TRUMPING YOUR PARTNER's ACE

Ask questions, discuss and debate your strategies, euchre polls and more
Post Reply
irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

TRUMPING YOUR PARTNER's ACE

Unread post by irishwolf » Fri Jul 29, 2022 10:44 am

The situation is this, QH is the upcard and the Dealer makes Hearts trump. You have no idea what those five cards were skill level. However, Eldest leads the AS. You at S3 hold KH QD AC KC 9C.

The question is should you trump that ace?

1) Will it lead to a higher EV trumping or sloughing you QH?

2) Will you not provide an advantage in spoiling more 'marches' by trumping?

3) Will you not provide more euchres by sloughing the QD?

Ray, if you do this on your simulator, when you successfully trump, you MUST lead the AC to the 2nd trick!

NOTE: You have no idea that S1 had AS singleton, doubleton or trip. And that for sure makes a difference. So it 3 simulations. But S2 must follow suit and have a Spade in this situation!

IRISH



Tbolt65
Posts: 820
Joined: Sun Jun 16, 2019 9:14 pm
Location: Las Vegas

Unread post by Tbolt65 » Fri Jul 29, 2022 3:51 pm

A time and place for everything.


With this configuration. You can slough the diamond or you can trump in with the king of hearts.


I will be sloughing more often than not in this situation with the specific up card. With this non-specific score. I have trumped in with the king before and treat it like a bare left. I'm more inclined to trump in here if the Right was picked up, just to force it out. However one can make an argument to make sure S4 drops something higher than a queen of trump on it. But with creating a void in diamonds and having void in spades. I think sloughing here offers more of a chance for a euchre. Now with that said. If the score was say like 8-5 them. I won't gamble and I will try to get my stopper asap. If it was 6-3 them, I do gamble and go for the euchre and slough.


Tbolt65
Edward

irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

Unread post by irishwolf » Fri Jul 29, 2022 5:17 pm

Good comments and I think score and situation is important here. Score - do you need to hold them to one point or do you need a euchre. And probably critical is who is that player at the Dealer spot (aggressive or conservative). Partner tendencies as well.
As they say - IT ALL DEPENDS! (it was Gerry Blue, ES favorite)

This is why I would like to see what Ray's simulator with various hands of the dealer as well as that S1 AD holdings!

IRISH

raydog
Posts: 260
Joined: Thu Sep 16, 2021 6:56 pm

Unread post by raydog » Wed Aug 03, 2022 6:36 pm

To simulate this, I gave S3 the given hand and imposed QH as the turn card. I also gave S1 the AS but left the rest of the cards random. I then only looked at hands where 1) S4 called H with partner; 2) S1 led the AS; and 3) S2 followed suit. I think this is the most realistic way to isolate the subset of possible hands dealt when S3 has to make this particular decision.

Of 100,000 hands dealt, just over 19,000 met the above criteria.
QD played: EV = -0.067 (for S1/S3); EV = -0.108 the second time I ran it
KH played: EV = -0.471; EV = -0.480 the second time I ran it

So quite clearly more favorable to slough the QD.

S1 could have 1, 2 or 3 spades (or even 4), so I tried isolating those different scenarios, even though the results aren't actionable by S3 (who has no idea how many S his partner has).
I ran 100,000 hands twice.
S1 has 1 spade:
QD played: EV = +0.127 (5,760 games); EV = =0.157 (5,627 games)
KH played: EV = -0.343; EV = -0.333

S1 has 2 spades:
QD played: EV = -0.234 (11,152 games); EV = -0.213 (11,191 games)
KH played: EV = -0.643; EV = -0.611

S1 has 3 spades:
QD played: EV = +0.146 (1,598 games); EV = +0.250 (1,578 games)
KH played: EV = +0.043; EV = +0.130

In every case, it was better to play the QD than the KH. The last case is interesting, as the EV's actually turn positive again, but there are two problems with this data. First, the sample size is very small, so there is a lot of volatility. Second, and more important, it is possible that S1 will NOT lead the AS if they have 3 of them, and instead will lead some other suit, which screws up the comparison.

The obvious question is whether my program is playing the hands optimally. So I ran a few small batches to look for specific cases where leading the QD gave better results than leading the KH. I list two of those hands below, along with how my program played them. They look good to me, but let me know if you think they would have been played differently (for different results) by expert players.

[note that I didn't come across any cases where playing the KH was better, but my sample size for this exercise was very small. There almost certainly exists such a hand, but far more hands, it seems, where the opposite is true.]

Hand 1:
S1: AS + AH + A-9D + 10C
S2: K-Q-9S + Q-JC
S3: (see above)
S4: J-10S + 9H + J-10D (10D discarded for QH)

play: AS/9S/QD/10S; S1 wins
AD/JC/9C/9H; S4 wins
JS/10C/KS/KH; S3 wins
AC/QH/AH/QC; S1 wins
9D/QS/KC/JD; S4 wins
euchre!

play: AS/9S/KH/10S; S3 wins
AC/9H/10C/JC; S4 wins
JS/9D/KS/9C; S2 wins
QS/QD/QH/AH; S1 wins;
AD/QC/KC/JD; S4 wins
1 pt.

Hand 2:
S1: A-Q-9S + J-10C
S2: K-JS + 9H + K-10D
S3: (see above)
S4: 10S + A-10H + 9D + QC (10S discarded for QH)

play: AS/JS/QD/10H; S4 wins
9D/9S/KD/KH; S3 wins
AC/QC/10C/9H; S2 wins
KS/9C/QH/QS; S4 wins
AH/JC/10D/KC; S4 wins
1 pt.

play: AS/JS/KH/AH; S4 wins
9D/9S/KD/QD; S2 wins
KS/9C/QC/QS; S2 wins
10D/KC/10H/10C; S4 wins
QH/JC/9H/AC; S4 wins
2 pts. (sweep)

Finally, I ran the simulation a final time for 1,000,000 hands, and also recorded the sweeps/euchres:
188,830 games counted
QD: EV = -0.085 (9,446 / 118,642 / 60,742) [2 pts. / 1 pt. / euchre]
KH: EV = -0.483 (11,049 / 141,553 / 36,228)

So no matter what the score, you will stop more sweeps and get more euchres by playing the QD. IF my simulation is correct...
Last edited by raydog on Thu Aug 04, 2022 8:42 am, edited 1 time in total.

irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

Unread post by irishwolf » Wed Aug 03, 2022 8:29 pm

THANKS for running this - good stuff I need time to digest. Lot of stuff here.

I like Ed's comments as well.

-----------

After review of your results my comment is that I was surprised in that I predicted trump the AS with the KH would result in fewer Sweeps by the dealer but sloughing the QD would result in more Euchres. So only half right there. I will adjust my play as that hand surfaces. Actually, should be quite frequent regardless if S2 has a spade or not. I have no information that would refute your results, good stuff.

Just one thing comes to mind about your program. If the AS wins the first trick and S1 had both a 2nd Spade and a diamond, what does it lead to the 2nd trick? If that were me and I had the choice I would definitely lead a Diamond.

IRISH
Last edited by irishwolf on Thu Aug 04, 2022 9:58 am, edited 1 time in total.

irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

Unread post by irishwolf » Wed Aug 03, 2022 8:53 pm

RA

Fix this to S3 SLOUGHS THE QD: So quite clearly more favorable to slough the QH. I think there a couple of those - S3 can't play QH.

So much worst to trump the AS lead based on your EVs. So clearly that has to imply that the Dealer has a spade or even has 2 spades that presents him with issues.

A little surprised as I think this says you stop more sweeps by Sloughing.

IRISH

irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

Unread post by irishwolf » Fri Aug 05, 2022 12:30 pm

RAY

Still pondering you results. Now wondering the results if S3 had QD 10D KH AC QC with S1 leading the AS was led. Probably similar.

Most of the good Euchre players I know if they had a singleton AH or Left bower unprotected (any other holdings) would trump. Your simulator, I suspect, would say that is an incorrect play. To learn one has to also look other possibilities. I strongly suspect if S1 is leading from Ace headed by three trumping is better.
Problem is you as a convention, S1 should not do that UNLESS he had at least two trump going for a euchre you want your partner to trump high. And if I were S1 with AS KS XS, I am going lead the KS to get my partner to trump in HIGH. Your simulator, I suspect would say that is an incorrect play.

IRISH

raydog
Posts: 260
Joined: Thu Sep 16, 2021 6:56 pm

Unread post by raydog » Mon Aug 08, 2022 2:03 pm

I checked out the revised hand - add 10D (now 2 D, so not voiding in D; A-QC instead of A-K-9C): [1,000,000 hands]

play QD: EV = -0.198 (6,837 / 118,254 / 48,750)
play KH: EV = -0.389 (6,781 / 129,369 / 37,691)

So still clearly better to slough a D, BUT slightly fewer sweeps for S2/S4 is trump in with the KH - though such a slight difference that the significance is questionable).

I then tried a simulation with S1 holding A-K-9 S. The problem here is that S1 won't want to lead that AS, as it is such an unlikely winner (only 3 other S out there). Will only lead the AS if:
1) has 2 trump
2) has JD (L) plus another card
3) has 4 spades + trump

I did indeed run this scenario (well, at least w/ S1 having 3 spades), but ran it again, specifying that the spades were A-K-9. [1,000,000 hands simulated; only 7.5% met the criteria]:

QD: EV = -0.420 (7,208 / 51,296 / 16,988)
KH: EV = -0.542 (8,044 / 53,234 / 14,214)

So still better to slough than to trump.

By the way, if S1 leads the KS (while holding the AS), my program will have S3 trump in. BUT, they will play their lowest trump, if they hold more than one trump. I totally understand the concept of playing highest trump - especially if there is an understanding between partners - but that is a step too far with regards to this particular topic. Happy to pursue in another thread, is you propose a suitable hand.

This is really getting into the weeds, though I agree that they are important weeds, for the hard-core. But perhaps not something everyone wants to follow closely. And I am unsure whether my program plays these types of hands optimally, but would be happy to make the necessary modifications.

I propose that I send you 20 randomly generated hands, from the initial scenario, to have you play as you think best. We can then compare the results with how my program currently plays. This will both allow me to adapt my program to avoid clear sub-optimal play, and also allow you to gain insight into what the opponents are likely to hold, in a given situation.

Let me know if you are game, and I'll send you a personal email through the site.

irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

Unread post by irishwolf » Mon Aug 08, 2022 6:43 pm

Ray, Yes on this "This is really getting into the weeds, though I agree that they are important weeds, for the hard-core. But perhaps not something everyone wants to follow closely. And I am unsure whether my program plays these types of hands optimally, but would be happy to make the necessary modifications.

I propose that I send you 20 randomly generated hands, from the initial scenario, to have you play as you think best. We can then compare the results with how my program currently plays. This will both allow me to adapt my program to avoid clear sub-optimal play, and also allow you to gain insight into what the opponents are likely to hold, in a given situation.

Let me know if you are game," YES

IRISH

raydog
Posts: 260
Joined: Thu Sep 16, 2021 6:56 pm

Unread post by raydog » Thu Aug 11, 2022 2:59 pm

Irish, I sent you a private message on this site, please have a look.

Ray

irishwolf
Posts: 1319
Joined: Tue Apr 24, 2018 9:33 pm

Unread post by irishwolf » Thu Aug 11, 2022 4:13 pm

Okay, will do.

Also, check yours.

Post Reply