12-20-21 OE Monday - stopping lones

[Dlan]

Yes, I know this lone attempt was not successful but could it have been played differently?



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

[justme]

3rd seat wins the first trick on a Kc lead by 1st seat. A Jc second trick lead by 3rd seat leads to euchering the maker. No chance the loner attempt in this situation could succeed.

From the point of view as the dealer, I would not have gone alone. With a little help from my partner we might march for 2 points. The game is well in hand. No need to go big, just not going to pass this solid biddable hand to an opponents loner attempt.

[raydog]

I also found it was ever so marginally better to take partner along on this hand. But it's close, so I can't be definitive (just reading the results of my simulation).

Another curious thing I found was that it is better to discard the QC than the 10D. Both when bidding alone and when bidding with partner. But I suspect this may be due to a bias in my program. Clubs is a longer suit than Diamonds (6 cards vs. 5 cards), so the idea is that my opponents are more likely to favor those suits, when they have a choice, and discard their diamonds. Neither the Q nor the 10 seem likely to win a trick, though the Q is a bit better (the A and K may both be buried or with my partner); the overriding factor seems to be what suit(s) the opponents hold on to. Since all the players in my program are simply playing the odds, they will be more likely (all else being equal) to hang on to black suits.

I noticed the greatest advantage to discarding the QC over discarding the 10D when my partner in S2 bid. Since they would only do this if they had a fairly strong hand, they are most likely void in some suit, and are more likely to hold the AC than the AD, since the former is more valuable (more likely to win a trick). They are also more likely to be void in D, the shorter suit. So in that particular instance I think it makes sense. I'm not sure if players go through this analysis when they are dealer in this situation and are deciding what to discard. With Q and lower ranks, the more important decision may be the suit than the rank. Just a thought.

[Wes (aka the legend)]

IMO S1 led right. When given the choice between leading the KD or the KC S1 should lead the KD given that his P is slightly more likely to be void in that suit. As it turns out S3 WAS void in that suit but didn't trump in with an unguarded Left. I'm assuming that's a misclick.

[Wes (aka the legend)]

I also found it was ever so marginally better to take partner along on this hand. But it's close, so I can't be definitive (just reading the results of my simulation).

That's what I would've guessed, that calling would have a higher EV than going alone with this configuration. Keep in mind tho the score matters here. If the dealer's team is at 7 pts, like in this hand, loners are now only worth 3 pts. That changes the math. Did you account for that in your simulation? If not then I think we can probably clearly say that going alone with this holding up 7-1 is incorrect. If your data has calling marginally better with a successful loner worth 4 points, than surely calling will be significantly better if a loner is only worth 3 points. This of course is assuming that winning is all that matters. If one is playing in a tournament where total points matter, then the above doesn't apply.

I noticed the greatest advantage to discarding the QC over discarding the 10D when my partner in S2 bid. Since they would only do this if they had a fairly strong hand, they are most likely void in some suit, and are more likely to hold the AC than the AD, since the former is more valuable (more likely to win a trick). They are also more likely to be void in D, the shorter suit. So in that particular instance I think it makes sense. I'm not sure if players go through this analysis when they are dealer in this situation and are deciding what to discard. With Q and lower ranks, the more important decision may be the suit than the rank. Just a thought.

If S2 ordered I would discard a green King over a Next 9 for the reasons you've talked about above. I believe Wolf disagrees with me over that. However whether that's correct or not with unknown partners is a different argument than whether that's correct or now with known partners. Real life example: When me and Edward play as partners we always discard green when we have the choice assuming that green is a non-ace. This allows our partner to read our hand better and lead to our void with much greater accuracy. For instance, If I call red from the 2 seat 1st rd and Ed shows a club on the first trick, I can confidently lead a spade on the next trick when necessary knowing there's a great chance Ed is void in that suit and thus will trump in. This is important stuff when calling with marginal hands like R+1+0.

[irishwolf]

Several topics here.

Big mistake for S3 not to trump that KD lead with his Left bower. Might have even euchred the runner.

But to me it's a toss up leading the KC vs KD. Two singletons by S1, which is better and Why? The advantage of leading the KC is that it helps your partner to sort his hand. An example, what if the Dealer had he AC + xC. And your partner might have QC + XC and an off suit ace. Leading the KC he can save his doubleton, even sloughing the acde. I have done it before. And in the next suit, Dealer could have a singleton Diamond. Saving the douibleton, even if (S3) holds an ace, it all depends!

True that the advantage of leading KD is it's the shorter suit and 16.7% more likely for S3 to trump it (like he did not do). So its not conclusive, IMO. But as a warning, the dealer having a doubleton when going alone is greater than 16.7% So chose your decision wisely.

But I want to talk about the discard, 10D vs QC. The discard, IMO has nothing to do with one suit is shorter than another. Why? Because there are only two cards that stop the QC and three that stop the 10D. That is a statistical fact and has nothing to do (or should anyway) about tendencies either. So you chose!

However, that is a completely different situation and statistic if S2 assists. Of course if you have a known partner and SET UP A CONVENTION on what you agree to discard. Those are two different species.

Then the 3rd topic about going alone or or with your partner. Of course score matters but in general if you do not go alone, you MUST SCORE TWICE AS MANY SWEEPS to equal successful loners here. With than hand, a reasonable estimate of a successful march is about 30% (1 in 3 or 4 tries). So the question is, would taking your partner along score 60% sweeps? I SAY NO WAY! Which has a higher EV, you guess.

Except at 8 or 9, I would go alone and discard the 10D.

IRISH

[irishwolf]

So let's explore this statement, as one would think this Convention is the Holy Grail of Euchre:

"If S2 ordered I would discard a green King over a Next 9 for the reasons you've talked about above. . . .Real life example: When me and Edward play as partners we always discard green when we have the choice assuming that green is a non-ace. This allows our partner to read our hand better and lead to our void with much greater accuracy."

Situation Dealer has R + 1 (JH up) and dealer has JH has 10H QS 9s, KC 10D picks up JH. Luck would have it AC buried (16.7% it could be). S1 leaDs QC, S2 only has QH and trumps the trick. He as holds JS KS TS KD (or maybe the AD) what should he now lead?

Or what if S2 ordered KH up and the Dealer holds 10S 9S QS KC 9C, no Next to discard. S2 thinks the Dealer has a card in Next and will avoid leading Next?

I am not saying it is not a Convention but those type of agreements have to work at level to be Effective ploys? Go figure! Just know an Ace will be buried about 1 in six hands.

IRISH

[irishwolf]

GOING ALONE vs NOT: (POINTS) Most likely close because you are discarding the QC instead of the 10D. How is the program set as to what is led on a loner? If programmed to favor a next or always a green suit lead, that might be an inherent bias.

With additional thought about this one. The loner rate is much lower than I had initially thought. I say the loner rate will be 10 to15%, of course more euchres going alone but the Sweeps taking your partner along will not out perform going alone.

ALONE: Did you do it both ways (the discarding) on your simulator?

Irish

"I also found it was ever so marginally better to take partner along on this hand. But it's close. . ." WHOA!

[RedDuke]

IMO S1 led right. When given the choice between leading the KD or the KC S1 should lead the KD given that his P is slightly more likely to be void in that suit. As it turns out S3 WAS void in that suit but didn't trump in with an unguarded Left. I'm assuming that's a misclick.

Why in the world would S3 trump in with an unguarded left? He knows that the maker has the right and is also slightly more likely to be void in diamonds than in clubs.

The only possible reason to trump in with an unguarded left here would be if he wanted to force the maker to play the right or lose the trick. That makes no sense when S3 does not know where the ace-trump is.

There's no possible way for this lone to succeed. Even if both S1 and S3 played like idiots, they would have stopped it. The only way for it to have succeeded from the maker's point of view is for Ace-King-Queen of trump to all be buried or with the partner. If either opponent has two trump, the lone will fail. It's very likely that at least one opponent has two trump. It would have been better to take the partner along and try for march.

[irishwolf]

Now RedDuke - "Why in the world would S3 trump in with an unguarded left? He knows that the maker has the right and is also slightly more likely to be void in diamonds than in clubs."

This is one of the ABSOLUTES, must do in Euchre to trump with the unguarded Left. But I will you addressed to Wes so I will let him tune in on that. lol

IRISH

[Wes (aka the legend)]

Why in the world would S3 trump in with an unguarded left? He knows that the maker has the right and is also slightly more likely to be void in diamonds than in clubs.

The only possible reason to trump in with an unguarded left here would be if he wanted to force the maker to play the right or lose the trick. That makes no sense when S3 does not know where the ace-trump is.

S3 has 1 trump--the best possible one trump, the JD--and a chance to use it by trumping in on the first trick. If the maker has the AD or if your P has just the AH or KhXh trumping in will stop the loner everytime. And thats obviously the goal of S3, to do whatever he can to stop this loner. So of course S3 HAS to trump in here. After all holding onto his JD will do nothing for his team as it will be strategically worthless by 2nd street as the maker will surely clean him out. So It's now or never. First street is usually the only chance S3 has of using his trump to help his team, to stop that loner.

Part of me--actually a 100% of me--is shocked I even have to explain this. Of course one should trump in here. What else are they gonna do with their trump?!

Funny thing is almost this exact spot happened to me last night in Round 1 of my weekly tournament. My partner was a beginner. Bad luck for me. Dealer went alone on Hearts. I led the KD, my P was void in diamonds but didn't use her unguarded Left. She just inexplicably played off. The maker had the AD. So trumping in would've saved my team. That mistake cost us the game and knocked me out of that tournament. It also likely cost me the yearly championship as it's going down to the wire. Interesting to note, I've played probably over 30k games online + real life the last 4 years. I have seen every possible mistake--or so I thought--but yesterday was the first time I seen THAT mistake. This is more likely to happen on the app due to misclicks which I would chalk up as shit happens which is categorically different than somebody willfully making this mistake but still I've never even seen that. Everyone who plays this game for 5 minutes just knows to trump in here.

It's utterly surreal to be a victim of this mistake and it's utterly surreal I had to even explain this out to you the day after as if it needed explanation. Not trying to be mean but what are the odds. I wasn't aware there was even one human brain that would come to a different conclusion. In the last 24 hrs I have encountered two.

PS: I suppose it's three people lol if you count the unknown in that video. I'm still betting it was a misclick

[raydog]

Addressing Irish's questions:
I looked at my data again and was initially confused, as it didn't seem to jibe with what I had written previously. So I re-ran the scenarios, and here is a more complete description of my results (with raw data).

S4 bids alone: (100,000 hands; 86,816 bid by S4)
discard QC: 3,724 / 72,421 / 10,671 EV = 0.76 [+4 pts / +1 pt / euchred]
discard 10D: 4,857 / 70,656 / 11,303 EV = 0.78
So I find a higher loner success rate AND a higher euchre rate when discarding the 10D, but ultimately a higher EV when discarding the 10D. Makes sense, as far as the discard goes, but the euchre success rate is only 5.5%, far lower than what Irish predicted. Clearly, a lot of things have to go right to win all the tricks; I haven't attempted a back-of-the-envelope estimate, but am happy if my program says 5.5% - it's not too hard to play a loner hand.

S4 bids with partner: (100,000 hands; 86,638 bid by S4)
discard 10D: 11,940 / 64,834 / 9,864 EV = 0.80
discard QC: 14,194 / 64,577 / 7,867 EV = 0.89
So in this case I find more 5-trick sweeps AND less euchres with the QC discarded. Perhaps because neither the QC nor the 10D is likely to win on its own, so the key is getting void in the longer suit.

Just as a test, I also compared the two "better" strategies above, head to head (so comparing the same subset of hands):
(100,000 hands; 86,767 bid by S4)
S4 bids alone, discards the 10D: 4,774 / 70,795 / 11,198 EV = 0.78
S4 bids w/ partner, discards the QC: 14,129 / 64,742 / 7,896 EV = 0.89
Happy to see results very close to what I found previously; 100,000 appears to be a sufficient sample size.

In this last scenario I also looked at the occasions when S2 bid, since S4 discarded differently:
S2 bids w/ partner, S4 discards 10D: 6,750 / 5,287 / 19 EV = 1.56
S2 bids w/ partner, S4 discards QC: 7,517 / 4,527 / 12 EV = 1.62
So here I again find that discarding the QC gives better results. It is more likely here that S2 has a good hand, and very likely a void. That void is more likely in Diamonds than in Clubs, perhaps explaining the results.

[BTW, if you are trying to add outcomes to 100,000, know that S1 and S3 did bid on a few hundred occasions]

[irishwolf]

Thanks Ray!


IRISH