, so little time.

:arrow: I treat every unknown situation as people playing optimally. When they show me they are not I readjust so not as to incur too many extra euchres and I will be ordering more with their lack of calling. Its crucial because you need to maximize/minimize points/euchres with weaker partners. Doing so will add more wins. Best strategy in tough games will change depending on our opponents. There us a time and place for everything :arrow:

The arrows are to denote my new responses since I cant highlight them on my phone. I suppose I could hit the color first but after typing im not going back to retype. Next time I'll do it first.

Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 1:41 pm
by irishwolf
WOW! (QH 10H 9H AD KD - KH UP)
I am looking at the raw data of just what S3 makes for points, your data Wes.

Calling gets euchred: 25/70 = 35.71% -71 + 43 + 42 = + 14
Calling gets 1 pt: 30/70 = 42.86%
Calling gets sweep: 15/70 = 21.43%

Calling gets euchred: 30/100 = 30% - 60 + 48 + 44 = + 32
Calling gets 1 pt: 48/100 = 48%
Calling gets sweep: 22/100 = 22%

It is a little surprising such a difference in euchre rate and makes sense to combine the so the Euchre rate of (combined) 55 / 170 = 32.35%, 78 / 170 = 45.88 and 37 / 170 = 21.76. Then based on 170 hands:
COMBINED 170 HANDS : - 64.7 + 45.88 + 43.52 /170 = + 24.7
CORRECTED to + 24.7

The difference, is pretty big between the samples. However, in my mind that is a good strategy to Order (24.7 points / 100 hands) just based on S3 ordering not considering passing and who occurs after that move.

And now to passing, that is reflected with a huge difference in standard deviation. Your large difference is more reflective of passing than just on S3 raw data of calls of ordering. It becomes (my experience) very complex more on S1 hands of calling than the dealer ordering the KH. And toss in Ed's 50 which is close to the results you got.

So my point is this, for drawing conclusion on S3 pass, S1 making trump you need a lot more data in convincing (at least me) that + 24.7 positive results is a bad call. It becomes complex, WILL S1 more likely call next or JUMP THE FENCE. I would not trust data of A LEAP OF FAITH even though that may occur with players like yourself. And with two diamonds at S3, you will not get very many solid Next hands compared to Crossing Suit hands. Complex, but I think wise, first just look at data first, in the true sense of calling with solid Next or Crossing the suit hands. Then you can access what if just called Next, as a leap of faith. Just me talking! It's your test but I don't know for sure how to interpret what occurs after S1 & S4 passes? And at what Score or Who is Playing.

I also think having one less off suit card (JD vs the hand of JS for AS/KS or AS/9S) is also significant (16.7%) as this prevents opponents from trumping and what the dealer might hold.

Then Ed says, I saw a lot of loners. I reviewed the loners. Successful loner are hard to get, of 75 hands, with S3 hand, I saw 8 (3 with Diamonds, 2 Clubs and 2 Spades). One was 50:50 as S2 got squeezed with two aces. 10.7%. Just look at the loners from Friday's 11 games, 6 of them is what I counted, all hands. Normally, 10.2 hands on the average for a game is 11 is 112 hands - 6 / 112 = 5.4%, right on EV for many games played. So this hand lends to more loners but be careful on the assumption.

From what I see in my hands, S1 has euchre hands, Pass-Pass much of the time. If he calls weak, expect a high euchre rate nullifying many of those loners. It continues to be a complex call if S3 passes. Just saying.

~Irishwolf

I hope I have not confused the issue(s).

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 5:43 pm
by justme
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
WOW! (QH 10H 9H AD KD - KH UP)
I am looking at the raw data of just what S3 makes for points, your data Wes.

Calling gets euchred: 25/70 = 35.71% -71 + 43 + 42 = + 14
Calling gets 1 pt: 30/70 = 42.86%
Calling gets sweep: 15/70 = 21.43%

Calling gets euchred: 30/100 = 30% - 60 + 48 + 44 = + 32
Calling gets 1 pt: 48/100 = 48%
Calling gets sweep: 22/100 = 22%

It is a little surprising such a difference in euchre rate and makes sense to combine the so the Euchre rate of (combined) 55 / 170 = 32.35%, 78 / 170 = 45.88 and 27 / 170 = 15.88. Then based on 170 hands:
COMBINED 170 HANDS : - 64.7 + 91.76 + 31.46 / 170 = + 34.6

The difference, is pretty big between the samples. However, in my mind that is a good strategy to Order (34.6 points / 100 hands) just based on S3 ordering not considering passing and who occurs after that move.
"Euchre rate of (combined) 55 / 170 = 32.35%, 78 / 170 = 45.88 and 27 / 170 = 15.88. Then based on 170 hands:
COMBINED 170 HANDS : - 64.7 + 91.76 + 31.46 / 170 = + 34.6"

Should the above actually read as follows:

Euchre rate of (combined) 55 / 170 = 32.35%, 78 / 170 = 45.88 and 37 / 170 = 21.80 Then based on 170 hands:
COMBINED 170 HANDS : - 64.7 + 45.88 + 43.60/ 170 = + 24.78

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 05, 2021 7:40 pm
by irishwolf
Justme

Corrected as above, Yes Houston we had an issue!

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 06, 2021 6:17 am
by Wes (aka the legend)
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
WOW! (QH 10H 9H AD KD - KH UP)
I am looking at the raw data of just what S3 makes for points, your data Wes.

Calling gets euchred: 25/70 = 35.71% -71 + 43 + 42 = + 14
Calling gets 1 pt: 30/70 = 42.86%
Calling gets sweep: 15/70 = 21.43%

Calling gets euchred: 30/100 = 30% - 60 + 48 + 44 = + 32
Calling gets 1 pt: 48/100 = 48%
Calling gets sweep: 22/100 = 22%

It is a little surprising such a difference in euchre rate

I'm not surprised by the difference actually. KhQhThAdKd vs a 9H getting euchred 30% while QhTh9hAdKd vs a KH gets euchred around 35.71% seems totally plausible to me. Having 3 trump above the upcard should produce better results than 3 trump under the upcard.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
and makes sense to combine the so the Euchre rate of (combined) 55 / 170 = 32.35%, 78 / 170 = 45.88 and 37 / 170 = 21.76. Then based on 170 hands:
COMBINED 170 HANDS : - 64.7 + 45.88 + 43.52 /170 = + 24.7
CORRECTED to + 24.7

I wouldn't combine these hands. I don't think that's useful. They are two different hands.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
The difference, is pretty big between the samples. However, in my mind that is a good strategy to Order (24.7 points / 100 hands) just based on S3 ordering not considering passing and who occurs after that move.
I wouldn't feel safe going with a rule of thumb on the EO of a 3S call strategy. We can do better than that by simply putting the hand to the test and seeing if calling does better than passing or not.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
And now to passing, that is reflected with a huge difference in standard deviation. Your large difference is more reflective of passing than just on S3 raw data of calls of ordering. It becomes (my experience) very complex more on S1 hands of calling than the dealer ordering the KH.

Yes how S1 plays is the key to this sample, but it's not very complex to me. I just play every seat/spot the way I would play it. So these results basically tell us how a S3 pass/call will do in a tough game.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
And toss in Ed's 50 which is close to the results you got.

Ed's hands aren't relevant to this discussion imo since he tested the outside suited As9s while we are talking about the outside AdKd. Two very different hands.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
So my point is this, for drawing conclusion on S3 pass, S1 making trump you need a lot more data in convincing (at least me) that + 24.7 positive results is a bad call.

I think it's always good to be skeptical, especially if you can't easily replicate my results. What I suggest to you if you run this same test is play every seat/spot as yourself and see what happens. Your play is close enough to mine that I would predict you too will find that passing QhTh9hAdKd is better than calling (I'm getting a little ahead of myself since I haven't even reached a 95% confidence interval yet but let's assume it sticks).

What's most important to me tho is not convincing you but actually convincing Edward. Edward is my P in the cash game. If passing a certain hand or calling a certain hand from S3-R1 is +EV when I'm in S1 Edward needs to know about it and make the necessary adjustments. For example, Ed now knows thanks to you that calling QhTh9hAsKs vs Kh is the +EV play when I'm in the S1. Before he was passing as was I. Now presumably he wont. Since Edward plays very similar to me, I'll be calling that hand too confidently expecting the same kind of result.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
It becomes complex, WILL S1 more likely call next or JUMP THE FENCE. I would not trust data of A LEAP OF FAITH even though that may occur with players like yourself. And with two diamonds at S3, you will not get very many solid Next hands compared to Crossing Suit hands. Complex, but I think wise, first just look at data first, in the true sense of calling with solid Next or Crossing the suit hands. Then you can access what if just called Next, as a leap of faith. Just me talking! It's your test but I don't know for sure how to interpret what occurs after S1 & S4 passes? And at what Score or Who is Playing.

I don't really think it's inherently complex. Just play every spot as you would play it. That's all I'm doing. No guesswork. If there's a true 50-50 spot flip a coin or side with the play that benefits the strategy you're trying to disprove to make your result more conservative. I don't even see it as inherently complex if a non-expert is in S1. We just have to formulate his calling/passing/alone standards before we run the test (and I think we all have a good idea on this type of play style since it's the style we see the most of on apps/websites) and then everything becomes automatic again. This is a test I'll have to do in the future to refine my play in my tournament. There's 4 guys in my tournament including Edward I feel confident passing QhTh9hAdKd vs Kh from the 3rd Seat-R1. Vs everyone else I would call cuz I can't trust them in the 2nd rd. But just for the helluva it someday I will test this hand with a non-expert in S1 and see what happens. That's way down the priority list tho.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
I also think having one less off suit card (JD vs the hand of JS for AS/KS or AS/9S) is also significant (16.7%) as this prevents opponents from trumping and what the dealer might hold.

Yea I'm sure that's a factor.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
Then Ed says, I saw a lot of loners. I reviewed the loners. Successful loner are hard to get, of 75 hands, with S3 hand, I saw 8 (3 with Diamonds, 2 Clubs and 2 Spades). One was 50:50 as S2 got squeezed with two aces. 10.7%. Just look at the loners from Friday's 11 games, 6 of them is what I counted, all hands. Normally, 10.2 hands on the average for a game is 11 is 112 hands - 6 / 112 = 5.4%, right on EV for many games played. So this hand lends to more loners but be careful on the assumption.

Having no jacks in my hand was part of the reason I assumed bagging QhTh9hAsKs vs a Kh would do better than calling since there's not only a decent chance S4 would order but also that S1 could be sitting on a 2nd rd loner. Now that we can test hands and falsify such thinking, however reasonable and plausible it may sound, I'm vary wary of any assumptions underlying my play. My ultimate goal is to test out every 1st round grey area scenario. Eliminate any play based on assumption. Every play backed by real evidence. This could take awhile tho, maybe years idk. But that's the goal. We can't do this in the 2nd round imo, but this can be done for the first round. The first round can be completely solved in my mind. It's just gonna take time.
irishwolf wrote:
Mon Apr 05, 2021 1:41 pm
From what I see in my hands, S1 has euchre hands, Pass-Pass much of the time. If he calls weak, expect a high euchre rate nullifying many of those loners. It continues to be a complex call if S3 passes. Just saying.

~Irishwolf

I hope I have not confused the issue(s).
I definitely cross the river a lot, potentially running into a lot of euchres in this spot, and I also call Next super weak depending on my other cards which can also result in quite a few euchres. As I've talked about before, the primary factor governing my play in S1-R2 is NOT Hoyle, but rather what suits I have blocked or not. Hoyle of course plays a factor in my decision making, it's just not the primary factor. The fact that I run into a lot of euchres jumping the fence in this spot or calling Next super weak actually makes the results of my sample that much stronger those times my data reaches the conclusion that passing is better than calling for S3-R1.

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 06, 2021 2:14 pm
by irishwolf
WES, You are correct that one cannot combine test results on Different hands.

I miss read the test trump and the upcard. Quite similar hands in that AD/KD & 3 trumps but not the upcard. Scratch what I said.

I think it would help to start a new thread for each different hand with the hand in the subject heading.

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 06, 2021 11:45 pm
by Tbolt65
Wes (aka the legend) wrote:
Tue Apr 06, 2021 6:17 am




What's most important to me tho is not convincing you but actually convincing Edward. Edward is my P in the cash game. If passing a certain hand or calling a certain hand from S3-R1 is +EV when I'm in S1 Edward needs to know about it and make the necessary adjustments. For example, Ed now knows thanks to you that calling QhTh9hAsKs vs Kh is the +EV play when I'm in the S1. Before he was passing as was I. Now presumably he wont. Since Edward plays very similar to me, I'll be calling that hand too confidently expecting the same kind of result.

As I've said before my euchre isn't was it once was. Since taking 7 years off. This may be an area I may be weak in terms of not remembering correctly and what I had come in terms off ordering and why. What is clear to me is this though. Ordering with an expert vs non expert will vary because you know the likely hood of passing in Seat 1 is very high when you hold 3 trump with A-X. That's why I don't order Left-x or Left-X, off ace. trump ever with you as my partner from the 2nd Seat. I gave mentions to the only cases that I do, but I allow you to go alone or semi donate in dealer seat because thats your job, Not mine in the second seat. I give my partner the chance to call, plus I bag heavy on dealer, if I know the tendencies of dealer and if I know my partner is competent.



Having no jacks in my hand was part of the reason I assumed bagging QhTh9hAsKs vs a Kh would do better than calling since there's not only a decent chance S4 would order but also that S1 could be sitting on a 2nd rd loner. Now that we can test hands and falsify such thinking, however reasonable and plausible it may sound, I'm vary wary of any assumptions underlying my play. My ultimate goal is to test out every 1st round grey area scenario. Eliminate any play based on assumption. Every play backed by real evidence. This could take awhile tho, maybe years idk. But that's the goal. We can't do this in the 2nd round imo, but this can be done for the first round. The first round can be completely solved in my mind. It's just gonna take time.

For you maybe. Here is a question I'll ask you. Why do you go alone with left-9-10 A-K at certain times? When you get that answer you will start to understand why certain gambles are worth taking at the higher level vs not doing the same plays with weaker players from various positions around the table.


Also to take into consideration. About all the data we are taking in from 4 different people. For one we are all playing it differently. There are chances people are playing biased. I've caught myself personally to play not as perfect as I normally would so that I could get that march, euchre or point. I made sure I didn't consciously or unconsciously to favor one way or the other. So I corrected myself as it was unfolding. In Euchre the little things make all the difference and that includes the data. For example I was a good 10% above everyone else in the Q-10-9h As-9s hand. Nearly at 73% sucessful point/march rate. In some case more than 10% above. Now I played how I would play from every position. Not how anyone else would play. That should be very, very telling. With that said, after I get done responding to posts tonight Im going to get another 30 hands in on top of what I did last night for passing in 3rd seat with Q-10-9-h As-9s and continue with that.
I definitely cross the river a lot, potentially running into a lot of euchres in this spot, and I also call Next super weak depending on my other cards which can also result in quite a few euchres. As I've talked about before, the primary factor governing my play in S1-R2 is NOT Hoyle, but rather what suits I have blocked or not. Hoyle of course plays a factor in my decision making, it's just not the primary factor. The fact that I run into a lot of euchres jumping the fence in this spot or calling Next super weak actually makes the results of my sample that much stronger those times my data reaches the conclusion that passing is better than calling for S3-R1.
[/quote]

This is where you should be calling for your partner more often, seat 1 and seat 2 equally. You can go against hoyle. I do but it really has to be strong to do so. Other wise I'm I'm going to take my A-J suited with that 10-9d X off suit after a heart is turned down and call Diamonds/Next all day. Or similar hands of such. Even if I've got one in next with same thing Im calling it unless I have double black with decent cards in both depending on my partner I may just pass there and let it get to my partner and not get us euchred and give up needless points. Some times passing is good and vs weaker players you could get the call to your partner and they call their best and now your in a better spot to help out. Against tougher opponents? Sure you possibly can let them score a point but its better than getting your team euchred and giving up 2pts. The team that makes the least mistakes over all and mistakes in terms of bleeding points away and those who find a way to capitalize on donates for points, euchre's and loners. This is what people need to focus on to put themselves above most as euchre players.

Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Wed Apr 07, 2021 10:06 am
by Wes (aka the legend)
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
As I've said before my euchre isn't was it once was. Since taking 7 years off. This may be an area I may be weak in terms of not remembering correctly and what I had come in terms off ordering and why. What is clear to me is this though. Ordering with an expert vs non expert will vary because you know the likely hood of passing in Seat 1 is very high when you hold 3 trump with A-X. That's why I don't order Left-x or Left-X, off ace. trump ever with you as my partner from the 2nd Seat. I gave mentions to the only cases that I do, but I allow you to go alone or semi donate in dealer seat because thats your job, Not mine in the second seat. I give my partner the chance to call, plus I bag heavy on dealer, if I know the tendencies of dealer and if I know my partner is competent.
As far as L+1+A in the 2S-R1. My claim,or perhaps I should say hypothesis, is that you should order that even with me as your P if you only block 1 out of 3 2nd rd suits. This is something we need not argue about as this spot is totally testable. It's just a matter of when I get the time to do it. So we'll find out about this sooner or later. For now it's to be continued.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
For you maybe. Here is a question I'll ask you. Why do you go alone with left-9-10 A-K at certain times? When you get that answer you will start to understand why certain gambles are worth taking at the higher level vs not doing the same plays with weaker players from various positions around the table.
I suspect L-10-9 with an outside suited AK is a loner from the dealer spot at any score save 8 or 9. That's a bold claim I know. And that's actually one of the loner hands I can't wait to test.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
Also to take into consideration. About all the data we are taking in from 4 different people. For one we are all playing it differently.

That's not a problem to me. In fact I see it as a potential strength. I.E. the fact that you, me, Irishwolf don't play exactly the same. That means if our data reaches the same conclusion, that conclusion is that much stronger. Other than that, my data specifically is tailor made for you since it gives you potentially perfect information on what is +EV and -EV when I am your P. Since we play together all the time, my data has tons of value for us as a team. That's pretty cool.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
There are chances people are playing biased. I've caught myself personally to play not as perfect as I normally would so that I could get that march, euchre or point. I made sure I didn't consciously or unconsciously to favor one way or the other. So I corrected myself as it was unfolding.

Unconscious bias is a potential problem. To combat that problem all you can do is play every hand as you really would play it and in any true 50-50 spot flip a coin or choose the pathway that benefits the strategy you're trying to disprove to make your sample more conservative. What's important here is not to care about the results. Be a slave to the numbers. Ignore your preconceived notions.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
In Euchre the little things make all the difference and that includes the data. For example I was a good 10% above everyone else in the Q-10-9h As-9s hand. Nearly at 73% sucessful point/march rate. In some case more than 10% above. Now I played how I would play from every position. Not how anyone else would play. That should be very, very telling. With that said, after I get done responding to posts tonight Im going to get another 30 hands in on top of what I did last night for passing in 3rd seat with Q-10-9-h As-9s and continue with that.
I don't really see how that's "very, very telling" at all but either way good stuff. That's all you need to do. Play every position as yourself. So far your data appears to strongly favor calling QhTh9hAs9s from S3-R1 which is a cool result.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
This is where you should be calling for your partner more often, seat 1 and seat 2 equally. You can go against hoyle. I do but it really has to be strong to do so. Other wise I'm I'm going to take my A-J suited with that 10-9d X off suit after a heart is turned down and call Diamonds/Next all day. Or similar hands of such. Even if I've got one in next with same thing Im calling it unless I have double black with decent cards in both depending on my partner I may just pass there and let it get to my partner and not get us euchred and give up needless points. Some times passing is good and vs weaker players you could get the call to your partner and they call their best and now your in a better spot to help out. Against tougher opponents? Sure you possibly can let them score a point but its better than getting your team euchred and giving up 2pts. The team that makes the least mistakes over all and mistakes in terms of bleeding points away and those who find a way to capitalize on donates for points, euchre's and loners. This is what people need to focus on to put themselves above most as euchre players.
[/color]
Again, I have to wonder if you know who you are talking to. I call Next and reverse Next more than anyone in the game. Period. Most likely to a fault even.

Assume dealer turns down a spade and I'm in S1. I'm calling next (clubs) with the following hands:

(Card_Q-S) (Card_10-S) (Card_9-S) (Card_10-D) (Card_10-H)

Note: I can only have the above hand if the dealer turned down the JS

(Card_A-D) (Card_K-D) (Card_A-H) (Card_10-H) (Card_10-S)

(Card_9-C) (Card_10-S) (Card_9-S) (Card_9-D) (Card_9-H)

(Card_9-C) (Card_A-D) (Card_9-D) (Card_9-S) (Card_10-S)

That should clue you into how often I'm calling Next. My next calling range is SUPER wide.

Now if I have this hand I'm NOT calling Next:

(Card_K-C) (Card_A-D) (Card_Q-D) (Card_9-D) (Card_10-S)

I'm calling diamonds. IMO anyone who calls Next in that spot is taking Hoyle to a religious degree.

Re: ITT We Test 3rd Seat Hands

Posted: Wed Apr 07, 2021 3:51 pm
by Tbolt65
Wes (aka the legend) wrote:
Wed Apr 07, 2021 10:06 am
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
As I've said before my euchre isn't was it once was. Since taking 7 years off. This may be an area I may be weak in terms of not remembering correctly and what I had come in terms off ordering and why. What is clear to me is this though. Ordering with an expert vs non expert will vary because you know the likely hood of passing in Seat 1 is very high when you hold 3 trump with A-X. That's why I don't order Left-x or Left-X, off ace. trump ever with you as my partner from the 2nd Seat. I gave mentions to the only cases that I do, but I allow you to go alone or semi donate in dealer seat because thats your job, Not mine in the second seat. I give my partner the chance to call, plus I bag heavy on dealer, if I know the tendencies of dealer and if I know my partner is competent.
As far as L+1+A in the 2S-R1. My claim,or perhaps I should say hypothesis, is that you should order that even with me as your P if you only block 1 out of 3 2nd rd suits. This is something we need not argue about as this spot is totally testable. It's just a matter of when I get the time to do it. So we'll find out about this sooner or later. For now it's to be continued.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
For you maybe. Here is a question I'll ask you. Why do you go alone with left-9-10 A-K at certain times? When you get that answer you will start to understand why certain gambles are worth taking at the higher level vs not doing the same plays with weaker players from various positions around the table.
I suspect L-10-9 with an outside suited AK is a loner from the dealer spot at any score save 8 or 9. That's a bold claim I know. And that's actually one of the loner hands I can't wait to test.

I think your missing what I'm getting at.
Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
Also to take into consideration. About all the data we are taking in from 4 different people. For one we are all playing it differently.

That's not a problem to me. In fact I see it as a potential strength. I.E. the fact that you, me, Irishwolf don't play exactly the same. That means if our data reaches the same conclusion, that conclusion is that much stronger. Other than that, my data specifically is tailor made for you since it gives you potentially perfect information on what is +EV and -EV when I am your P. Since we play together all the time, my data has tons of value for us as a team. That's pretty cool.

but only from one viewpoit of ordering and in euchre when we are crunching pts to save on euchres and maximize making/marching. Everything is that much valuable. The lower the rate, the higher the euchres which in turns leads to bleeding unnecessary pts. Plus you got to factor in the pluses of passing which I am now showing in my other thread.


Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
There are chances people are playing biased. I've caught myself personally to play not as perfect as I normally would so that I could get that march, euchre or point. I made sure I didn't consciously or unconsciously to favor one way or the other. So I corrected myself as it was unfolding.

Unconscious bias is a potential problem. To combat that problem all you can do is play every hand as you really would play it and in any true 50-50 spot flip a coin or choose the pathway that benefits the strategy you're trying to disprove to make your sample more conservative. What's important here is not to care about the results. Be a slave to the numbers. Ignore your preconceived notions.

I've done that.

Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
In Euchre the little things make all the difference and that includes the data. For example I was a good 10% above everyone else in the Q-10-9h As-9s hand. Nearly at 73% sucessful point/march rate. In some case more than 10% above. Now I played how I would play from every position. Not how anyone else would play. That should be very, very telling. With that said, after I get done responding to posts tonight Im going to get another 30 hands in on top of what I did last night for passing in 3rd seat with Q-10-9-h As-9s and continue with that.
I don't really see how that's "very, very telling" at all but either way good stuff. That's all you need to do. Play every position as yourself. So far your data appears to strongly favor calling QhTh9hAs9s from S3-R1 which is a cool result.



it could be that Ive had a good sample. However it also could mean I'm playing my cards better as a team to get such a discrepancy.

Tbolt65 wrote:
Tue Apr 06, 2021 11:45 pm
This is where you should be calling for your partner more often, seat 1 and seat 2 equally. You can go against hoyle. I do but it really has to be strong to do so. Other wise I'm I'm going to take my A-J suited with that 10-9d X off suit after a heart is turned down and call Diamonds/Next all day. Or similar hands of such. Even if I've got one in next with same thing Im calling it unless I have double black with decent cards in both depending on my partner I may just pass there and let it get to my partner and not get us euchred and give up needless points. Some times passing is good and vs weaker players you could get the call to your partner and they call their best and now your in a better spot to help out. Against tougher opponents? Sure you possibly can let them score a point but its better than getting your team euchred and giving up 2pts. The team that makes the least mistakes over all and mistakes in terms of bleeding points away and those who find a way to capitalize on donates for points, euchre's and loners. This is what people need to focus on to put themselves above most as euchre players.
[/color]
Again, I have to wonder if you know who you are talking to. I call Next and reverse Next more than anyone in the game. Period. Most likely to a fault even.

I know who I'm talking to.
I'm talking to person who plays nearly just like I do but math oriented. I'm trying to shine up the facets of your diamond playing skills. Maybe I am not conveying properly so you understand and maybe your not grasping because your not wired the way I am.



Assume dealer turns down a spade and I'm in S1. I'm calling next (clubs) with the following hands:

(Card_Q-S) (Card_10-S) (Card_9-S) (Card_10-D) (Card_10-H)

Note: I can only have the above hand if the dealer turned down the JS

(Card_A-D) (Card_K-D) (Card_A-H) (Card_10-H) (Card_10-S)

(Card_9-C) (Card_10-S) (Card_9-S) (Card_9-D) (Card_9-H)

(Card_9-C) (Card_A-D) (Card_9-D) (Card_9-S) (Card_10-S)

That should clue you into how often I'm calling Next. My next calling range is SUPER wide.

Now if I have this hand I'm NOT calling Next:

(Card_K-C) (Card_A-D) (Card_Q-D) (Card_9-D) (Card_10-S)

I'm calling diamonds. IMO anyone who calls Next in that spot is taking Hoyle to a religious degree.

Re: ITT We Test 3rd Seat Hands

Posted: Fri Apr 09, 2021 5:53 am
by Wes (aka the legend)
Reached a 95% confidence interval after 96 hands.

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

Vs (Card_K-H)

Sample size: 96
1,-1,0,-1,0,-1,0,-1,-1,1,-1,-1,0,1,-4,-2,-2,-1,-1,0,-1,-1,-1,-3,0,1,0,-3,-1,-1,-1,2,-1,-1,-1,1,1,1,-1,-1,3,-1,0,2,2,3,0,2,-1,-2,-1,0,-2,0,4,-1,1,1,-1,3,0,0,-3,-1,0,-1,-1,-1,-1,-1,1,0,3,2,0,0,-3,-1,0,0,-1,0,-1,-1,0,-1,-2,-1,0,-1,0,-1,-1,3,0,-1
Mean of calling: -.2917
Standard deviation: 1.4358
P value: .0494

Calling gets euchred: 28/96 = 29.17%
Calling gets 1 pt: 49/96 = 51.04%
Calling gets sweep: 19/96 = 19.79%
EO of calling: .3229
EO of passing: .6146
EV of calling: -.2917

Interesting to note, calling QhTh9hAdKd actually outperformed this hand:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-S) (Card_K-S)

Vs (Card_K-H)

And it even outperformed the strongest possible calling hand from this configuration!:

(Card_K-H) (Card_Q-H) (Card_10-H) (Card_A-D) (Card_K-D)

vs (Card_9-H)

The calling EO of QhTh9hAdKd vs Kh was +.3229. The calling EO of QhTh9hAsKs vs Kh in my sample was +.2571. And the calling EO of KhQhThAdKd vs 9h was .32

IOW calling with:

(Card_Q-H) (Card_10-H) (Card_9-H) (Card_A-D) (Card_K-D)

probably ran a little hot in my sample and YET passing with that hand still clearly did better. I'm not surprised by this result. As I predicted, having a hand that hits a Next call changes things.

What IS surprising: Ok we know the EO of calling ran roughly the same between QhTh9hAdKd vs Kh and KhQhThAdKd vs a 9h (.3229 vs .32). But the difference in the EO of passing for each hand is pretty significant. The EO of passing QhTh9hAdKd vs Kh was +.6146 but the EO of passing KhQhThAdKd vs 9h was only +.25. Notice both hands are exactly equivalent if the hand gets to the 2nd rd, but KhQhThAdKd vs 9h should do slightly better than QhTh9hAdKd vs Kh when the dealer picks up. So we would expect KhQhThAdKd vs 9h to have a slightly higher passing EO but that wasn't the case. So what's going on here. Is KhQhThAdkd vs 9h simply running bad after passing or is QhTh9hAdKd vs Kh simply running hot after passing. Which sample approximates reality better?

How am I going to utilize this information: Well since the best possible calling hand from this configuration KhQhThAdKd vs 9h reached a statistical tie between calling and passing after 100 hands, and passing beat out calling after reaching a 95% confidence interval with QhTh9hAdKd I feel reasonably safe passing all hands from this configuration for now (assuming I have a partner I can trust).

The opposite is true with an outside suited green ace. Since calling clearly beat out passing with QhTh9hAsKs vs Kh when we reached a 95% CI, and I reached a statistical tie between calling and passing with the worst possible calling hand from this configuration, QhTh9hAs9s vs Ah, I feel reasonably safe calling all hands from this configuration for now.

Re: ITT We Test 3rd Seat Hands

Posted: Sun Apr 11, 2021 7:33 pm
by irishwolf
Okay, back to testing results.
One hand outperforming another. I think you have to be very careful about drawing hard conclusions about one hand outperforming another. I have been doing hands and analysis for 20 some years now. I have never done results for less than 150 hands. I will give you explanation below. And it is the reason I do these in sets of 25 hands, hands for each seat in columns and play each seat . Dealing does not always (actually seldom does) give you an even statistical spread and distribution of the KEY cards Jacks, aces and trump cards. For example, take a Jack, over 25 hands, it should be in a perfect world, have a distribution for any specific card per column of 25 hands like this: 6, 6, 7, & 5 to the stock. Take all four Jacks and it will almost never be that way. It could be for that Jack as low as 4 and as high as 9 to one player for 25 hands. So how treat this, to get a completely good statistical test is as follows. Guaranteed that some of those significant cards will vary statistically. Can't be helped.

So on a standard notebook paper, write down the six known cards at the top, draw lines for four columns, 3 for the stock, each column for the 5 cards to each of 3 players. Number the columns 1 , 2, 3. When I do the analysis, I always have 1 vs 2&3, 2 vs 1&3, 3 vs 1&2, all combinations for each hand. (PLAYING FOR AND AGAINST IN ALL COMBINATIONS then average the results. This will smooth out any extremes of variation. But it's more work.) So in effect you have 75 different hands for S1 (with test hand played against S2/S4). S2 and S4 are switched because it does make a different which one dealer and the discard. I have found this smooths out variability in the best fashion.

Thus, with those significant cards you get a complete picture that fairly represents a random distribution. (of course pulling out hands that S2 would have assisted, etc. etc.) But that is just me, I am anal about my tests so I can go back and revisit if something looks suspicious.

Even with just 100 hands you can get some results that can be skewed but less likely the way I have handled this. I tabulate results by 25 sets, so I now have results and believe me, it can vary for each group of 25. Combined, now have a better distribution and results. I hope that makes sense for you.

But there could also be another real reason. With xxx ADKD, the JH is more likely to be with the S1 and JD not so much. But take xxx ASKS or xxxAS9s, the JH is still more likely to be with S1. But JS/JC evenly distributed. This would make weaker Next calls more successful than Black. You would have to have saved all the hands to go back and inspect? Don't know!

Just my thoughts about results that you are surprised about.

~IRISHWOLF

Wes said, " What IS surprising: Ok we know the EO of calling ran roughly the same between QhTh9hAdKd vs Kh and KhQhThAdKd vs a 9h (.3229 vs .32). But the difference in the EO of passing for each hand is pretty significant. The EO of passing QhTh9hAdKd vs Kh was +.6146 but the EO of passing KhQhThAdKd vs 9h was only +.25. Notice both hands are exactly equivalent if the hand gets to the 2nd rd, but KhQhThAdKd vs 9h should do slightly better than QhTh9hAdKd vs Kh when the dealer picks up. So we would expect KhQhThAdKd vs 9h to have a slightly higher passing EO but that wasn't the case. So what's going on here. Is KhQhThAdkd vs 9h simply running bad after passing or is QhTh9hAdKd vs Kh simply running hot after passing. Which sample approximates reality better? "

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 13, 2021 6:33 am
by Wes (aka the legend)
irishwolf wrote:
Tue Apr 06, 2021 2:14 pm
I think it would help to start a new thread for each different hand with the hand in the subject heading.
Forgot to respond to this. The problem is I want to do so many hands I think it would clutter up the forum. I personally think it's best to basically have 4 threads. One thread for testing first seat hands, one for 2S etc.

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 13, 2021 6:46 am
by Wes (aka the legend)
irishwolf wrote:
Sun Apr 11, 2021 7:33 pm
Okay, back to testing results.
One hand outperforming another. I think you have to be very careful about drawing hard conclusions about one hand outperforming another. I have been doing hands and analysis for 20 some years now. I have never done results for less than 150 hands. I will give you explanation below. And it is the reason I do these in sets of 25 hands, hands for each seat in columns and play each seat . Dealing does not always (actually seldom does) give you an even statistical spread and distribution of the KEY cards Jacks, aces and trump cards. For example, take a Jack, over 25 hands, it should be in a perfect world, have a distribution for any specific card per column of 25 hands like this: 6, 6, 7, & 5 to the stock. Take all four Jacks and it will almost never be that way. It could be for that Jack as low as 4 and as high as 9 to one player for 25 hands. So how treat this, to get a completely good statistical test is as follows. Guaranteed that some of those significant cards will vary statistically. Can't be helped.

So on a standard notebook paper, write down the six known cards at the top, draw lines for four columns, 3 for the stock, each column for the 5 cards to each of 3 players. Number the columns 1 , 2, 3. When I do the analysis, I always have 1 vs 2&3, 2 vs 1&3, 3 vs 1&2, all combinations for each hand. (PLAYING FOR AND AGAINST IN ALL COMBINATIONS then average the results. This will smooth out any extremes of variation. But it's more work.) So in effect you have 75 different hands for S1 (with test hand played against S2/S4). S2 and S4 are switched because it does make a different which one dealer and the discard. I have found this smooths out variability in the best fashion.

Thus, with those significant cards you get a complete picture that fairly represents a random distribution. (of course pulling out hands that S2 would have assisted, etc. etc.) But that is just me, I am anal about my tests so I can go back and revisit if something looks suspicious.

Even with just 100 hands you can get some results that can be skewed but less likely the way I have handled this. I tabulate results by 25 sets, so I now have results and believe me, it can vary for each group of 25. Combined, now have a better distribution and results. I hope that makes sense for you.

But there could also be another real reason. With xxx ADKD, the JH is more likely to be with the S1 and JD not so much. But take xxx ASKS or xxxAS9s, the JH is still more likely to be with S1. But JS/JC evenly distributed. This would make weaker Next calls more successful than Black. You would have to have saved all the hands to go back and inspect? Don't know!

Just my thoughts about results that you are surprised about.

~IRISHWOLF
I have no doubt your method is better than my method. But I have to say this is the best I can do for now. I mean overall I still think my method is pretty good. I think running samples until I reach a 95% CI is still gonna way outperform even expert intuition. So my testing approach is still gonna refine my game big time for the better. I agree about the "not drawing hard conclusions" part and the problem of "Dealing does not always (actually seldom does) give you an even statistical spread and distribution".

As said before tho, I still feel confident my approach will beat out expert intuition. To combat the "less than 150 sample/uneven distribution problem" I could simply do more samples for each hand I'm testing. The problem is I'm battling time here. I first wanna reach as many conclusions as I can in the shortest time possible. So when I reach a 95% CI with at least 50 hands, I'm moving to the next hand to test. Right now I wanna plug as many possible leaks in my game as I can asap. Down the road I can always add to my samples to refine things further.

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 13, 2021 12:28 pm
by irishwolf
I AGREE, beats expert intuition and "subjective probability," which humans are not so good at. I have not tried this but possibly a short cut would be in my method but short cut it to 15 hands, then played all combinations. You now have 90 hands played in total. If you do not reach 95% CI then keep going with another until you get there. Just a thought.

Wes said, "As said before tho, I still feel confident my approach will beat out expert intuition. To combat the "less than 150 sample/uneven distribution problem" I could simply do more samples for each hand I'm testing. The problem is I'm battling time here. I first wanna reach as many conclusions as I can in the shortest time possible. So when I reach a 95% CI with at least 50 hands, I'm moving to the next hand to test. Right now I wanna plug as many possible leaks in my game as I can asap. Down the road I can always add to my samples to refine things further."

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 26, 2021 5:35 am
by Wes (aka the legend)
I just did a 50 hand sample of this R1-3S hand:

(Card_A-S) (Card_J-S) (Card_9-S) (Card_Q-C) (Card_9-H)

vs (Card_10-S)

N = 50
-1,3,-1,-1,3,0,-1,-1,-1,0,-3,1,0,-2,0,-4,-4,-4,2,2,1,-1,0,0,0,-1,0,-1,1,0,0,0,-1,4,0,1,-1,-1,-1,0,-1,3,-1,-1,0,1,0,-1,0,1

Mean of calling: -.22
Standard deviation: 1.6449
P value: .3489

Calling gets euchred: 6/50 = 12%
Calling gets 1 pt: 35/50 = 70%%
Calling gets sweep: 9/50 = 18%
EO of calling: .82
EO of passing: 1.04
EV of calling: -.22

So far passing/bagging is beating out calling, but there's a ways to go before we reach a 95% CI. To be continued. This sample assumes all players play like me, so S1 is always calling something in the 2nd rd veering towards Next if he doesn't block reverse Next. Note, there are scores S1 can't do this like up 9-8, or 8-8. So this sample would not apply in those specific situations. Up 9-8 or 8-8 I would recommend calling with this holding regardless of what my sample will eventually conclude.

Re: ITT We Test 3rd Seat Hands

Posted: Mon Apr 26, 2021 11:42 am
by Tbolt65
I agree with the 9-8, 8-8. Got to call. I would assume you would agree an 8-9 is still a pass?

I also agree, until you find out your partner does not call in seat 1. Its a call.

Tbolt65
Edward

Re: ITT We Test 3rd Seat Hands

Posted: Tue Apr 27, 2021 3:04 am
by Wes (aka the legend)
Tbolt65 wrote:
Mon Apr 26, 2021 11:42 am
I agree with the 9-8, 8-8. Got to call. I would assume you would agree an 8-9 is still a pass?
Yes it's a pass down 8-9 although one can point out that the EO of passing wont be as good at that score becuz S1-R2 4 pt loners are only worth 2 pts at that score (Not true in our tournament obv but true in a normal game where winning is all that matters). Still that wont make a significant enough difference to change the overall EV of passing. S1-R2 4 pt loners just don't happen that often. Even if S1 has that hand, the dealer can still always mess that up by calling in the first round. In my 50 hand sample, there has only been one S1-R2 made 4 pt loner. Ofc this all assumes the current conclusion of my sample holds up. It very well may not.
Tbolt65 wrote:
Mon Apr 26, 2021 11:42 am
I also agree, until you find out your partner does not call in seat 1. Its a call.

Tbolt65
Edward
Yep. Looking at that nice EO of calling so far (+.82). That's a lot. You simply cannot pass that up if you can't fully trust your P in the 2nd rnd.