Is "finesses wrong for declarer" a good reason to double a (bridge) contract?

Question

Suppose you have a hand like (s) KJx (h) xxx (d)Axxx (c) xxx, sitting "West."

The (Standard American) bidding has been S 1 spade, N 2 hearts, S 2 spades N 4 spades. You and partner have passed so far, and E and S pass to you. Vulnerability is equal.

A book I read opined that this is a good time to double because "finesses were wrong for declarer." You have a minor spade tenace behind declarer, declarer did not support dummy's hearts, meaning your partner has "some."

I count two trump tricks and one side trick in your hand; one heart trick and one other side trick for partner, a potential two-trick "set."

Is this, in fact, a good time to double? Or are there reasons against doubling that I've overlooked?

Answer 1

Good question. One caveat to what you mention is that with every losing finesse by declarer you are potentially endplayed. If the defensive strength is predominantly in one hand, best not to double; make declarer guess whether the endplay is on or not, as he may have alternative lines that fail.

Also, there are only specific circumstances in which it is correct to double for a one-trick set. As a double is so informative that Declarer can often play the hand one trick better, double only when you expect to set the contract at least two tricks.

Further, only double when the correct opening lead is already obvious (to you or partner, as the case might be). Opponents who double and then agonize over the opening lead have put themselves in the pickling jar.

The best known (and most common) counter-example to this is in matchpoints when the opponents are vulnerable in a partial that rates to give you a bad board even if it goes down one undoubled. In that case one doubles to try and get +200, as the last hope to save the board.

Answer 2

Just to add to what Peter said:

You should double a contract if:

1. It's unlikely to go down if you don't. (Lead-directing doubles)
2. What you stand to gain if it goes down outweighs what you stand to lose if it doesn't, remembering that your double is going to help declarer.

This is situation #2, and there are no hard and fast rules to apply here. You need to listen to the auction, and figure out how many tricks you expect them to go down, then compare the value of that to what they get if they make.

On the hand you give, you expect them to down one at most, and that's if partner produces a trick, which is far from a guarantee, and you take all of your potential tricks, which is far from certain. (This isn't a weak auction. A delayed game raise like this should be showing actual spade support and a heart suit worth showing. If you find as little as S Qxx H AKxxx in dummy, you are going to be very unhappy with your double.)

Why You Lose At Bridge, by SJ Simon, has a lot of useful advice on this subject (and many others).

Answer 3

On this particular hand, North rates to have 3 spades, 5 hearts, and no minor-suit shortness to go along with minimum game-forcing values. If North's hand is xxx KJxxx Kx KQx, then you can safely exit a heart every time you're on lead, likely taking one or two hearts, two spades, and a diamond (Note that this is a hand that stretched to make game, and even this dummy may allow declarer to make, e.g. when holding AQTxxx -void- QJxx Axx).

If, on the other hand, dummy shows up with Qxx AKQJx Qx xxx, you're in trouble. Your opening heart lead may have given declarer a chance to pitch two quick minor-suit losers, and you're only taking one trump trick.

So a reason not to double with this hand is this: On the auction, partner probably has two important cards. It would be surprising for partner to have two aces, and if one of their cards is the diamond king or queen, it's probably being picked up.

---

Let's look at the odds for doubling in various situations. This table gives scoring form on the left (total points and international matchpoints), and vulnerability on the top. In each box is W/L/P, where W is the score gained when doubling and defeating the contract with one undertrick (vs not doubling), L is the score lost when letting 4Sx make with no overtricks (vs 4S undoubled making), and P is the probability that you defeat the contract by one to make doubling a worthwhile action.

|    |      NV      |      V        |
| TP | +50/-170/77% | +100/-170/63% |
|IMPS|  +2/-5/71%   |   +3/-5/63%   |

At matchpoints, of course, if everyone is in 4S, you need a 50% chance of setting or better to make doubling worthwhile.

Of course this is a simplification; it doesn't take into account the chance of taking two undertricks or the opponents making with an overtrick. On this hand, I haven't done any simulations but I would say that you certainly won't defeat opponents on three deals out of four, and probably not on two deals out of three, but may one deal out of two; I would consider doubling when playing matchpoints but not at other forms of scoring.