Page 1 of 1

OE - is this hand winable

Posted: Tue Feb 16, 2021 7:23 pm
by LeftyK
This was the last hand of the night. Can the dealer team keep the game alive?

Image

https://worldofcardgames.com/#!replayer ... %3A1%7D%5D

Re: OE - is this hand winable

Posted: Tue Feb 16, 2021 11:42 pm
by Dlan
This is the exact type of thing that the Euchre Workshop is designed for.

Here you would use the 'Select individual seats' option

You only need to enter the cards once. After the hand is played the first time, the 'Replay the above hand" button will reload the hand as many times as you'd like. This will allow the testing of different ways to approach the hand

Image

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 12:05 am
by irishwolf
No simulation needed. Not winnable if opponents don't get stupid.

An interesting seven games. Many donates, many unwarranted. A few mistakes but not many.

Many attempted loners, only two successful. Guess who?

But some great games, enjoyable!

~Irishwolf

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 3:15 am
by Wes (aka the legend)
LeftyK wrote:
Tue Feb 16, 2021 7:23 pm
This was the last hand of the night. Can the dealer team keep the game alive?
The question you are asking is not relevant. IOW you should not care what the answer to this question is becuz the answer cannot make you a better euchre player. Whatever the answer is will just be result oriented noise. What matters here is what is the best line for the maker in the long run after he takes the first trick? Should he lead offsuit on 2nd street or trump?

According to this site, the maker should lead offsuit from 3 trump, 2 suited, no off ace, configurations: https://ohioeuchre.com/E_ThreeTrumpTwoSuited.php

I don't agree with everything from this site but I do agree with that. Last time I checked, Edward disagrees with the site recommendation and believes the maker SHOULD lead trump on 2nd street in this spot. Unfortunately we can't really KNOW what is best until this spot is put to the test in a good simulation.

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 10:00 am
by Dlan
All questions relating to euchre are encouraged on this forum. The OP asked because he wanted to know. It is up to us, as those providing answers, to be respectful of those that asked, and respectful in our answers.

As far as ‘good’ simulators go, anyone that has played with/against bots knows they have limitations.

In regards to the Euchre Workshop, It is not, nor was it intended to be an intelligent euchre game simulator. It is a way to automate the process of sitting at a table, dealing out hands, and trying to decide the best play.

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 1:13 pm
by Wes (aka the legend)
Dlan wrote:
Wed Feb 17, 2021 10:00 am
All questions relating to euchre are encouraged on this forum. The OP asked because he wanted to know. It is up to us, as those providing answers, to be respectful of those that asked, and respectful in our answers.
I agree. Just to be clear, I wasn't being mean. Knowing the right questions to ask is often more important than knowing the right answer becuz the latter is often unknown or unknowable. Just trying to get OP in the right mindset to be the best player he can possibly be.
Dlan wrote:
Wed Feb 17, 2021 10:00 am
As far as ‘good’ simulators go, anyone that has played with/against bots knows they have limitations.
In theory it should be easy to design a good simulator. Just have one bot team always lead trump in that spot and the other bot team to always lead their offsuit, everything else being equal. Then run as many hands/games as necessary to get a statistically significant result and see which bot team did better. That's one way of doing things.

Or a more efficient way: keep the dealer's hand, I.E. the maker's hand type fixed (3 trump, 2 suited, no off ace), then randomize all other hands (eliminating hand samples where S1,S2 or S3 had a calling hand to keep their passing ranges true) then run it over and over for each strategy (lead trump or lead offsuit) to see which strategy scores more points in a statistically significant sample. That should get us close enough to the truth where we can use the "know" word.

That said, it is often the case that the best evidence we have is riddled with limitations. That's the messy world we live in. All we can do is comport our beliefs with the strength of the evidence. And a good simulator, however limited, would be the strongest evidence we could possibly have here.
Dlan wrote:
Wed Feb 17, 2021 10:00 am
In regards to the Euchre Workshop, It is not, nor was it intended to be an intelligent euchre game simulator. It is a way to automate the process of sitting at a table, dealing out hands, and trying to decide the best play.
I haven't looked at this yet. I'll mess around with it soon.

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 1:51 pm
by Dlan
Wes (aka the legend) wrote:
Wed Feb 17, 2021 1:13 pm
In theory it should be easy to design a good simulator. Just have one bot team always lead trump in that spot and the other bot team to always lead their offsuit, everything else being equal. Then run as many hands/games as necessary to get a statistically significant result and see which bot team did better. That's one way of doing things.
"Just" :)

then program in the other 8000 hands and you have it

Here's a little bit of useless information, The Euchre Workshop used over 1300 lines of code.

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 4:16 pm
by Tbolt65
LeftyK wrote:
Tue Feb 16, 2021 7:23 pm
This was the last hand of the night. Can the dealer team keep the game alive?

Image

https://worldofcardgames.com/#!replayer ... %3A1%7D%5D
I have not read other comments and I am NOT, saying this because Ive already saw the outcome. Wes can back me up because we have had this discussion. I advocate to leading trump and leading High here. Its your best chance at possibly getting two points, or promotes yours and your partners off suit. Plus it yanks trump out of everyones hands. So if your up against a lot of trump in the opponents hands this will help you make them bump heads instead of holding back.

With this configuration you gotta pull trump, asap.


Tbolt65
Edward

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 4:20 pm
by Tbolt65
Tbolt65 wrote:
Wed Feb 17, 2021 4:16 pm
LeftyK wrote:
Tue Feb 16, 2021 7:23 pm
This was the last hand of the night. Can the dealer team keep the game alive?

Image

https://worldofcardgames.com/#!replayer ... %3A1%7D%5D
I have not read other comments and I am NOT, saying this because Ive already saw the outcome. Wes can back me up because we have had this discussion. I advocate to leading trump and leading High here. Its your best chance at possibly getting two points, or promotes yours and your partners off suit. Plus it yanks trump out of everyones hands. So if your up against a lot of trump in the opponents hands this will help you make them bump heads instead of holding back.

With this configuration you gotta pull trump, asap.


Tbolt65
Edward
P.s.

The hand is both still winable and loseable but you have to give yourself the best shot at winning the point, or even two.

Tbolt65
Edward

Re: OE - is this hand winable

Posted: Wed Feb 17, 2021 5:13 pm
by Dlan
Wes (aka the legend) wrote:
Wed Feb 17, 2021 1:13 pm
Or a more efficient way: keep the dealer's hand, I.E. the maker's hand type fixed (3 trump, 2 suited, no off ace), then randomize all other hands (eliminating hand samples where S1,S2 or S3 had a calling hand to keep their passing ranges true) then run it over and over for each strategy (lead trump or lead offsuit) to see which strategy scores more points in a statistically significant sample. That should get us close enough to the truth where we can use the "know" word.
Image

Yes, the workshop will help here as well. Use the "Select dealer + Turn" option. Enter your starting hand plus the turn card. Use "Random deal E-N-W" to deal the rest. If a hand dealt doesn't meet the required criteria, simply hit "Replay w/new E-N-W". You may use that as many times as you'd like. The workshop will remember the dealer's hand and the turn card each time.

Re: OE - is this hand winable

Posted: Thu Feb 18, 2021 2:06 pm
by LeftyK
irishwolf wrote:
Wed Feb 17, 2021 12:05 am
No simulation needed. Not winnable if opponents don't get stupid.

An interesting seven games. Many donates, many unwarranted. A few mistakes but not many.

Many attempted loners, only two successful. Guess who?

But some great games, enjoyable!

~Irishwolf
Yes it was quite fun adjusting to your style of game, although I think I only went 3/8 which is just below my 500 avg with this tough crowd.

Re: OE - is this hand winable

Posted: Thu Feb 18, 2021 4:46 pm
by RedDuke
Wes (aka the legend) wrote:
Wed Feb 17, 2021 1:13 pm
Dlan wrote:
Wed Feb 17, 2021 10:00 am
All questions relating to euchre are encouraged on this forum. The OP asked because he wanted to know. It is up to us, as those providing answers, to be respectful of those that asked, and respectful in our answers.
I agree. Just to be clear, I wasn't being mean. Knowing the right questions to ask is often more important than knowing the right answer becuz the latter is often unknown or unknowable. Just trying to get OP in the right mindset to be the best player he can possibly be.
Dlan wrote:
Wed Feb 17, 2021 10:00 am
As far as ‘good’ simulators go, anyone that has played with/against bots knows they have limitations.
In theory it should be easy to design a good simulator. Just have one bot team always lead trump in that spot and the other bot team to always lead their offsuit, everything else being equal. Then run as many hands/games as necessary to get a statistically significant result and see which bot team did better. That's one way of doing things.

Or a more efficient way: keep the dealer's hand, I.E. the maker's hand type fixed (3 trump, 2 suited, no off ace), then randomize all other hands (eliminating hand samples where S1,S2 or S3 had a calling hand to keep their passing ranges true) then run it over and over for each strategy (lead trump or lead offsuit) to see which strategy scores more points in a statistically significant sample. That should get us close enough to the truth where we can use the "know" word.

That said, it is often the case that the best evidence we have is riddled with limitations. That's the messy world we live in. All we can do is comport our beliefs with the strength of the evidence. And a good simulator, however limited, would be the strongest evidence we could possibly have here.
Dlan wrote:
Wed Feb 17, 2021 10:00 am
In regards to the Euchre Workshop, It is not, nor was it intended to be an intelligent euchre game simulator. It is a way to automate the process of sitting at a table, dealing out hands, and trying to decide the best play.
I haven't looked at this yet. I'll mess around with it soon.
Wes,

I actually am a Python programmer. I've got a Ph.D. in statistics and data science.

I wasn't able to come up with any way to code in every possibility without crashing my 32 gb RAM 16-core rig. So good luck finding anyone that can code a bot that can play out all these hands perfectly. The technology is just not there yet.

With that said, I don't think there was any realistic way for the dealer to win this. Picking up was suicide since the opponents had the big trump. Passing was also pretty dumb since the dealer had no good second round call. Overall, this was pretty much a Catch-22 for the dealer. I think the dealer made the right call here, but the card layout pretty much promised that he would lose.