[lamford]

An interesting suit combination occurred yesterday night at the Young Chelsea Bridge Club, which was 987654 opposite Q32 with plenty of entries to each hand, at IMPs. You lead from the 9 and the 10 or J appears. You can afford two losers but not three. Over to you.

[WellSpyder]

 lamford, on 2014-March-08, 05:35, said:

An interesting suit combination occurred yesterday night at the Young Chelsea Bridge Club, which was 987654 opposite Q32 with plenty of entries to each hand, at IMPs. You lead from the 9 and the 10 or J appears. You can afford two losers but not three. Over to you.

The lack of a response up to now may suggest the novice/beginner forum isn't necessarily the best place for suit combinations. I suspect novices and beginners are reluctant to post either because they expect they will get it wrong or simply because they don't think these sort of problems are particularly relevant to the game at their level, while more advanced players don't want to spoil the problem for beginners. Anyway, here is my attempt at an answer:

Spoiler

You are missing A and K (call them H for high honour) and J and 10 (call them L for low honour). The relevant holdings for LHO are HHL, when playing Q wins, and HLL when ducking in dummy wins. There are equal numbers of these holdings, so that won't help you make a decision. Similarly, I'm not sure the principle of restricted choice helps. Certainly LHO has a choice of low honours from HLL, but they also have the choice of playing H and then L from HHL.

So I would make my decision on the basis that it is probably clearer to me than to LHO that I can afford 2 losers but not 3, so in practice they are likely to play H from HHL. I therefore play them for HLL and duck in dummy, being prepared to congratulate them if they have found the play of L from HHL....

[1eyedjack]

I would probably go up with the Q, on the grounds that if I cannot avoid three losers I go one fewer down than if I sustain 4 losers, which would be the case with AKJT onside if I ducked, and all other scenarios being equal.

[SteveMoe]

Hmmmm...Suit Play suggests that low to the Q and low ducking offer exactly the same success percentage - 53.1304% for 4 tricks (i. e. two losers) Looks like extra information about distribution or Hand strength (any bidding?) would help tilt the decision one way or the other. We need more detail when decisions are this close. If we have no more information from the bidding or play we're left with a guess.

[1eyedjack]

 SteveMoe, on 2014-March-11, 22:39, said:

Hmmmm...Suit Play suggests that low to the Q and low ducking offer exactly the same success percentage - 53.1304% for 4 tricks (i. e. two losers) Looks like extra information about distribution or Hand strength (any bidding?) would help tilt the decision one way or the other. We need more detail when decisions are this close. If we have no more information from the bidding or play we're left with a guess.

I suspect that what he is getting as is that while the numbers are equal against perfect oppos there is in reality a psychological increased chance that LHO might rise with a high honour in front of Q where suitplay suggests low. But that does not seem like a N/B issue.

It could be that the entire point of the thread is that it may be counterintuitive that it makes no long term difference what you choose to do on the first round.

[billw55]

I don't think this one is all that complicated.

Spoiler

It is pretty easy to list all possible trump divisions. Let's say LHO plays the jack:

1. J - AKT
2. JT - AK
3. KJ - AT
4. AJ - KT
5. AKJ - T
6. AJT - K
7. KJT - A
8. AKJT - void

In cases 1-4, the play from dummy is irrelevant. In case 5, playing the queen gains. In cases 6 and 7, ducking gains. In case 8, playing the queen gains, but three tricks are still lost so this is irrelevant. So it looks like ducking is better, even before considering any psychological aspects. Although it is hard to believe SuitPlay is wrong ... am I missing something obvious?

[johnu]

 billw55, on 2014-March-13, 07:08, said:

I don't think this one is all that complicated.

Spoiler

It is pretty easy to list all possible trump divisions. Let's say LHO plays the jack:

1. J - AKT
2. JT - AK
3. KJ - AT
4. AJ - KT
5. AKJ - T
6. AJT - K
7. KJT - A
8. AKJT - void

In cases 1-4, the play from dummy is irrelevant. In case 5, playing the queen gains. In cases 6 and 7, ducking gains. In case 8, playing the queen gains, but three tricks are still lost so this is irrelevant. So it looks like ducking is better, even before considering any psychological aspects. Although it is hard to believe SuitPlay is wrong ... am I missing something obvious?

You are missing the AK10 - J case.

[Cyberyeti]

 johnu, on 2014-March-13, 12:19, said:

You are missing the AK10 - J case.

Difficult for the J to be played from AK10

[johnu]

 Cyberyeti, on 2014-March-13, 12:50, said:

Difficult for the J to be played from AK10

The basic question was whether the SuitPlay a priori odds of 53.1304% given in a prior post for both lines of play is correct.

Once the J or 10 are played, a type of restricted choice comes into play. With AKJ or AK10, you have to play J or 10 (if you don't play high), with XJ10, you can play the J or 10. I assume based on the given percentages that SuitPlay assumes random play among equal rank cards, otherwise the percentages would have been different.

Thus you have to look at the case where the 10 is played first to get the whole picture.

[1eyedjack]

 Cyberyeti, on 2014-March-13, 12:50, said:

Difficult for the J to be played from AK10

Would you play differently from dummy had the J appeared instead of the 10, then?

[kenrexford]

This reminds me of one of my favorite suit combinations from actual play.

I had the functional equivalent of the exact same layout but only five cards in Dummy. I also played the functional equivalent of the 9 from Dummy and saw the 10 or Jack from RHO. I tried the Queen, losing as expected, to the King.

After 9-J/10-Q-K, I needed to lose only one more trick in the suit. This of course seems impossible. However, there is a line, sort of.

What I did was to fake a finesse by crossing to Dummy to lead small toward AKQ in my hand in a side suit, finessing the Queen to make LHO think RHO had the King.

I also cleared LHO's suit for him (notrump contract) so that if he got in he would have cashers against me.

I then tried a small card toward Dummy's 8.

LHO was caught in a problem. If I was missing the King in that side suit, then I must have the Ace in this suit for my points. Hence, I must be trying to sneak past his J-x. He hopped the Jack, smothered by his partner's now tight Ace, and the two losers compressed into one, with RHO helplessly unable to lead LHO's suit for him.

The actual layout you provide does not have this extra chance. However, 98xxx opposite Qxx is not a layout where you simply do the math. You also have to add in the possibility of catching KJx or K10x to the left, having RHO play small from AJ or A10, and being able to induce a hop of the Jack at round two by LHO.

[lamford]

Suitplay is indeed correct that both queen and low are the same at around 53%. I would always play low however, as it is only identical against god, who plays low from AKT and AKJ 100% of the time. If he varies one iota from that strategy, then you gain by ducking. Some average players would see no point in playing the jack from AKJ, so I expect quite a big gain from ducking. There was no opposition bidding to help you, and the relevant distributions are all 3-1 onside, so no clue from the carding in any other suit.

[Zelandakh]

My take on this is that (since we are in N/B) the Q is best if the jack appears and ducking is best if the ten appears. The reason being that a pure novice will play the ten from HJT almost every time and will play a high one from AKx in some non-zero percentage. Against a "tricky" player (one who always plays the J from HJT) or one who has learned to vary equal honours we can instead always duck.

[helene_t]

The "math" is quite easy and really shouldn't require the use of SuitPlay - just realize that it only matters if RHO has a singleton, and obviously a priori the chance that the singleton is A or K is 50%. So it boils down to whether LHO would play high from AK(J/T), whether he would play high from (A/K)JT, and which card he is more likely to play if holding both J and T and deciding not to play high.

I agree with WellSpyder that this kind of problems shouldn't be posted in the N/B forum. It might be too trivial for the Expert forum but I/A would be appropriate.

[JLOGIC]

 helene_t, on 2014-March-14, 03:30, said:

The "math" is quite easy and really shouldn't require the use of SuitPlay - just realize that it only matters if RHO has a singleton, and obviously a priori the chance that the singleton is A or K is 50%. So it boils down to whether LHO would play high from AK(J/T), whether he would play high from (A/K)JT, and which card he is more likely to play if holding both J and T and deciding not to play high.

Agreed, people confuse themselves when they think "well he played the jack not the ten"... It even happened in this thread. Of course you can also look at it this way, in which case it is a case of restricted choice, after the J is played it is 50 % of KJT, and 50 % of AJT, vs 100 % of AKJ which is also 50/50 but is a more convoluted way of doing it. The way you described is simple and clear, does not require suitplay or advanced math knowledge, and automatically does not fall for restricted choice traps. Low to the queen wins against stiff J or stiff T on your right, and loses to stiff A or stiff K, ez game.

[billw55]

 JLOGIC, on 2014-March-14, 17:43, said:

Agreed, people confuse themselves when they think "well he played the jack not the ten"... It even happened in this thread.

Guilty