[kfay]

Here's a play problem I'm absolutely ashamed to say I got wrong.

It came up in the World Junior Teams during my team's (USA) semi-final match against Israel. Up to this point in the tournament I'd played about as well as I would expect of myself, and then this hand came along and reminded me why I suck at bridge.

CONSTANT VIGILANCE!!!!!

---

How do you play on the lead of the ♦2 (3/5)?

[inquiry]

hide my answer since the hand was meant for B/I forum (see kfay comment below).

[hidden]Percentage play in hearts for one loser (start ♥Q), then after trumps are pulled, protect against 4-0 club split (play ♣A first). [/hidden]

This post has been edited by inquiry: 2010-October-27, 13:00

[kfay]

Oops, I meant to post this in B/I. Whatever, though.

[wyman]

what's viligance?

[kfay]

The result of typing too fast.

[Bbradley62]

Unhiding, since others are unhidden... Thanks for the insight, Inquiry...

With 9 including all the spots, isn't cashing ♥A equal to leading ♥Q? Leading Q loses to KJx(x) in West; and cashing A loses to KJx(x) in East. Leading Q wins if singleton J in West; cashing A wins if singleton K in East. Equal for other positions. For some reason, I want to start with A. But definitely start ♣ with A.

[mfa1010]

Best heart play is ace and another. Running the Q wins in symmetrical cases except that it has an extra losing case when K is stiff behind the ace.

An alternative line in trumps is low to the T and then run the Q. This is marginally worse in theory, but in practice LHO sometimes rises K from Kx.

If we claim that it is just a mistake to rise K from Kx, then we should remember this when we have to play the suit for no losers. Say same hand with ♣Kxx in the closed hand. It is twice as likely that LHO has ♥Kx than RHO has ♥K, so maybe the best line is to go for a swindle and play small heart T2 to the Q instead of banging down the ace or running the Q.

[gnasher]

Another possibility is ♥A, ♠A, spade ruff, diamonds throwing a club, trump. That guards against RHO's holding ♥Kx and a club void. It loses when they've led a singleton diamond, or a deceptive 6th or 7th best.

BTW, I think this deal is in the right forum.

[gnasher]

gnasher, on 2010-October-27, 13:59, said:

Another possibility is ♥A, ♠A, spade ruff, diamonds throwing a club, trump. That guards against RHO's holding ♥Kx and a club void. It loses when they've led a singleton diamond, or a deceptive 6th or 7th best.

Regarding the possibility that LHO has a singleton diamond, I don't think this is much of a risk. If he had ♥Kxx or ♥KJx, he probably wouldn't lead his singleton. If he's 1=4 in the minors I'm down anyway. Hence I'm only losing to something like 7213 without ♥K.

I think that makes this line rather better than just drawing trumps.

[pooltuna]

There are two suits you can misplay one mostly just bad luck the other just bad play. There are 2 ways to play the ♥ suit but the difference between is negligible but for ease of hand flow I would lead the A♥ and another rather than try to double finesse the KJ. The ♣ suit is another matter you can handle 4-0 with rho having 4 but are screwed if LHO has 4 but you have to lead to the A first. Leading the K or Q is just careless. On a good day you will catch RHO with the singleton ♥K and LHO with 4♣

[phil_20686]

mfa1010, on 2010-October-27, 13:51, said:

Best heart play is ace and another. Running the Q wins in symmetrical cases except that it has an extra losing case when K is stiff behind the ace.

This isn't true. If you run the queen and if it loses to the K behind run the ten. Wins vs all split honours, loses only to KJ tight and KJx(x) offside.

So the losing case is KJ tight, they are the same in all other cases. This is the only case where ace and another is better. Running the Q is correct in all cases holding holding 8 or fewer cards including:

Axxx QT98
AT98 Qxxx

Axx Qt98
AQx T98x (finesse q first for one loser)
Ax QT98x(x) (often forgotten)

QT9 Axx
AQx T98

AT9 Qx


Returning from our tangent: The lines differ by only one case out of twelve of 2-2 breaks is about 3% or so, and if you have any inference the rho might be longer in hearts that would be enough to sway it. Here if you judged that the 2 was likely from 5 as he might lead into the strong nt hand from his longest collection of small cards, then you might run the Q.

[mfa1010]

phil_20686, on 2010-October-29, 08:58, said:

This isn't true. If you run the queen and if it loses to the K behind run the ten. Wins vs all split honours, loses only to KJ tight and KJx(x) offside.

So the losing case is KJ tight, they are the same in all other cases. This is the only case where ace and another is better
...

Well, I disagree. If we run the Q to the K, then we should play to drop the J next time. We are analyzing a 9-card fit. K to Jxx is one specific division, and so is KJ to xx. The specific 2-2 is more likely than the specific 1-3.

[WellSpyder]

mfa1010, on 2010-October-29, 09:07, said:

Well, I disagree. If we run the Q to the K, then we should play to drop the J next time. We are analyzing a 9-card fit. K to Jxx is one specific division, and so is KJ to xx. The specific 2-2 is more likely than the specific 1-3.

Agreed.

Indeed, if you are going to finesse twice, I think it is better to run the 10 the first time round, since you can then change your mind if RHO shows out. Finessing twice starting with the 10 loses to KJ offside or KJx offside, but not KJxx offside. Starting with A loses to KJx onside and KJxx onside, but not KJ onside. The KJxs balance, but the specific 2-2 is more likely than the specific 4-0, so it looks like cashing A is better.

But putting LHO to the test with Kx by leading small from hand may be better still....

[phil_20686]

mfa1010, on 2010-October-29, 09:07, said:

Well, I disagree. If we run the Q to the K, then we should play to drop the J next time. We are analyzing a 9-card fit. K to Jxx is one specific division, and so is KJ to xx. The specific 2-2 is more likely than the specific 1-3.

this is not true. there are 8 3-1 breaks comprising 50% = 6.25% each, there are 12 2-2 breaks comprising 40% = 3.4% each, so one 3-1 break is greater than one 2-2 break. or "specific doubleton is less likely than a singleton" as some have learned in another context.

[mfa1010]

phil_20686, on 2010-October-29, 10:00, said:

this is not true. there are 8 3-1 breaks comprising 50% = 6.25% each, there are 12 2-2 breaks comprising 40% = 3.4% each, so one 3-1 break is greater than one 2-2 break. or "specific doubleton is less likely than a singleton" as some have learned in another context.

I disagree again.
There are 8 3-1/1-3 breaks but only 6 2-2 breaks.

this is not true

That expression is starting to annoy me a little...

[kfay]

phil_20686, on 2010-October-29, 10:00, said:

this is not true. there are 8 3-1 breaks comprising 50% = 6.25% each, there are 12 2-2 breaks comprising 40% = 3.4% each, so one 3-1 break is greater than one 2-2 break. or "specific doubleton is less likely than a singleton" as some have learned in another context.

It's true that, when discussing suit breaks, any 1 evenly-divided configuration is more likely than any 1 other more-unevenly-divided configuration.

[phil_20686]

mfa1010, on 2010-October-29, 10:21, said:

I disagree again.
There are 8 3-1/1-3 breaks but only 6 2-2 breaks.

this is not true

That expression is starting to annoy me a little...

Oh, that is true actually. just thought "Choose one form 4 and one from 3": forgot have to divide by 2(2!) to avoid the double count of permutations. Now its absurdly close isnt it: 6.6% vs 6.25% or something. At any rate - both those lines are inferior to A and another by 6.x%

[mfa1010]

phil_20686, on 2010-October-29, 10:52, said:

Oh, that is true actually. just thought "Choose one form 4 and one from 3": forgot have to divide by 2(2!) to avoid the double count of permutations. Now its absurdly close isnt it: 6.6% vs 6.25% or something. At any rate - both those lines are inferior to A and another by 6.x%

This time I agree.

[Cascade]

phil_20686, on 2010-October-29, 08:58, said:

if you have any inference the rho might be longer in hearts that would be enough to sway it.

Or that RHO was more likely to have the strength.

[l milne]

I think we can make two simplifications here:

1. we can't draw any inference from the lead, as it could be 3-5 or 5-3 in ♦ either way;
2.I'm not going to worry about pressure plays on LHO with Kx ♥, as that will also put unnecessary pressure on myself when there is a good line available.

There are two lines in trumps:
A: Ace first then lead towards the QT, losing only to KJx and KJxx with RHO;
B: Run the queen. If this loses to the King, play the Ace, losing to KJx and KJxx with LHO, as well as K opposite Jxx.

A is better, not only because it picks up all the K opposite Jxx combinations, but also because having to guess on the second round sucks. Even though playing for the drop on the 2nd round in line B is best, it is very close. Beating myself up on later hands about these isn't fun.

One final consideration: if LHO has JTxx ♣, then we must lose a trick in clubs (roughly 5%). In this case, we would need to play trumps for no loser, in which case lines A and B are equal. But we would take line B, as vacant spaces suggests J opposite Kxx is more likely than Jxx opposite K. However, given that 4-0 clubs with LHO is only ~5% and K opposite Jxx (the extra losing combo for line B ) is over 6%, line A in trumps is best.

Obviously, we play the clubs the correct way as well.