Sunday, March 23, 2014

Staying with happiness

Greed can be a tricky thing at bridge. You hold ♠87 K9732 AQ42 ♣J8 at unfavorable vulnerability. The opponents have a normal, uncontested auction to 4 and you will be on lead. Should you double? You do have tolerably nice trumps but you're also in danger of being endplayed every time you're on lead. Partner may not have much so it's possible that the opponents have close to 30 high card points.

Let's say that there are two more or less equally likely outcomes: making exactly and going down one. And let's further assume that, should you make the double, it will not help declarer make the contract (frequently it does help him). At the rubber bridge table, you are risking 170 to gain 50 (3.4 against). At IMPs, the odds are 5:2 (2.5) against (assuming your teammates are in 4 not doubled at the other table). These are not good odds. At matchpoints, a double in this situation will typically change your result from somewhere near average to either a top or a bottom, depending how it goes. The odds improve a bit if there's a healthy chance of a two-trick set. At IMPs (again we assume 4 at the other table), the outcomes are +5, +2 and -5. That's much better. But with the hand hand shown at the top, you have no expectation of five tricks at all, so why rock the boat? You're happy that they're playing the contract in your suit, right? Stay with happiness.

I've mentioned this general principle a couple of times before, especially in What makes a good penalty double? where I violated the principle myself. This time, I'd like to go into a bit more detail because there are so many ways that a double can work out badly and relatively few ways it can work out well.

Let's start by taking a rather lucrative looking example. At favorable vulnerability you hold ♠3 KJT85 782 ♣9753 and partner opens 1♠. RHO bids 2, you pass and LHO bids 3♣. Partner passes, RHO bids 3NT and LHO closes the auction with 4. Should you double? You have the best possible arrangement of assets -- partner has the entries and you have the trumps. Your hearts are sitting over declarer (and LHO probably doesn't have stellar hearts), they don't have more than about 24 or 25 HCP between them. You know exactly what to lead (your stiff spade). It's very unlikely that they can make the hand. If ever there was a time for a juicy penalty double, this is probably it. But is it without risk? Not at all.

The assumptions that we were using before for IMP calculations was that the other table would be playing the same contract undoubled. If RHO had dealt and opened the bidding with 1NT followed a transfer sequence carrying them to game, that would be a reasonable assumption. But that's not what happened here so there are several reasonable contracts: part-scores in hearts, notrump contracts, maybe your opponent bought the contract with a weak two spade bid. It makes the calculation of odds quite difficult. But bear in mind the following aspect of team play: if the contract is the same at both tables then the only scope for IMPs to flow is in the play and defense. But if the auction or play  at the other table is different, you may already be winning (or losing) the board. It's sometimes very hard to know. But making the wrong decision now can either add to the damage (or reward) or can cancel some of it out.

Let's say your counterpart's partner decided not to open his marginal hand and your teammates had an uncontested auction to 3NT by the hand on your left. Perfect defense would have set it one trick but the opponent on lead quite reasonably led his fourth-best spade and your teammates ended up making an overtrick for +630. In other words, you're already ahead. [Yes, I know that you may be playing the board first, but it makes no difference what the actual order of play is.] If they go down one at your table, you will win 12 IMPs (730). If they go two down doubled, you will gain 15 IMPs (1130). That's a profit of only three!

But we should always be on the lookout for a few extra IMPs, right? What's the downside? Maybe left to their own devices, they'd drift one off. But your double tips declarer off to a way of playing the contract that actually makes. Oops! You were winning 12 IMPs before the double but now you're losing 4. That's a negative swing of 16 IMPs! Another less drastic way that your double might backfire is if they decide to run to 4NT and you have the same defensive problems as their pair at the other table. You just gave back the12 IMPs that you were due.

There's another possibility. What if they "send it back" (redouble)? The hand quoted at the beginning was played by Alexander's opponent at a recent club game. This time, there were several factors why doubling was not a good idea. First, her trumps were not nearly as good as in my hypothetical example. Secondly, almost all of the partnership assets were in the one hand (and partner had passed throughout). Thirdly, with partner not having bid, it's quite possible that they have the values for slam but didn't bid it -- although in this case you might have been fixed already. Finally, the vulnerability was not going to give quite the big bonus that might be hoped for (5 instead of 7, 10 instead of 12, 12 instead of 15).

The actual result at the table? 4XX making by Alexander (i.e -880 and 0/11 matchpoints) instead of the possible (likely even) +50 for 7 matchpoints. In practice, of 12 declarers in hearts (9 were in game, of which four were doubled, one redoubled, 5 played the contract without comment). Of those that were doubled, 2 made 10 tricks, 1 made 9 and one 8. Of those not doubled two made it, two didn't). Of the part-scores (none were doubled) two made ten tricks, the other nine.

Although it was a match-point event, I'd like to look at it as if it were IMP pairs. Here, the datum would be -180 for the defenders and so +50 would score 6 IMPs while -420 would score -6 IMPs  The actual score for Alexander's opponent would have been -12 IMPs  It's hard to know what would have happened in the hypothetical case of no double, although I think there's a good chance declarer would have gone down without the warning of a bad split. If he was going to make anyway, then the double cost "only" 6. But if it tipped him off to the right play then it cost a whopping 18!

The moral of the story is, of course, Stay with happiness.

No comments:

Post a Comment