Wednesday, December 30, 2009

A poker player's guide to bridge

The other day, I had the pleasure of partnering Kathy, my sister-in-law at the bridge club.  Kathy has had quite a bit of success at the club and at tournaments, but isn't as expert at bridge as she is at poker (she's been a dealer and player).

A hand came up which I think typifies the importance of raising partner with support as soon as possible, in much the same way that you raise in poker when you're in the right position and have a good hand but one that might easily become less good after the flop, or the next "street".

Here is the hand in question: ♠QT842 AT8 54 ♣Q65.  All vulnerable, RHO deals and passes.  You pass and LHO bids 1.  Partner bids 2♣.  RHO now doubles and the time has come.  What do you call?

Before deciding on that, let's consider what's going on here.  LHO has made a 3rd-hand opening which at this point may or may not be based on a solid opening hand.  RHO didn't have enough to open but has enough to double, presumably she has both majors and enough points to force the partnership to the two level.  Whose hand is it?  Who's got "the nuts"?  Well, right now, we have the best two hands at the table (we have at least half the deck and what's more, we've got a fit).  We're ahead, but we might not be for long.  It's our duty with the assets we have to make it as difficult as possible for the opponents to draw that magic card on "the river".  In other words, we must raise!  With all our defensive values and being vulnerable too, it makes no sense to go jumping around preemptively.  But it seems to me that an immediate raise to 3♣ would be normal, even automatic.  It's reasonably likely that we would buy the hand for 3♣, making.  Or RHO might feel that it was necessary to show the good spades (5431 distribution) and bid 3♠ which we can double for 200 or 500.

What actually happened was that the hand shown passed, LHO rebid 2, confirming a decent suit, and RHO now showed her shape and strength (11 hcp in fact) with 2♠.  Kathy, who held this hand, now bid 3♣, but the damage was done.  RHO finally supported opener with a 3 bid and there it rested.  It's unusual to be punished so harshly for such a relatively minor lapse but -110 turned out to be worth exactly 0 matchpoints.  The other 10 tables were all going down in three or four spades (I suspect most of the dealers opened 1♠) or, in one case letting us go down quietly in 4♣ (the second worst score for our direction).

The old bridge adage of support with support really is a good one!

Somehow we landed in 6NT

The title refers to the excellent and entertaining book of that name by David Bird.  It's full of interesting 6NT hands from top level play and contains many instructive quizzes on how to play such hands, generally based on constructed deals.

The general rule that I've worked out for deciding when to bid 6NT at matchpoints, as opposed to 6 of a suit, is this (leaving aside for now any indications such as a Lightner double):
  • if you're in 6 or 6♠, and it looks like you'll take the same 12 tricks in NT (and you have good stoppers in any enemy suit, if any), then bid 6NT;
  • if you're in 6♣ or 6, stay where you are, unless you're fairly sure that you'll make the same number of tricks (or more) in notrump.
The reasoning behind this is that most players, at the club rather than the Blue Ribbon pairs, will not find the minor suit slam in the first place so you might as well maximize your chances of making it by playing in a suit (you usually have decent trumps when you're at the six level in a minor).  On the other hand, even the weakest pair in the field will be in 6 of a major when they have a fit with 30+ hcp.  That's when you have to get an edge with the extra 10 points.

A couple of interesting hands came up last night at the club.  On the first hand, all are vulnerable and you LHO deals and passes.  Your hand is ♠9432 AQ942 AQ2 ♣K.  Partner bids 1♠ and RHO passes.  Without going into detail, you discover that partner has both black aces and at least Kxx, a void in diamonds and is enthusiastic about slam.  You will be declarer if the contract is in notrump.  Let's say you know that he doesn't have the ♠K.  Which slam do you want to play in?

I think this is a clear case for 6NT.  If they lead a diamond, your otherwise wasted Q will become a trick.  If they don't you should have time to develop at least four spades, five hearts, a diamond and two clubs.  As it happens, the ♠K was onside and we made 1460.

Of the 16 results on this board, 2 pairs failed to reach slam at all (one of these was a good pair who perhaps had a misunderstanding).  13 pairs played either 6 or 6♠.  Only one pair bid 6NT.  Unfortunately, not us.  So, bidding 6NT was worth 6 matchpoints out of 15.  Why only 6?  Two declarers in 6♠ apparently eschewed the trump finesse because a heart ruff was looming.  That's the only explanation I can come up with for not taking all the tricks.

The second interesting board had a somewhat shorter auction.  Red versus white, you pick up ♠T76 AJ52 QJ76 ♣AK and again you are fourth-in-hand.  This time, dealer does not pass.  In fact, he opens with 5♣!  Partner contributes 5 and RHO passes.  What's your call?  Obviously nobody will be bidding again unless you choose 6♣.

Assuming that partner isn't a lunatic, you're surely going to bid a slam of some sort.  Which is it to be.  My favorite partner (the one I'm married to) suggested 7 which would have scored a nice 14.  My actual partner chose a more conservative 6 which was worth 10.5.

I'd like to think that if I had held this hand at the table, I'd have bid at least 6NT, if not 7NT, which would have scored either 14 or 14.5, respectively.  After all, I have two stoppers in the enemy suit and nothing that partner could ruff in my hand, and not much that could be profitably ruffed in partner's hand.

I would have been interested to follow this board around the room to see exactly what transpired at each table.  The deal itself looks like something from one of those goulash tournaments that are popular online:


AK5

87

AT985432




2

Q93



T98765432


QJ9843

KT64

K

QJ


T76

AJ52

QJ76

AK


Just look at that club suit in the West hand!  Nine of them to the T - two straight flushes back to back!  Had I held this hand, there's no way I'd have opened only 5 clubs!  Without the Q, I'd be bidding 7♣, but with the actual hand, I think 6♣ is about right.  That would make it quite awkward for North, although I think 6 would still be the choice of the brave.  But getting to play 6X would have been worth 10 matchpoints for the preempter.


If you're still with me, there's just one other interesting hand that might also make 6NT on really really bad defense.  I held this hand: ♠T732 T543 4 ♣JT86 as dealer with nobody vulnerable.  Not much to write home about!  I passed, although a psychic 1♠ or 1NT couldn't possibly have yielded a worse score.  LHO opened 1NT (11-13 if I recall correctly), partner showed a single-suited hand and they ended up in 3NT after opener admitted to at least four hearts.  Partner led a diamond (what else?) and dummy came down with ♠AQJ6 J6 QJ8 ♣K932.  The Q was played from dummy and I followed suit.  Declarer played the 9.  The J was led to partner's Q.  He was looking at an original diamond holding of KT7652 with another certain entry (the A).  If my holding was 43 as appeared possible, another low diamond would force the A and all the diamonds would now be good.  Unfortunately, declarer was hiding the 3 and so was able to make three diamond tricks when partner continued the suit.

But now take another look at my hand.  Do you see that I'm triple-squeezed at trick 3?  Fortunately, I know my squeeze defenses and I was careful to give declarer a trick in his hand (by pitching a heart) so that the squeeze didn't become progressive.  As it turns out, with partner holding the HA, the hand can never make 12 tricks but that fatal weakening of our hearts did allow declarer to duck a heart and claim the rest.  Kudos to the declarer for a well-played hand and a deserving top.

Never a dull moment at the bridge club!

Tuesday, December 22, 2009

Suit Combinations and the Principle of Least Commitment

Are you one of those people who has learned all 656 of the bridge encyclopedia's suit combinations?  And you know how to extend the methods to all those that it doesn't cover?

Neither am I.

That's why I have been tinkering over the last few years with something I call the Principle of Least Commitment.  Basically, it works like this.  Instead of trying to remember the proper play at the table, just make sure that you take advantage of any useful sequences that you have.  Choosing a card from a sequence means you aren't committing yourself as much as if you play an isolated card.  Failing any useful sequences, finesse the lower of two significant cards first.  Again, you don't make such a big commitment at your first try.  For the full discussion, follow this link.

For example, you have a suit laid out thus: AQT2 opposite 753 and you are in a normal contract wanting to take the most tricks at matchpoints.  It may not seem to matter much whether you finesse the Q or T first, but it does.  If both honors are wrong you're going to lose two tricks in the suit.  If the honors are divided and you guess wrong then you'll lose one trick.  But if the honors are both on your left, then everything's roses, you say?  Well not quite.  Suppose you finesse the Q.  It wins.  Now you come back to finesse the T.  LHO has the K and J and can happily split his honors.  But if you finesse the T first, it too win and now when you come back to finesse the Q, it also will win and you've now got three tricks (four if the enemy cards split evenly) from the suit.

A suit combination came up the other evening at the club like this: 9 opposite AQT7643.  The contract was a part-score at our table and this was the trump suit.  It was a very normal contract so it was important to take the most tricks.  There is a sequence here: the T9.  But is it significant?  Well, the 8 is a significant card in the defenders' hands so it seems reasonable that the sequence is significant.  But is it sufficiently significant to outweigh the fact that we have only one chance to finesse and therefore should take the "obvious" finesse of the Q?

A careful analysis shows that it is better to run the 9 than to finesse the Q, but only slightly.  Here are the layouts that matter with the number of tricks from this suit (layouts assume that the length is in hand, the 9 is in dummy):

LHO/RHO layout
Probability
Q tricks
9 tricks
Jx – K8x
6.8%
6
5
J8 – Kxx
3.4%
6
5
K – J8xx
2.8%
4
5
Kx – J8x
6.8%
5
6
K8 – Jxx
3.4%
5
6
KJ8x – x
5.7%
4
5

As you can see, what you do only matters in 37.5% of the situations.  The top two cases both favor the finesse of the Q, while the lower four cases all favor running the nine.  Running the nine will win overall 8.5% of the time and is the right strategy for making the maximum number of tricks (your expectation overall is 5.47 tricks).

Unfortunately, I wasn't able to apply my principle of least commitment at the table because these hands were held by my opponents.  And even more unfortunately, the declarer was someone who knew his business.  But, fortunately, it was one of those times when the optimum play failed (it was in fact the first case mentioned in the table above).  Every other declarer must have finessed the Q because we were the only pair sitting our way to take four tricks.  And the defense on the hand is relatively obvious (the long hand was always going to be declarer and the opening leader is always going to start with his AKxx in a side suit).

Watch this space for more exciting developments in the principle of least commitment!

Sunday, December 20, 2009

Discipline, discipline

You pick up the following hand in second position, not vulnerable vs. vulnerable: ♠QJ86542 6 K6 ♣QJ5.  What's your call?

You do have a seven-card suit so naturally you think about opening 3♠.  But let's look a little more carefully.  The suit quality is pretty ragged and you have quite a bit of defensive strength outside your suit.  Also, in second position, you should be the most disciplined when preempting.  After all, one of your opponents has already denied an opening hand and partner could be sitting there with a moose.  In fact you almost have an opening bid of your own, but again you recall that you are in the most disciplined seat.  Maybe 2♠ would be the right opening?  But again, in this seat you should have a perfect weak two, not some maybe 1, maybe 3 bid.

So, you decide to pass and await developments.  LHO passes too and it is indeed your partner who has an opening hand: 1♣.  RHO now comes in belatedly with 1 and you clearly have enough for a (somewhat) forcing 1♠.  This is met with double on your left and everyone passes.  Apparently, this is a penalty double (our opponents are the GIBs) and when LHO leads 3, you're not terribly surprised to find partner devoid of spades.  He has AQ952 A82 ♣87432, not the most robust fourth seat opening (it doesn't even come close to the famous Cansino (Pearson) 15 count.  You could be in a spot of bother here, but there's one huge consolation: you didn't open 3♠ and get doubled there!

You don't play it perfectly (i.e. guessing that RHO holds T9 doubleton of spades) and you only make your contract, but it's still good for 6.3 imps.  Discipline really does pay!

Wednesday, December 16, 2009

A good showing at the club (championship)

Len and I were mostly on the same page last evening, resulting in a 1st E/W, 2nd overall at the NWDBC club championship game.  We didn't play or defend one doubled contract all evening.  We got to two good 6 contracts that earned a 14 and a 15 (15 top).  Our opponents bid 6NT twice, once making for about 32%, once going down 2 for a top.  We also missed a good heart slam on 25 hcp.

Still we had a few too many avoidable errors, and some which were perhaps just a little too hard to work out at the time.  For example,you hold ♠J64 762 AKJ ♣K952 after RHO has opened 1NT (14+:17) and has shown a five-card spade suit to reach 4♠.  You lead the A and dummy comes down with ♠QT53 JT4 Q83 ♣AT75.  This fetches 3-2-6.  Now what?  If you don't switch to a low club now, you're going to be -620 for a very poor score.  I woodenly continued with a diamond, setting up the queen for a club discard.

One instructive issue came up.  You hold ♠85 AK73 987 ♣5432.  LHO passes and partner opens 1♣ and you respond 1.  LHO now comes in with 1♠ and partner jumps to 2NT.  You decide to bid 3♣ (not clear whether this is forcing or not, given the competition) and partner bids 3.  What's your call?

The first issue is this: how many hearts does partner have?  We play support doubles so 2NT denied as many as three hearts.  Or did it?  I think it should, and I believe that's the way most experts play.  So 3 should show a good two-card holding, such as Qx.  In fact that is what it dummy had.  Thanks to some helpful defense, we bid and made 4 on the super-Moyesian fit for an 11.  As it turns out, 4NT and 5 were also making because both of the spade honors were in the overcaller's hand.

Here is the whole hand:


AKT92

962

63

KT9


85

AK73

987

5432


Q43

Q5

AKJ

AQJ87


J76

JT84

QT542

6

Deep Finesse says that we (E/W) can't make a heart contract at all.  But it turns out that's an error (not the first that I've found with Deep Finesse).  E/W can make 3.

Monday, December 7, 2009

Showing a fit - splinter bids

The conventional wisdom states that if you show a splinter and then ask for key cards yourself, that keycard ask is exclusion keycard and of course your short suit is a void.  The reason behind this is that a splinter should be a limited bid and that if partner signs off in game, you now show extra good news (it might not actually be all that good for partner) with the void.

I've never really liked this conventional wisdom but now I think I see the reason.  A problem arises if partner takes charge and bids a small slam.  If you have a really good hand, you know that a grand will probably make but without a void in your short suit, you can't override partner's decision in case there is a loser there.

Here's an example of a hand that is getting close to this maximum strength for a splinter.  Sitting South with the GIBs on BBO, I held ♠AQJ9 K643 3 ♣K542.  My GIB dealt and opened 1.  With no particular understanding with Mr. GIB, I simply bid 4, which seemed the obvious bid.  My GIB took charge and, after finding that I had two key cards and the ♣K, put us in a grand slam, taking 13 tricks and as many imps.  It actually wasn't the greatest grand slam and indeed only one other GIB even took 13 tricks.  Given that my GIB had ♠2 AQ9852 AQ64 ♣A6, where the Q is worthless, he might have been slightly more conservative.  OTOH, he has a magnificent hand outside the diamond suit and he can almost count 13 tricks.  Ruffing that fourth diamond might be tricky.

No other pair reached 7 and only half of the rest even bid a small slam.  I thought I'd look to see which other players (all human) with my hand had bid 2NT and which had splintered to see if the splinter was the key to bidding the grand.  The answer absolutely astounded me!

No other South splintered!  Only two bid 2NT!  What other bid is even possible?  Well, here goes: four votes for 2♣ (including one where RHO had overcalled 1♠);  four votes for 1♠; two for 4; one for 3; and, unbelievably, two for 2!

I realize that many of the people who play with the GIBs probably can't get good partners of their own, and it's quite likely that several of the humans don't know that GIB plays 2/1.  But this demonstrates to me that the ignorance of people regarding how to show a fit for partner is unbounded.

Friday, December 4, 2009

A knotty problem

The scoring table demands that when looking for a minor-suit slam, we should be able to stop in 4NT, thus scoring 430 (hopefully) instead of 400 (or 630/600).  For example, 1♣ (p) 2♣ (p) 4♣ (p) 4♠ (p) 4NT is to play where 4♣ asked for keycards and 4♠ shows not enough

What about when the fit is first established at the four-level and that bid itself is not a keycard ask (or when playing kickback)?  Partner and I had the following auction this evening (opponents silent throughout): 11 – 1♠ – 2♣ – 22 – 3 – 4 – 4NT.  (1) 11-16 hcp, at least two diamonds, may have longer clubs up to 5; (2) fourth suit forcing to game.  After my club rebid, I showed at least 8 cards in the minors (usually 9) and therefore at least three diamonds.  After my heart "raise", I showed either 0445, 1435 or perhaps 1444 shape.

My 4NT bid was intended as non-forcing suggesting that we play in 4NT because my hand is going to be a disappointment for a diamond slam.  In fact my hand was ♠3 AQT7  A97 ♣KT632.  My partner thought that 4NT was a cuebid showing a heart control (since 4 would have been the keycard ask).

While it makes sense for 4NT to be a cuebid when our agreed suit is a major, I don't think it makes sense when our agreed suit is a minor, for the reason stated in the first sentence above.

If I was enthusiastic about slam, I could bid 4 (keycard ask), cuebid 4♠, cuebid 5♣.  Admittedly, there is now no way to show a heart control, but it seems to me that we gave that possibility up when we agreed to play Kickback.  How often would I have the hand with two quick losers in the unbid suit (hearts) such that I wouldn't feel comfortable going straight to the keycard ask?  Especially when I actually "raised" hearts earlier myself.

Comments welcome.

Thursday, December 3, 2009

Daytime bridge

Gotta love those day-time bridge players!  Warning: the following is likely going to sound mean and snobbish.

Here's a hand from a recent day-time game.  Imagine yourself holding this hand: ♠A986543 –  72 ♣J432.  Nobody is vulnerable and RHO deals and passes.  Do you bid 3♠?  If you do (I wouldn't) you might just escape for -200.  Incidentally, why do I think this is not a good preempt?  Because you don't have the ♠Q!  Your Ace is likely to win a trick whatever contract you're in.  Besides that, partner might have ♠K72 65432  A3 ♣AQ5 and you're odds-on to make 4♠ (you might conceivably take all the tricks on a club lead!) but partner is probably going to pass you out.

So, in any case, you decide to pass.  LHO now starts proceedings with a bid of 1♣ (playing standard American more or less).  Your partner overcalls 1, and RHO now bids 2 (natural).  Now, do you bid 3♠?  Of course not!!  There are several reasons why your hand is even less suitable for a preempt now:
  • RHO has significantly limited her hand (she passed initially but now shows 10+ with a good diamond suit) – this means that LHO has a very good idea what the total assets of their partnership are and whether to bid on or nail you.
  • Your partner has shown good values in hearts, the one suit you don't have.  On a bad day (this hand, in fact) you can never even get to dummy to cash partner's two heart tricks)!
Nevertheless, despite your hand going seriously down the toilet since you picked it up, you now venture a 3♠ call.  LHO decides to double and you know that you're in trouble now.  Partner, apparently is catching the spirit of let's make as many mistakes on one hand as possible, and now bids 4.  What this means is hard to fathom but naturally, RHO now doubles this.  What now?  Could it be that diamonds is actually our best spot?  We do have two of them!  No, you didn't come here to while away the hours as dummy!  So you bid 4♠ and, predictably, this is doubled on your left.  You take half of the ten tricks you've contracted for and lose 1100 points.  And for what?  LHO wasn't even going to bid 3NT (with 13 hcp, badly placed heart honors and a 4414 pattern, he was thinking that 2NT might be high enough, although it turns out that 9 tricks can be made with careful play in notrump for a score of 400).  Incidentally, 4X might have saved a trick (-800 instead of -1100) but double-dummy defense will probably still get 1100.

The dramatis personae?  All I will reveal is that I was "LHO" and my hand was: ♠KQT7 QT97  6 ♣AQ96.  I don't think I've ever felt quite so confident about a penalty double of a 3-level bid.  The fact that they went on to the four level was a case of Christmas come early!