Wednesday, September 15, 2010

The Principle of Least Commitment

Do you know all 656 of the suit combinations in the Bridge Encyclopedia?  Neither do I. Can you always visualize exactly what's out against you and evaluate every line in terms of its percentage success rate?  Neither can I.

So begins the article that I've been working on for several years which might eventually make it into some bridge magazine if I can ever perfect it.  This is the current state of the article. It seems to me to be a valuable principle, as much so as the principle of restricted choice, to which it is related, for example.  But I've never heard anyone mention anything like it.  Am I missing something?  Is it so blindingly obvious that I'm the only person to think it worth writing down?

I was reminded of it last night at the bridge club because there were two PLC transgressions at our table, at least that I noticed.  Here's one: you are in a 3NT contract with 24 hcp and you have the following suit to play: AQT94 in dummy opposite 65 in the closed hand.  You have the tempo and sufficient entries to both hands.  How do you play the suit to maximum advantage?  Well, you finesse the 9/T.  If the K and J are split then you are simply guessing.  If they're both guarded offside you're doomed to lose two tricks in the suit regardless.  But here's the case where it matters: KJx on your left and xxx on your right.  By finessing the T first, you pick up the entire suit.  If you finesse the Q first, you must give up a trick.  Least commitment.  As it happened, KJxx was on the left so it didn't matter but the declarer didn't give himself quite the best chance.

Here was the second case: You're in 3S and your trump suit is 632 in the dummy and AQT874 in the closed hand.  At first sight you might say, aha, just like last time, let's finesse the T first (as the actual declarer did, losing to Jx).  But here there are only four cards out, as opposed to the six in the last example.  You should expect to be finessing once only (not twice as before).  This despite the fact that you have an extra card in the short hand with which to finesse.  In "normal" layouts of the suit (2-2 or 3-1 splits), the cards T and below are essentially irrelevant here.  Correct play is to take the obvious finesse of the Q which has a 27% chance of picking up the entire suit (essentially, you need the K onside and a 2-2 split or some other fortuitous event like singleton J offside).

In this case, there were three losers outside the trump suit.  Our opponents had stopped in 3S where some might have been in game.  Thus, there might be something to be said for taking the safety play for five tricks.  However, as is often the case when we have no sequences of our own (here, they have none either), the "least commitment" strategy is to bang down the Ace (that takes no guesswork at all!).  Now, you increase your chance of taking 5 tricks to 83%.

If you can offer any suggestions for my description of the PLC, I'd appreciate it.

1 comment:

  1. The Principle of Least Commitment is certainly known in AI circles, and not just in adversarial situations. Planning, interpreting visual scenes, solving puzzles (what I'm up to today). It's often about filling in the details that are certain or forced before making guesses. Googling for the phrase is how I got here.