Inferences from the auction

May 4th, 2013  ~  Jeff Lehman  ~  No Comments 

I am often amazed at the strength of inferences one can draw from the game of bridge.

Inspired by hands that partner and I held at a recent club IMP pairs event, see what inferences you can take from the auction you hear.

Here is the system my partner and I are playing: 12-14 opening 1NT, 15-17 rebid of 1NT, and two forms of checkback for responder after opener has rebid 1NT.  The two forms of checkback are 2♣ invitational checkback and 2♦ game forcing checkback; opener, in reply to checkback, follows a “hearts first” approach, showing three hearts before four spades if responder had responded 1♥, and showing four hearts before three spades if responder had responded 1♠.

Inference #1.  What is opener’s likely distribution when you, as responder, hear this auction: 1♣-1♥-1NT-2♣-2♠?

And, thus, what bids would you consider when holding –, JT963, AQJ3, J964?

Inference #2.  What is responder’s likely distribution when you, as opener, hear this auction: 1♣-1♥-1NT-2♣-2♠-3♦?

And thus, what bids would you consider when holding K653, K, KT86, AK73?  (Side issue: do you agree with opener’s choice to rebid 1NT?)

(My) answers follow.

Inference #1.  Opener is almost assuredly 4=2=3=4.  He might possibly be 4=2=2=5, but would be likely to rebid 1♠ with that hand unless an unduly high amount of his strength were in the doubleton suits.  He might also possibly be 4=1=4=4, but with that hand, he might well have chosen either to open 1♦ or to rebid 1♠.

Responder might choose to rebid either 3♣ or 2NT.  Which of the two bids work out best is likely to be determined by the strength of opener’s four card spade suit: the stronger the spade suit, the better 2NT might work out.  3♣ seems safer, I suspect.

Inference #2.  Responder is likely to have great red suit length in his invitational strength hand.  Five hearts and six diamonds is certainly possible, while 0=5=5=3 is another possibility.

Opener has unexpectedly long diamonds in his hand.  If he had less strength in his spade suit, surely a diamond raise would be in order, even opposite an invitational strength in hand.  On the actual hand, a 4♦ call is still reasonable, as might be a pass of 3♦.

So … what happened at the table?

Responder failed Inference #1 test.  He rebid 3♦ rather than 3♣ or 2NT.

And Opener (whose choice to rebid 1NT seems pretty reasonable to me) failed Inference #2 test.  Rather than settle in diamonds on the auction he heard, he rebid 3NT.

Here are the actual hands.

You might notice that 3NT is actually makeable, by opener holding off on the first two rounds of spades.  Because the hold off voids RHO of small spades and because all of the defender’s entries are held by RHO, declarer can force out the two high heart honors and limit losses to two tricks in each major suit, while creating slow heart winners to accompany his ♠K, and top minor suit winners.  At the table, opener did not want to chance RHO gaining the lead and leading a spade through the king, and so chose to win the first round of spades and then pray for someone to hold queen-doubleton of clubs.  Prayer not answered and down two was the result.

Is down one good bridge?

April 29th, 2013  ~  Jeff Lehman  ~  6 Comments 

Board 6 was the last board of a two-session regional pairs event (Friday, April 26, Hyannis, MA) for partner and me.  I would disclose our weak notrump system auction to 3NT, were it not for fear that potential teammates might read this.

With hand locations reversed for convenience, I was declarer at a nonvulnerable 3NT contract:

The opening lead was the ♥T.  I looked at the opponents’ convention card to see if they were playing one of my favorite conventions of leading different cards from interior sequences that contain a higher honor than from interior sequences that do not contain a higher honor – favorite convention, that is, for the opponents to play.  Alas, these opponents provided no such clue to help my declarer choices.  I ducked in dummy – a finesse can always be taken later – and won in hand with the ♥K.  I played a club to the queen, losing to RHO’s king.  RHO switched to a spade, to my jack, the queen, and dummy’s ace.  I played another club from dummy, small from RHO, eight from my hand, surprisingly winning the trick as LHO followed suit small.  I continued with a diamond to dummy’s king and a spade to my ten, each winning.

I advanced another diamond and LHO rose with the ace and switched back to spades.  RHO won both the ♠K, all following, and the thirteener spade as I pitched hearts from each hand.  A diamond was returned to dummy’s queen.

I returned to my hand with ♣A, as LHO showed out.  I had reached a two-card end position, having taken seven tricks and lost four tricks:

I know that RHO owns the ♣J.  The other outstanding cards are the ♥Q9, and ♦J.

I led my heart and LHO played the nine.

Who owns the ♥Q?  If LHO owns the ♥Q, a finesse of the ♥J will land the contract.  But if RHO owns the ♥Q, playing the ♥A will fell the ♥Q and the ♥J will win the last trick.  If there is a clue from the play to date, the clue escapes me.

So … let’s see what happens if I guess wrong – something with which I have great experience.  If I play for the drop of the ♥Q and am wrong I will be defeated by one trick.  If I play for the finesse against the ♥Q and am wrong I will be defeated by two tricks because RHO will next cash the ♣J.

That’s the tie-breaker; down one is better than down two.  I play the ♥A.

Is down one good bridge?

Two part bidding logic quiz

April 10th, 2013  ~  Jeff Lehman  ~  4 Comments 

Your partner opens 2♣, strong and artificial.  Playing controls responses, you bid 2♠, artificial and promising an ace and a king.  IMP scoring. 

Now your partner bids 7♦!

Two questions: What do you bid?  What hand would you construct for partner?

What do you bid?  Oh, I forgot to show you your hand.  Does it matter?  Clue: your partner just skipped a zillion levels of bidding; it does not sound as though he is asking you for an opinion.  Conclusion: Just pass.  Shrug of shoulders, optional; thinking about what to do, waste of time.

What hand would you construct for partner?  If you fully trust partner, I think you can construct some possible hands for partner, by piecing together a number of logical clues.

Clue: Partner should be able to identify exactly which ace and which king you hold.  Conclusion: Partner holds the other three aces and three kings.

Clue: Partner did not ask about the quality of your diamond holding.  Conclusion: Partner has some length and great intermediates in the diamond suit.

Clue: Partner did not bid 7NT.  Conclusion: Partner is likely to have some slow potential loser, with hopes that such loser, if not covered by a secondary card in your hand (in which case 7NT makes, and is actually safer than 7♦), can be ruffed in your hand.

Combining these clues, you can construct possible hands for partner, such as the following (non-diamond suits interchangeable):

Opener can hardly hope to learn about responder’s possession of the ♠J, ♥Q, or ♣Q.  So, he hopes you have one of those cards.  And just in case you have none of them, he hopes for a ruff, in a layout something like this one: 

Not just bad luck

March 16th, 2013  ~  Jeff Lehman  ~  3 Comments 

At a club Swiss event yesterday, I held this hand following partner’s weak 2♦ opening as dealer: ♠AKQJ97, ♥A4, ♦T, ♣T973.

I responded 3NT.  One can envisage many hands opposite with better play for nine tricks in notrump than play for ten tricks in spades.  And partner held one of those hands.

Unlucky, I was thinking, not only that my table opponent found a heart lead but that spades were 5-1.

But more careful analysis proves that bad play contributed more than bad luck to my losing 11 IMPs on this hand.

I doubt I would have even recognized my error had spades split, but surely it is better to have played spades at Trick 2 than to have played diamonds.  When I get the bad news about spades, I can next try for ♦Qxx onside.  When that effort proves lucky, I would emerge with eleven tricks and gain 2 IMPs instead of losing 11 IMPs.

What Really Causes Slow Play

February 15th, 2013  ~  Jeff Lehman  ~  11 Comments 

Our local club played a Fast Pairs event yesterday.  The event confirmed my opinion about the real causes of slow play.

Allow me to provide some contextual background.  I play and defend slowly; OK, sometimes very slowly.  My partner, the same.  And yet, my pair had no problems maintaining the compressed schedule for the Fast Pairs event.

In my opinion, the reason my pair had no problem keeping the schedule confirms my belief that playing and defending slowly is not generally the most significant major contributor to overall slow play.  The more significant contributors to overall slow play generally are:

• Not moving quickly from the completion of one board to the commencement of the next board.  Players who review and announce each score on the Bridgemate, or who do not begin sorting their cards for the next hand until they have completed a post mortem on the previous hand generally waste a greater amount of the table’s time than players who play or defend slowly.
• Neglecting to consider anticipated issues before they occur.  Players who delay thinking about their next bid until it is their turn in the auction or delay thinking about their opening lead until such time as the auction is completed also waste a great deal of table time, by their failing to take advantage of the (possibly considerable) thinking time afforded them by the time spent by the bidders who precede them.

How did the Fast Pairs event reduce the impact of the two contributors above?  By programming the Bridgemate to not show the scores of pairs who had already played the boards, the time between boards was substantially reduced.  And because the director constantly reminded players of the time restrictions, not only were players more attentive to bidding and leading on time, but also their opponents felt “enabled” to politely remind the potential offenders of matters such as it is their turn to make the lead.  By the Fast Pairs event taking or enabling those steps, the compressed schedule presented no problem for even a pair such as mine, comprised of two players who declare and defend slowly.

For a better, and more complete, discussion of steps that can be taken to speed the game see http://bridgewinners.com/article/view/respecting-the-duplicate-bridge-clock/

Squeezy Peezy

February 2nd, 2013  ~  Jeff Lehman  ~  1 Comment 

I was having grand (slam) delusions on this hand from Thursday morning’s club game, when a (new) partnership suffered a bidding misunderstanding.  Partner played 4NT and not 7NT.

Perhaps my thinking about a grand was too much, since my partner’s owning the ♣J was surely not guaranteed and neither is the running of the spade suit, but that slam optimism is sort of normal for me, I admit.  And it can make interesting blog material.  But since he did own both black jacks, 7NT must have looked good.  Still, as another player I know has advised me, when you are within one of your contract, always think squeeze.

The squeeze chances on this hand happen to be extinguished if you receive a heart lead, but assume that you get a more passive lead instead.

Win the spade lead with the jack (or diamond lead with the A; it does not matter which), and run clubs first.  Oops, clubs aren’t running.  No matter, run the other top two clubs, pitching one from each red suit from dummy.  Next your remaining spade to dummy.  Cash the two top diamonds and two more of the spades.  That’s four spades, four clubs, and two diamonds, leaving a three card end position.

Lead dummy’s last spade.  West will keep the club master to discard a heart.  The club card in your hand having performed its job, you can discard the club.  East must keep a diamond and so will discard his second heart.  Voila!  The double squeeze works and the ♥2 in dummy (surely, you showed off by discarding the ♥6 on an earlier spade, right?) will win Trick 13 after the ♥A takes Trick 12.

A simple squeeze earned us a poor score on another board.

I am not a fan of the 2NT call or the 3NT call, but then no one asked my opinion.  And I wouldn’t have been so aggressive as to have led the ♥4 either, but again no one asked my opinion.  Besides, who cares, since I was the goat on this hand anyway?

Pard did lead the ♥4.  Looking at the ♥2 in dummy and the ♥3 in my hand, I could ascertain that pard was leading from a four card suit, most likely one that is missing the ♥A.  For some reason, I thought it might help our side if I could hide the 4-4 heart split from declarer.  So thinking, I won the ♥K and returned not the standard ♥3, but a deceptive ♥7.  (This is unlikely to fool partner, who would be unlikely to think that this opponent had bid 2NT when holding a four card heart suit.)  Partner won the ♥Q and then surprised me by next playing the ♥A and then his fourth heart.  So much for my deception.  Reasonably enough, pard exited with a diamond.  Two more diamonds followed, then a spade to declarer’s ace.  I could see that I had to protect spades and so had no problem discarding clubs (first the ♣K)  on the ♦J and then ♦Q, but when declarer proved to own the ♣8 (dummy’s four clubs had been played on the third and fourth rounds of hearts, the first round of clubs and the fourth round of diamonds), -400 was my score.

Partner quickly saw that I could have broken the squeeze by leading a club before I played my heart (and, less understandably, so could have he), but I was not thinking that declarer had no heart stopper and, frankly, never even considered switching to a club.  I choose to blame partner.  No, not for his not switching to a club before cashing out the heart suit – his play looks reasonable to me – but for his not having been dealt the ♣8.  How much is that to ask from the guy?

Hard slams to bid

January 29th, 2013  ~  Jeff Lehman  ~  8 Comments 

Using a standard base (to be described), how would you bid these two hands from yesterday’s club game?

Board 10.

Board 25 (should be vulnerable)

As a “bonus” …

Board 11.

Slam spelled backward, Part 2

January 26th, 2013  ~  Jeff Lehman  ~  3 Comments 

In an earlier blog entry, I referred to some bad luck in reaching good slams, only to discover that an unfriendly lie of the opponents’ cards doomed the slams to failure.  http://jeff.bridgeblogging.com/2012/12/04/slam-spelled-backward/

 (Slam spelled backward is “mals” and “mal” in Spanish means “bad”.)

 This blog entry is a cousin to the earlier entry.  In today’s two slam failures, the auctions suggested – at least it suggested to me – that one slam would be playable and the other would not.  I was wrong both times!

 Board 1. 

I opened 2♣.  Using control responses, partner responded 2♥.  This is artificial and shows two kings or one ace.  Having played bridge before, I deduced the latter was more likely.  

Although I had no certain way to take care of my slow spade losers and no assurance that partner would produce any diamond support, I took a flyer at slam and rebid 6♦.  At least no one can accuse me of slowing the auction.

East smartly led the ♣A.  When the diamond finesse lost, I was scoring -50.  Three pairs of the twelve who played the board reached 6♦.  One made, the other and my pair scored 0.5 mps out of 11.

Board 12. 

Not too shabby, eh?  In second chair, I again opened 2♣.  This time partner responded 2♦, artificial and showing no aces and at most one king.  This partnership does not play Kokish and so a natural, and forcing 3♣ call was my next bid.  Partner now bid 3♠.

If partner owns the ♠K and has some club support, 6NT seems like a great spot.  On the other hand, what if partner’s spades are something like QJxxx (and he has enough other “stuff”, such as one of the two other missing queens, to have disqualified his hand from making a 3♦ second negative call)?  Then, I will be locked in my hand and will lose at least one spade and one club.  So thinking, I signed off in 3NT.

Winning the ♠K in dummy and taking a losing club finesse,  Claiming +690 for 5 mps.

Is this double for takeout or for penalty?

January 26th, 2013  ~  Jeff Lehman  ~  1 Comment 

What is the meaning of opener’s double in the following class of auctions?:

Opener             Responder       Opener’s RHO

1 grape             1NT                 2 bananas

Double

Is opener’s double a penalty double or a takeout double?  Have you discussed this with your partner?  If  not, don’t you think you should?

This blog entry is motivated by a thread in Bridge Winners website. http://bridgewinners.com/article/view/should-i-play-doubles-your-way/

One might play the double in multiple ways:

• Always as penalty – this will teach the whippersnappers to interfere with my auction!
• Always as takeout – who nowadays plays any low-level doubles as penalty?; we need to compete!
• Sometimes as penalty and sometimes as takeout – it all depends upon the tempo and emphasis of my double; who, me cheat?
• Sometimes as penalty and sometimes as takeout – it all depends upon which suit is grapes and which suit is bananas.  Yes, this is the right approach … but how to play this approach is the subject of this blog entry.

In part because I have read through the Bridge Winners thread, I can think of several ways of using the suits bid by opener and opener’s RHO to define when the double is for penalty and when the double is for takeout:

• The double is for takeout when grape is a major, and for penalty when grape is a minor.
• The double is for takeout when grapes outrank bananas and for penalty when bananas outrank grapes.
• The double is for takeout when responder has not bypassed 1 banana to bid 1NT and is for penalty when responder has bypassed 1 banana to bid 1NT.

Agreeing on any of the above three options is better than having no agreement at all.

Because the third bullet above is the most comprehensive, let’s assume that that has become the partnership’s agreement and let’s explore how that works.

1♠        1NT     (2♣/2♦/2♥)

Dbl

Because responder did not bypass 1♣/1♦/1♥ to respond 1NT, opener’s double is for takeout.  A typical hand pattern for opener would be five spades, one in the suit of the overcall, and 4-3 in the other two suits.  Responder’s most likely hand type is one where he has length in at least one of the two unbid suits, but had insufficient strength to bid a new suit at the two level.  Doubling is the action by opener that will inform responder to compete in his longest suit.  Let’s say that the overcall was 2♦ and that opener holds this hand AQxxx, KJx, x, KJTx.  Opener doubles 2♦ and finds responder with xx, Qxx, Qxxx, Axxx.  Responder bids 3♣, and opener and responder have competed over the opponents’ eight card fit at the two level to their own eight card fit.  All the Law of Total Tricks fans will approve.

1♣        1NT     (2♦/2♥/2♠)

Dbl

Because responder did bypass 1♦/1♥/1♠ to respond 1NT, opener’s double is a penalty double.  (The inference from responder’s having bypassed 1♦ is much less strong than the inference from responder’s having bypassed 1 of a major suit, admittedly.)  Why does responder’s having bypassed those suits mean that the preferred meaning of opener’s double is for penalty?  Because responder’s being short in the overcalled suit (at least short enough to have skipped responding in the suit) means that opener is increasingly likely to be long in the overcalled suit.  If opener, who sits immediately “over” the overcaller, has both length in the overcalled suit and strength in the overcalled suit, a juicy penalty double might await.  Let’s say that the overcall was 2♥ and that opener holds this hand AQx, KJ9x, xxx, Kxx.  With a probable three trump tricks, the penalty double will pay handsomely even when responder has some nondescript 1NT response such as Kxx, Qx, QJxx, xxxx.  Oh, don’t tell me that your opponents wouldn’t overcall 2♥ on xx, AT8xxx, Axx, Ax.  I have seen many, many more dangerous overcalls than this one.  And down two doubled at any vulnerability, when your side has two balanced hands and only 21 HCP combined, is very tasty.

As is the case with any bridge agreement, one of the most significant “ancillary” benefits of the agreement is derived from the negative inference when the agreement is not employed.  When opener passes the overcall in a situation where opener could have made a takeout double, responder can infer that opener has some length in the suit of the overcall.  If responder, too, has no shortness in the suit of the overcall, responder can often pass out the overcall, feeling comfortable that the opponents have settled in a contract where declarer will find his side’s tricks hard to come by.  Similarly, when opener passes the overcall in a situation where opener could have made a penalty double, responder can infer that opener does not possess both length and strength in the suit of the overcall.  With no special length or strength in the suit himself, responder can compete, feeling comfortable that each side can make a low level contract and pressuring the opponents to either choose to defend or to compete to a higher, more dangerous level in the overcalled suit.

Signaling Partner

January 26th, 2013  ~  Jeff Lehman  ~  2 Comments 

Playing in the last round of a club Swiss, where your team has a chance to win the event and is, in fact, playing the leaders, you are faced with a defensive signaling issue  … one that is influenced not only by partnership agreements, but, more importantly, by the attribute of your putting yourself into the shoes of your partner:

(Please disregard the “N” and “W”; the designations “East” and South” are correct.)

Partner leads the ♦A and declarer calls for a small card from dummy.  What is your defensive plan?  How does that plan affect your Trick 1 signal?

Dummy seems to have overbid his hand by treating it as a limit raise.  And, worse, seems to have recognized that by making an undisciplined bid of 5♠ over your partner’s raise to game.  If you can set 5♠, your chances of winning the event are surely enhanced.

You hope that partner has raised with only four diamonds so that his ♦A survives Trick 1.  Where are your other two tricks coming from?  You can begin with an assumption that declarer has no spade losers.

Are there quick losers in dummy that can be pitched on a second suit of declarer?  Since you own the ♥A, you can tell that declarer will not be able to immediately pitch clubs on hearts.  Can he pitch three hearts on clubs?  If declarer has AKQxx, he will be able, if allowed, to draw trumps and then pitch three hearts on the clubs, to win 12 tricks – this is giving declarer a hand such as AKxxx, xx, x, AKQxx, certainly a possibility.  You can hold declarer to eleven tricks if partner can be induced to lead a heart at Trick 2.  That seems to be the best you can do on this hand.  How can you induce partner to lead a heart?

Well, if partner knows that another diamond is not cashing, then perhaps your Trick 1 signal should be a suit preference signal.  Will a highish diamond get partner to lead a heart?

Perhaps.  But only if partner and you are on the same wavelength.  And why should partner be expecting suit preference to be the signal of choice?  From the perspective of your partner, you have not promised five diamonds.  The double of an artificial call of 2♦, when the vul opponents have a 9+ card fit in spades, is unlikely to be punished and you might well have taken advantage of that knowledge to have doubled on a four card suit.  No, to my way of thinking, you cannot make a suit preference signal at Trick 1 – at least in this situation – when your normal signals when partner leads are attitude.

OK.  So if your Trick 1 signal is attitude and not suit preference, should you encourage or discourage?  Well, if you encourage, you are likely to get a diamond return.  You know that that is not constructive to the defense because, assuming declarer does not ruff Trick 1, he will surely ruff Trick 2.  Your heart winner might go away, giving declarer an overtrick.  So, let’s try the effect of discouraging on the diamond lead.  That might get partner to switch suits.  But will he switch to hearts?

From partner’s perspective, he might be fearing that clubs will be pitched on hearts and not vice-versa.  The declarer hand that partner might fear is something like AKxxx, AJxxx, x, xx.  If that is the case then partner (who, in this construction, owns one of the top two club winners), will lead a club, so that your side can cash the first three tricks.  From partner’s perspective, if he does not lead a club, declarer will draw trumps, run hearts to pitch dummy’s two club losers and win twelve tricks in the form of five spades, five hearts, and two club ruffs.  Remember, partner does not know that you own the ♥A.

So … what should you do?

I think you should encourage a continuation of diamonds.  That’s because the one thing that an encouragement of diamonds should accomplish is discouraging partner from leading clubs.  And the lead of a club can be giving away the contract should declarer own AKxxx, xx, x, AQxxx.

Interestingly, at least to me as one who is a big fan of Granovetters’ Switch in Time, a defensive signaling system built on Obvious Shift principles will work great on this hand.  First, that is because the Granovetters teach that no signal at Trick 1 is a suit preference signal – that eliminates a concept that can cause partnerships to suffer misunderstandings.  Second, that is because clubs is defined as the Obvious Shift suit (if the hearts were headed by only one honor [in which case, up to three heart tricks might be available if the layout were different], hearts would be defined as the Obvious Shift suit).  Therefore, the Trick 1 attitude signal would revolve not around diamonds, the suit led, but around clubs, the suit of the Obvious Shift.  If you own either the ♣A or the ♣K, you would discourage on the diamond lead, in order to encourage a shift to the Obvious Shift suit of clubs.  And if you own neither the ♣A nor the ♣K, you would encourage on the diamond lead.  Yes, encouragement is likely to get you a diamond continuation rather than a heart switch.  At matchpoints, perhaps that is a problem since the ♥A might go away if declarer can pitch dummy’s three hearts on his projected AKQxx of clubs and you will find that you are saddled with -680 instead of -650.  But at IMPs, I would contend that your signaling should revolve around getting partner to avoid switching to a club.

So … what happened at the table?  Well, South discouraged.  North, who owned the ♣K, switched to a club.  Declarer did, in fact, own a 5=2=1=5 hand with solid spades and AQ-fifth of clubs and brought home a contract that has three tricks “off the top”.  Alas, while South’s defensive signal was not well-considered, North must share the blame.  North’s held J98x of hearts and should have been able to figure out that there was, in spite of South’s signal of discouragement, no compelling reason to switch to clubs.  A diamond continuation would have set the contract.  And, yes, probably have won the event.  Sigh!