[lexlogan]

A recent thread discussed whether to respond 2D or 2C over 1S holding:
♠A ♥Qx ♦Kxxxx ♣AKQxx .

One poster said a 2♦ response was necessary in order to reach 6♦ opposite Axx in partner’s hand; I replied that it might be hard to avoid 6♦ opposite Qxx, and that on balance this was not a suit I would be eager to suggest as trumps for slam.

I decided to analyze this in more detail. Let’s assume partner has 3 diamonds, and that if we don’t bid the suit he won’t either, so we’ll stop in 3NT, which is a 100% contract. If we do bid diamonds, partner will support us at some point, one of us will sooner or later use a key-card bid (RKCB, Minorwood, Kickback, etc.) and we will land in 6♦ if we have 2 of the top 3 trumps and 5♦ if we don’t. We will have plenty of tricks, so the contract will hinge on not losing too many trump tricks and avoiding a defensive ruff. Here’s a table:

“pard” shows partner’s trumps opposite our Kxxxx.
“#” shows how many combinations apply, varying the x’s (or the missing honor in the last line.)
“bid” is the presumed contract we will land in if we respond 2♦.
“good” is the chance we won’t lose too many trump tricks.
“mp” is our matchpoint expectation in diamonds, weighted by fraction of cases.
“3NT” is our expecatation at 3NT.
“risk” is the weight of cases for which we can't prevent a ruff
(N.B.: obviously a lot of factors have to come together for the defense to score a ruff, it’s just the risk is a lot higher if we lose the first round of trumps.)

Bottom line: with half the field stopping at 3NT and half investigating slam, thereby landing in 5D or 5D, we rate to score less than 1% better trying for slam in this suit than settling for 3NT, and that’s assuming nothing else goes wrong, such as a ruff which we can do nothing to prevent 61% of the time.

Varying the percentage of bidders in 3NT vs. diamonds had negligible effect, best case was 100% of the field in 3NT, advantage 50.7% to 50.0%; worst case 0% 3NT, advantage to diamonds 50% to 49.3%.

pard # bid good mp 3NT risk
xxx 4 5D 33% 2.4% 4.8% 7%
Txx 6 5D 39% 3.7% 7.0% 11%
Jxx 6 5D 50% 4.0% 6.7% 11%
Qxx 6 6D 14% 3.4% 7.3% 11%
Axx 6 6D 67% 6.3% 4.4% 0%
JTx 4 5D 85% 3.3% 3.8% 7%
QTx 4 6D 37% 3.1% 4.0% 7%
QJx 4 6D 67% 4.2% 3.0% 7%
ATx 4 6D 82% 4.7% 2.4% 0%
AJx 4 6D 84% 4.8% 2.4% 0%
AQx 4 6D 96% 5.2% 1.9% 0%
HHH 4 6D 96% 5.2% 1.9% 0%
Tot 56 .... .... 50.3% 49.7% 61%

Percentage of field in 3NT 50%
Percentage of field in 5♦ or 6♦ 50%

I feel justified in claiming this is a poor suit to suggest for slam purposes, unless you have a better method of evaluating your combined trump holding than RKCB or such. I’m mostly interested here in technical criticism, i.e., did I botch the math or are there factors I’ve overlooked. But if you have methods to reach a slam with QJx but stay low with Qxx, speak up.

In computing the “matchpoints”, I more or less assumed that 3NT making and 5D making were equivalent; the possibility of making 10 tricks at notrump would shift the results against investigating slam, unless we could land at 4NT instead of 5D after, say, Minorwood.

[FrancesHinden]

lexlogan, on Dec 10 2007, 03:56 AM, said:

In computing the “matchpoints”, I more or less assumed that 3NT making and 5D making were equivalent; the possibility of making 10 tricks at notrump would shift the results against investigating slam, unless we could land at 4NT instead of 5D after, say, Minorwood.

I would have thought that if you have slam-investigatory values and a fit, you also have room to work out whether the best spot is 4NT or 5 minor, so you only lose out when 9 tricks is the limit in NT and 10 the limit in the minor, and you can't play in a 5-2 major suit fit.

Your table seems to assume the only possible contracts are 5m or 6m once you've gone past 3NT.

[pclayton]

So I guess we should never open a suit of Kxxxx for fear that pard holds a monster with Qxx of trump.

Makes sense to me.

[lexlogan]

FrancesHinden, on Dec 10 2007, 09:38 AM, said:

lexlogan, on Dec 10 2007, 03:56 AM, said:

In computing the “matchpoints”, I more or less assumed that 3NT making and 5D making were equivalent; the possibility of making 10 tricks at notrump would shift the results against investigating slam, unless we could land at 4NT instead of 5D after, say, Minorwood.

I would have thought that if you have slam-investigatory values and a fit, you also have room to work out whether the best spot is 4NT or 5 minor, so you only lose out when 9 tricks is the limit in NT and 10 the limit in the minor, and you can't play in a 5-2 major suit fit.

Your table seems to assume the only possible contracts are 5m or 6m once you've gone past 3NT.

I think it's clear that, for a minor suit, it's important to have agreements on how to stop at 4NT. But many partnerships have no key card ask other than RKCB, and by that time you're too high. This would especially be true for major suits. So the decision of whether to mention a suit such as Kxxxx in an obvious slam hunt (when you have a rational alternative) will depend heavily on your methods. The basic point here was to illustrate the potential problem -- on average, Kxxxx opposite 3 card support just doesn't cut it for the 5 or 6 level. RKCB has been a huge advance in terms of avoiding poor slams, but it doesn't completely solve the problem of poor trumps. Ideally, you should avoid asking for key cards without at least two of the top four honors. Unfortunately, there isn't always a clear-cut way to tell partner, "I love my hand except for the trumps!"

[lexlogan]

pclayton, on Dec 10 2007, 04:13 PM, said:

So I guess we should never open a suit of Kxxxx for fear that pard holds a monster with Qxx of trump.

Makes sense to me. 

Don't be absurd. The quoted hand had 18 hcp in response to a 1♠ opener; slam was extemely likely. If we respond 2C instead of 2D, we'll likely miss a 5-3 diamond fit, if it exists, but not 5-4. Even if partner rebids his suit with 6-4 we can bid the diamonds ourselves the next round. My point is that missing a 5-3 diamond fit under these circumstance may be a good idea, if the likely result of finding such a fit would be to land you at 5D or 6D. If you can stop at 4NT instead, or weed out some of the bad cases, bidding 2D has more to recommend it.

[jdonn]

I don't agree with the analysis because when partner has xxx, Txx, and probably Jxx he will or should not support directly, then he will preference on the third round, then neither player will mention the suit again. I think Qxx is the only true problem holding for partner, and maybe JTx.

[pclayton]

lexlogan, on Dec 10 2007, 03:21 PM, said:

pclayton, on Dec 10 2007, 04:13 PM, said:

So I guess we should never open a suit of Kxxxx for fear that pard holds a monster with Qxx of trump.

Makes sense to me. 

Don't be absurd. The quoted hand had 18 hcp in response to a 1♠ opener; slam was extemely likely. If we respond 2C instead of 2D, we'll likely miss a 5-3 diamond fit, if it exists, but not 5-4. Even if partner rebids his suit with 6-4 we can bid the diamonds ourselves the next round. My point is that missing a 5-3 diamond fit under these circumstance may be a good idea, if the likely result of finding such a fit would be to land you at 5D or 6D. If you can stop at 4NT instead, or weed out some of the bad cases, bidding 2D has more to recommend it.

Sorry. But I hope you see the fallacy in your argument.

I think anytime one hand has a long, weak 5 bagger, and the combined values between the two hands add up to slam, this will be a problem. It need not be Kxxxx. I can see Axxxx or Qxxxx being a problem. Sometimes we fit a key Jack or 10-9 in the other hand and sometimes we don't.

[lexlogan]

pclayton, on Dec 10 2007, 11:59 PM, said:

lexlogan, on Dec 10 2007, 03:21 PM, said:

pclayton, on Dec 10 2007, 04:13 PM, said:

So I guess we should never open a suit of Kxxxx for fear that pard holds a monster with Qxx of trump.

Makes sense to me. 

Don't be absurd. The quoted hand had 18 hcp in response to a 1♠ opener; slam was extemely likely. If we respond 2C instead of 2D, we'll likely miss a 5-3 diamond fit, if it exists, but not 5-4. Even if partner rebids his suit with 6-4 we can bid the diamonds ourselves the next round. My point is that missing a 5-3 diamond fit under these circumstance may be a good idea, if the likely result of finding such a fit would be to land you at 5D or 6D. If you can stop at 4NT instead, or weed out some of the bad cases, bidding 2D has more to recommend it.

Sorry. But I hope you see the fallacy in your argument.

I think anytime one hand has a long, weak 5 bagger, and the combined values between the two hands add up to slam, this will be a problem. It need not be Kxxxx. I can see Axxxx or Qxxxx being a problem. Sometimes we fit a key Jack or 10-9 in the other hand and sometimes we don't.

Of course Axxxx and Qxxxx are problems -- I focused on Kxxxx as a particular example. The point is that it's an avoidable problem on the type of hand under consideration. Either you have methods to land in 4NT when the trumps are poor, or it may be better to avoid bidding the suit as your first response, when practical to do so in a slam-going auction. This obviously applies even more strongly to Qxxxx. As one poster suggested, if opener does not raise immediately with xxx or Txx, that shifts the balance toward responding 2♦.

The modern style is to teach beginners to bid on shape, without regard to suit quality. This is generally correct for part-score and game auctions, but not so good for slam bidding. I recall an auction where, with a known triple fit, partner carefully selected the 4-4 heart fit: Axxx opposite Kxxx, losing two trump tricks when the suit broke badly. Either spades or diamonds would have predictably made seven, since both were solid and all that was required was to pull trumps with one suit and pitch the losers on the other. But he had never encountered the idea that, at slam, you should strain to play in the strongest trump suit.