[gnasher]

JLOGIC, on Sep 9 2010, 08:13 PM, said:

Why do you think a spade is best if partner has not signalled (or do you think that)?

I suppose it depends on how he has not signalled.

If the 9 is his lowest, he has A109 or K109. If it's A109, as you said it doesn't matter what we do. If it's K109 and declarer has Axxx Kxx QJx Qxx, a diamond lets it through, because we get strip-squeezed.

If the 9 is random, so that it might be a non-signal from 109x, we're back to a guess, I suppose. Even then, it's not clear that a diamond is better than a spade. A diamond is playing for declarer to have ♣Qxx or better and ♦J9x or worse. A spade is playing for declarer to have ♣Jxx and ♦J10x, which is less likely.

However, if partner is playing his cards at random, he's only playing the 9 from 1/3 of the 109x holdings, and I still get half of the time that he has K109. That probably doesn't swing it far enough in my direction, but I can't be bothered to do any arithmetic.

This post has been edited by gnasher: 2010-September-09, 13:34

[eyhung]

Fascinating discussion here on signalling, I'm glad I posted the hand. FWIW I do believe suit preference is not standard below the top-level, but if I were playing with a top partner I expect it to be standard. Just for completeness, on the actual hand partner had QJx of diamonds and Qxx of clubs (declarer had J9x of clubs), so a diamond or the club king (to induce a misguess) would have been fine.

[gnasher]

I didn't mention K98, where again a spade return is best.

PS: Sorry, Eugene, that we spent so much time talking about the incorrect deal and not about the one you actually wanted to post.

[awm]

I'm trying to work out why a diamond is better than a club (ignoring any suit preference implications from partner's carding).

It seems like if declarer has the queen of the minor you switch to, it's potentially bad (giving up a tempo). If partner has the queen and jack, then you made a great switch. If partner has the queen and declarer has the jack, then you did well when partner has the next lower card and badly when declarer does (i.e. if declarer has ♦JTx or ♣J9x then switching to that minor costs, otherwise it's good).

Overall it seems very symmetric to me, so I'm not sure why several people seem to be convinced that a diamond is much better.

[eyhung]

No problem, I obviously was more culpable in not getting my spots/plays right when I posted the hand.

[JLOGIC]

Gnasher, why would partner not give honest carding with AT9 or KT9? The need to falsecard is no longer there.

Anyways, enough about the carding.

Quote

Overall it seems very symmetric to me, so I'm not sure why several people seem to be convinced that a diamond is much better.

This is a good question. Here is my answer:

If declarer has a minor suit queen, we have to shift to the other suit obviously. So I think a diamond is better because:

1) If declarer has Qx of a minor and is open in the other minor, I am dead if I shift to a club, but he might have QT doubleton of diamonds and misguess (meaning partner led from AQxx instead of J9xx which I think is reasoanble), or he might even have QT8 of diamonds and end up misguessing twice. If I play a club and he has Qx or Qxx he cannot misguess

2) If declarer has QJ of a minor, he would almost definitely have gone after diamonds first with QJT, but since he can't have the ten of clubs he might have chosen to go after spades first with QJx of clubs.

3) It's possible declarer with JT8 of diamonds will misguess later even though not likely. Returning a high diamond might be a good play for this reason.

Minor things but imo they make D>C.

Edit to add:

I forgot, if declarer has NO queen then he might well have gone after diamonds first with JT9 of diamonds (the correct play). That eliminates a combination where playing a diamond back is worse than a club.

[gnasher]

JLOGIC, on Sep 9 2010, 11:17 PM, said:

Gnasher, why would partner not give honest carding with AT9 or KT9? The need to falsecard is no longer there.

He can't afford to consistently play the 10 from A109, because declarer might have KQ8x.

With K109, it's still not a good idea to tell declarer which suits are breaking.

[JLOGIC]

gnasher, on Sep 9 2010, 05:48 PM, said:

He can't afford to consistently play the 10 from A109, because declarer might have KQ8x.

So? Declarer will still play him for AT if he plays the ten, and A9 if he plays the 9. What's the problem?

Yes if declarer knows he plays the 10 100 % from AT it is a very small problem, in that case he should play it 90 % or w/e, in reality it does not matter if you play it 100 % of the time.

[bluecalm]

Quote

Partner's ♠9 is interesting. I'd expect suit-preference here - count is far more likely to help declarer than the defence. I also don't think he'd play the 9 unless he had either the 10 or the 8 to go with it, or he had no choice.

+1 for suit preference. Many people here had given up smith in favor of suit preference.

Quote

it was that I do not think you can assume partner is giving suit preference with no agreements to that effect.

Interesting. Playing with advanced+ partner without any agreements other than "no smith" I would assume this is suit preference. I guess we live in different bridge cultures.

Quote

Fascinating discussion here on signalling, I'm glad I posted the hand. FWIW I do believe suit preference is not standard below the top-level, but if I were playing with a top partner I expect it to be standard.

I think top level is overstatement. I usually play with bright people but by no means some top level guys and all of them play either smith or s/p or some combinations of those (smith if honor possible s/p otherwise etc.). Nobody dreams of giving count in declarer's suit when there is so much useful information to give.

Anyway, gotta think about the hand assuming ♠ spot is irrelevant.