Sunday, June 24, 2007

Showing a Stiff Honor?

Theory is often useful to study, even if from those who use techniques that are different from your own. A study of Precision bidding can give insights into a Kaplan-Sheinwold system; understanding canape thinking can help with problem hands in natural sequences.

So, I'm not sure how many out there use an unbalanced diamond opening, but one strange auction that seems to have recurred recently came up in a completely different context, suggesting, yet again, that theory can often be translated.

For those unfamiliar with an "unbalanced diamond" opening, I am not speaking of Mr. Miles' new "My System." I am speaking of a more mundane agreement where, in a 2/1 GF or standard/natural system, minors are not opened "better minor," or "short club" in the sense of opening clubs with 4-4-3-2 only. With the unbalanced diamond approach I use, a One Diamond opening is reserved for hands with 6+ diamonds or with a stiff or void somewhere. This makes a One Club opening possible with 3352, 4342, 3442, and any other balanced hand, regardless of the minor situation. One benefit of this approach is in knowing that partner has shortness in most auctions.

For instance, after 1D-P-1S-P-2S, Responder knows that Opener has shortness somewhere; he can bid 2NT to find out where. This seems at first blush to help Responder make game tries, which is true. The real benefit, however, is that Opener need not feel like he must stretch to bid 3S, instead, because of shortness -- that is already known. Jumping to 3S both hides the whereabouts of the stiff (and the existence of one also) and also may land us too high already. When Responder knows that the simple raise features a stiff (and thus extra strength above expectation already), he can ask when interested (passing on those "too high already" hands) and actually know the location of the shortness below three of the major.

So, let me give you a strange auction. 1D-P-1S-P-4S. What on earth is that? Surely Opener, if good enough for game, would have splintered to let us know where his shortness is, right? Or, he might also have made a 4D call, a Walsh Fragment showing spade support and a great six-card trick source in diamonds, right? How can he have neither call?
Well, the simple solution is that Opener must have a stiff, but that this stiff is the Ace or King of that suit. Rather than discouraging you from valuing honors in that suit, he simply jumped to game.
This auction found a parallel in a recent on-line discussion. 1H-P-1S-P-2C-P-2S-P-5S. If the partnership is using weak jump shifts, then 1S...2S would be constructive. The slam try must be based upon 3514 pattern -- how else can Opener have a comfort zone at the five-level but could only bid 2C earlier? So, why no diamond splinter? Using the theory parallel, it seems that the 4D splinter was a bad option because the diamond stiff was the Ace or King of diamonds. Responder, who held AQxxxx-x-Kxxx-10x, can now feel good about accepting the slam try.

Something to consider for your personal system.

Tuesday, June 5, 2007

Interesting Question from a Reader

The following Q&A may be of interest to some:

I just read your book. But it was not clear to me what your requirements were for the auction 1D-2C-2S-4H. (Do I need HHxx, x, xx(x), HHHxx(x)? That is what it suggests but this seems too restrictive and low frequency.) Please read problem below. I would appreciate your thoughts. (This comes from a little private chat group we have. Pretty high quality. You would be welcome if you were interested.) Where did the following auction go wrong (standard bidding, 1D-2C not game forcing)?

Dealer: KQ9x, Axx, AKTxxx, --
opposite Txxx, x, Jx, AKQxxx


Should responder have bid 4H after either 2S or 4D? Should opener have done something other than what he did? (Yes, I was dealer and AJx were offside. I know we have all been in worse slams, and no one, perhaps other than Fred, has a way to find the SJ.)

Ned: Asks if responder's 3S was weaker or stronger than 4S (I will have my observations on that below.) He agreed with responder's bids if it was an attempt to slow things down. He suggests that opener could have cuebid 5H over 4S and stop in 5S.

Jack: "Not every 'crime' has a perpetrator..." (As a lawyer, Jack should know better than this.) "If there was an overbidder, it was opener."

Jeff: "Opener drove to slam. He might have tried 4H over 3S suggesting club shortness..." When asked about his rules for cuebidding, Jeff says that "my first cuebid below game is never shortness." "2C is fine."

Fred: "Not much wrong with the auction." He does acknowledge that playing his forcing pass relay methods he would be able to find the SJ and CJ by the five level (I am pleased to discover he did not disappoint me in my prediction), but not playing strong Club.

Barry: I don't think the auction went wrong."

Len: "Responder showed decent restraint not bidding 4H, so opener drove to slam. I've been in worse, though."

Mike: At the table: I was the dealer. I drove to slam. This was a swiss. At the other table they played 4S. All three hands came from the same event. We missed the overalls.

Afterthoughts: I agree with responder's 2C response although, at the time, I thought that 2C followed by 3S confirmed an opening hand, else a 1S response. So, at the time I was hoping to catch a better "minimum". In fact, Txxx, x, Qx, AKQxxx makes 6D a fair slam (and better than 6S). But in retrospect 2C is the right bid. So, the auction starts 1D-2C-2S (yes, in our partnership this shows enough extra's to drive to game, not necessarily a full reverse.) Should partner rebid 3C (later planning to bid 4S), 3S, 4H (splinter) or 4S?

See Rexford's book again. He is big on "picture jumps" although he defines them very strictly. I believe he would show HHxx, xx, xx, HHHxx for 2C followed by 4S and HHxx, x, xx(x), HHHxx(x) for 2C followed by 4H. He would bid 3S with all other 4 trump raises. (I don't know whether he is being too rigid, as it strikes me that HTxx ought to be pretty good trump support.) I too am a believer in picture bids although I have never defined them quite so strictly and, at the time, thought that responder might have bid 4H over 2S. I am not of the "4S is weaker than 3S school" (and I don't think it has many supporters any more.) I also am a believer in "when you have trump you raise" so rebidding 3C would be wrong in my view. So, we have 1D-2C-2S-3S. What is 3N? Again, Rexford wants to treat it is a serious slam try. He gives up on 3N any time an 8 card major fit is found. I am not sure I am buying, but this auction could certainly be played this way. My agreement with partner on our 9 card fit major suit auctions is that 3N is "non-serious." That way, responder can sign off without cuebidding when there is no reasonable chance for slam. It seems that conveying information when slam is not likely helps the defense. (Perhaps someone can explain to me why serious 3N is superior to non-serious. I haven't seen it in print.) Ok, let's say that we play "my way" and that 3N is non-serious and you evaluate opener's hand as better than that. Jeff offered the rule that "first cuebid is never shortness." I would modify that to say "first cuebid is never shortness in partner's suit." But Jeff also suggest that 4H patterns out the hand. I am not sure I think that is best insofar as it seems more natural for dealer to rebid 4D before bidding 4H (see how valuable responder's Qx of D's could become). So, I would bid 4D. (We now have 1D-2C-2S-3S-4D.) Cuebids by responder below the level of game do not show extra's. Responder should bid 4H. Ok, this is a minimum 2C bid, the trump are lousy, but not to bid 4H is to lie about your hand. Could 4H be last train? Here you need a rule. Last train is only available to the partner making a serious slam try. Either 3N if you play it serious, or a cuebid if you play it not, should create a master-slave auction (unless the master subsequently gives up the command.) Now, opener can bid 5C, responder can bid 5S (had responder held Qx of D's, this would have been a good time to show it), and opener can pass having done his all. We end up with 1D-2C-2S-3S-4D-4H-5C-5S. Would responder have bid the same with Jxxx of spades (or JTxx)? Perhaps, so blasting could still be a winning gamble.

My Response:

The problem presented has a multitude of responses.

FIRST ISSUE: First, you will recall from reading my book that I have isolated the problem of the 1D-P-2C-P-2M problem and proposed a solution. That auction causes a loss of space when the major will be raised. The solution I proposed is a modified version of the Golady convention, wherein, simplistically, Opener transfers to his major at rebid. This would radically change the auction. Now, instead of 1D-P-2C-P-2S-P-???, where Responder must bid 3S or 4S, or make a Picture Jump, the auction would instead be 1D-P-2C-P-2H-P-???, which allows Responder to agree spades with a 2S call, 4S, or a Picture Jump. (3S would be a splinter in support of diamonds.) That extra round of bidding solves a world of hurt and also allows the Picture Splinter to have the great limitations on it. Without that step, the Picture Splinter is, as you mention, quite constrained. Using my auction (assuming 2C is allowed; you need 2C to be GF or to use the alternative that 2C…3C is invitational only, which works here), then one possibility is:

2S(agrees spades)-P-
3D(good trumps for bypassing 2NT, no club A/K/Q, two top diamonds)-P-
3H(heart A/K/stiff/void)-P-
3NT(serious interest, not three top spades)-P-
4C(two top clubs)-P-

At this point, Opener has a monster. The spade slam looks to be fairly safe. He could be really conservative or really practical, though, and bid 5D. As you will recall, a five-level call here, in a side suit that cannot be void, is a strange RKCB, where the “key cards” are the aces and the K-Q of the side suit. As Opener knows about the spade KQ but does not know about the diamond Queen, this seems to be the ideal solution. Had Responder held the diamond Queen instead of the diamond Jack, 6D would be found.

SECOND ISSUE: What does this alternative auction do to help the Picture Splinter?

2H(spades)-P-4H = HHxx-x-xx(x)(x)-HHHx(x)(x)(x)

2NT(poor trumps)-P-4H = spades still HHxx, but clubs weaker than HHHx+ because of the delayed action)

3minor-P-4H = Opener has now shown two top spades, so Responder cannot have HHxx. This auction handles the Hxxx-x-xx(x)(x)-HHHx(x)(x) hand.

Given space, the stuff works.

THIRD ISSUE: What if 2C is not GF (and not GF or one-suited invitational)?

When you are in an auction that is not GF, then Picture Splinters are not used. Hence, had 2S not forced game, 4H would be a simple Splinter and not a Picture Splinter. However, your 2S forced game, which means that 2D is apparently a temporizing call (I’ve played that before). This technique works OK, but the hidden cost is illustrated by this deal. My approach has no articulated solution for this problem if you do not adopt Golady. However, my feeling is that a reverse which establishes a GF is a problem auction for many reasons.

A big question I have is why 1D-P-2C-P-2S-P-4H would necessarily agree spades – you might also need this to agree diamonds in some situations, I would imagine.

Even if not, the cramping of space is a problem. One solution might be to use 2NT as Lebensohl here. That gives a lot more definition for Responder. Not sure.

However, without better techniques, I would not advise 4H as a Picture Splinter because, as you mention, the auction is different from a classic 2/1 auction and space is at a premium. Hope that helps. Please feel free to ask anything follow-up.

Saturday, June 2, 2007

For those who like 4441 openings

A post of mine from BBO Forum took a while and may be of interest, so I copied it here:

For about ten years, I used an idea initially in a strong-diamond canape system, later in a modified minor-suit structure precision system, and then later yet in a strong-club canape system, that handled three-suiters very effectively, both for game-inquiry and for slam-inquiry auctions. It will not directly help the 2♥ start (from the post the induced this one), but some parallels might be useful for your own approach.

The idea was for 2♣ to show a three-suited hand, with 2D as the initial asking bid. This is Part I.

In response to 2D, Opener would bid 2♥ with a minimum and a hearts suit (4414/4441/1444, or possibly the 5440 possibilities with hearts and a five-card minor). 2♠ showed a minimum without hearts (4144/4045/4054).

With maximums, Opener bids one-under the stiff (i.e., 2NT=4441, 3♣=4414, 3D=4144, 3♥=1444).

This resolved much of strength/shape relatively low, meaning not bypassing the "normal" signoff levels when Responder would have escaped but has just enough to think about game, and low enough for more interesting slam auctions.You will see that 2♣-P-2D-P-2♥ is unresolved. With disinterest, Responder can pass, convert to a pass-or-correct 2♠, try 2NT (optional), or try 3♣ (minors). Some difficult hands are handled with an immediate 2NT. With stronger values, Responder can bid 3D after the 2♥ rebid, Opener bidding 3♥=1444, 3♠=4441, 3NT=4414. Thus, again sort of one under.The one-under's also gain from occasional passing. Thus, one might pass 2♣-P-2D-P-3♥ if Responder needed a maximum, heart support, and shortness in a minor, or something like that.

Anyway, back to slam. The one-under bidding allows transfers, starting with the short suit, setting trumps.

For example, 2♣-P-2D-P-2NT(4441)-P-3D = transfer to hearts. This can be passed, if 3♥ is now the limit of the hand, of course.

Example #2 ("starting with the short suit" explained): 2♣-P-2D-P-2♠(4144)-P-? 3♣ = to play. 3♥ = transfer. 3♠ = transfer (♣).If Responder rebids after the transfer, this is a slam try.So, you could have, for example, 2♣-P-2D-P-2NT(4441)-P-3D(tr)-P-3♥-P-3♠(Q)-P-3NT(serious)... "Serious" might most logically mean a void, if 5440 is possible. However, perhaps in some auctions "non-serious" makes more sense, meaning without a void, when the known pattern is minors plus a major, allowing 4♣ or 4D to define a void and define which suit has the fifth card.

Your range can be larger than a two-step range, BTW, if 2♣ is the opening. Imagine 11-15 extended to 11-17. With 11-13, minimum. With 13-15, maximum. With 15-17? Two options:

If 2D asking, bid above highest answer (3♠+ = 15-17). If you want to avoid bypassing 3NT without both majors, bid the stiff (or the other minor) above 3NT, with 3♠ being 4144 and 3NT 1444. Or, limit the 15-17 hands to 44 majors. Something like that.

If simple pass-or-correct (e.g., 2♥), bypass the "correct" option (2♠) to bid next-up (2NT) with a non-fitting maximum, or above that with a fit, describing type(e.g., 3♣ = fit plus short club, 3D = fit plus short diamond, 3♥ = fit plus short spade).

The high option can also be much higher, a split-range (e.g., 11-15, or 19-21, whatever).You can also, alternatively, change the responses to 2D to handle a split range or larger range:

2♥ = same
2♠ = same
2NT = high range (3♣ asks for one-under shortness, Opener responding 3♠ for short clubs or 3NT for short diamonds)
3♣/D/♥ = same
3♠ = short clubs.

This alternative allows even more depth. You can also change the 3♣ option (short diamonds) to 3NT, allowing 3♣ to instead show an even higher range (11-19, or a higher possibility). 3D would ask for the shortness. You can avoid bypassing 3NT after the 3D call by having 3M show shortness in the other major and 3NT show both majors, 4♣ after 3NT then asking which minor is short. You also would need to slightly modify Opener's options after a sign-off pass-or-correct option.

The options are relatively unlimited. The first version I played featured 2♣ as 4441, any three, 5-card minor possible, with 10+ (unlimited) HCP's. 2D asked, with somewhat step responses in roughly 3-HCP ranges, tied more closely to LTC than to HCPs. This worked very well, even in the rare case of competition.

Compare that with Mini-Roman 2D, with a 2NT asking bid, a hopeless idea.

I mean, we even had the ability to play a correct 2NT quite often. Who can do that after a 4-by-1 opening????