The Submarine Ruff

by Matthew Kidd


This is board 4 (rotated) from the January 5, 2014 San Diego unit game at Adventures in Bridge. We ended up in 4 after an embarrassing auction that suggests we should spend more time discussing our constructive bidding agreements. Nonetheless, it is a good contract.

The lead was a diamond from the unappealing KJxx, a holding much maligned in David Bird and Taf Anthias' pair of books on opening leads based on double dummy simulations. Partner took the ace, hooked the ♠J, drew a round of trump, unblocked the ♠A, returned to the dummy with the J, and pitched the losing diamond on the ♠K. Hooking the ♣K brought her up to ten tricks.

We picked up an 87% for the board, which seemed like an undeservedly good result because eleven tricks were available double dummy. All 16 N-S pairs were in a heart contract, eleven in game which made five times and went down six times. The remaining pairs stopped in 3 with the exception of a single pessimistic 2 contract. A low diamond was led eight times. Not one declarer in this fairly strong field managed eleven tricks. And given how long I stared at the recap sheet, I’m not sure I would have seen the possibility either at the table.

The solution is simple; we teach it to beginners. Just ruff a club in dummy. But I think this is hard to see because most of us give up immediately on a ruff when presented with a 4-3-3-3 dummy. When we see a 4-3 side suit fit between dummy and our hand, we at best hope for the 36% a priori 3-3 split. Beating the 4-2 break with a ruff never occurs to us.

Seeking a club ruff does require care. Since two clubs must be lost before the last one can be ruffed and the J is needed as an entry at the right time, declarer can not draw any trump immediately; otherwise East can duck the club lead from dummy in the position below so that West can get in with the ♣J to lead the last trump. How many defenders will get this right?

Curiously, on the given layout, declarer is tightroped in Rodwellian parlance, required to find the one correct play, up to equivalent low cards, from both hands for the first eleven tricks until left with AK. This statement assumes the defender returns a trump when in with the ♣A or ♣J. If they force with a diamond, declarer can get away with drawing one round of trump on the given layout before continuing clubs. But doing so is anti-percentage because the third club may be lost to a defender who still has a trump to return.

It might be argued that not drawing any trump except the J transportation before cashing the ♠K, runs the risk of ruff. Since a 6-1 spade break is fatal regardless of how the hand is played, we can restrict our attention to the 5-2 and 4-3 spades splits which occur 30% and 70% respectively, given that West has four or five diamonds for the low diamond lead, and does not have a singleton club based on the failure lead it. When spades are 5-2, the diamond loser will still be discarded and declarer will gain back the lost trick when clubs are not 3-3 and the defender short in clubs either does not have the T or holds it singleton. I wrote a short program to loop over all distributions where West has two or five spades, four or five diamonds for the low diamond lead, and does not have a singleton club. Accounting for the relative probability of each distribution and location of the T, declarer gains the lost trick back about 32% of the time. So not drawing trump only hurts declarer about 20% of the time, i.e. 30% × (100-32%).

Declarer gains a trick when spades are 4-3, clubs are not 3-3, and the defender short in clubs either does not have the T or holds it singleton. A similar short program, shows that declarer gains a trick 14% of the time. So it turns out that the risk of a spade ruff slightly outweighs the reward of scoring a club ruff when clubs are not 3-3, despite the club ruff working well on the given layout.

If dummy has the T instead of the 9, prospects for the club ruff are much better. Declarer loses an unrecoverable trick to a 5-2 spade break 11% of the time while enjoying a 39% chance of gaining an extra trick. And this is this would not just be a matchpoint play. If the ♣K did not come home on a finesse, declarer would need the fourth round club ruff just to make the contract.

I’m not the first to point out a hand like this. I’m pretty sure Larry Cohen had a Real Deal column in the Bridge Bulletin several years ago with a similar theme. But I don’t remember him naming the play so I’m going to call it the submarine ruff. Four tricks deep, it hides well below the surface of the placid 4-3-3-3 dummy.