Defence bidding (1)

[bridge2k]

double.

[Blofeld]

Pass. Double should be P/C on general principles.

 

I'm willing to take the '1 in 6' line on doubles.

[han]

An interesting, though usually undiscussed aspect of this debate is length of match.

 

In a short Swiss match, a single board can easily win or lose the match - and if you double the opponents into game you lose 11.

 

In a longer match, such as a 26 board K.O., there is more time to recover.

 

In the big matches, where 90+ boards are played, there is more time still.

 

And imp pairs is a whole story in itself - kind of a mix of MPs and imps.

 

But it all comes down to risk/reward at team imps.

 

If you double 3H vul and it makes, you lose 11.

If it goes down 1, you gain an additional 2 imps: 3 verses 5.

Break even point is close to 6 good doubles to 1 bad.  As the object of the game is to go plus, breaking even like this isn't the answer.  For me, I think I'd like to get this up to more like a complete game swing: about 11 good to 1 bad.  I would gain 22 on the good and lose 11 on the bad for an overall pickup of 11.  Add to this the chances of down two and the reward increases while risk decreases.

 

So, at team imps I'd want to be about 91% certain of a 1 trick set with realistic prospects of a 2-trick set before I'd pull the trigger - and even then I'd be nervous.

 

Winston

 

Winston, I think that you have this all backwards :ph34r:. In a short match with VP-scoring it is more appealing to double the opponents. In a long match the percentages are as you say.

 

Consider a 6-board match with 5 flat boards. If you double and they make it, you lose by 11 IMPs, which is some number of VPs. If you double and they go down, you win by 2 or 3 IMPs, which is a smaller number of VPs, but not proportionally smaller (say -4 instead of -10). So at the 6-1 rate that you mention, the doubler is a huge winner at 30-VP's 6-board matches with no other swings (after 7 matches you'd be at +14 instead of the 0 you'd get by passing).

 

In very long matches however, it is only the IMPs that count and the 6-1 ratio is exactly right. So that means that you optimize your pay-off by making all doubles that appear to succeed more than 6 out of 7 times. There might be psychological reasons for making only 11-1 doubles, but no logical reasons.

 

Note that this discussion is a bit flawed because it assumes that the opponents are in the same spot undoubled. However, the lesson is correct, in swiss tournaments with short matches and VP-scoring the optimal strategy is more aggressive than in long knock-out matches. (of course, in extremely short knock-out matches it pays to make speculative doubles).

[Guest Jlall]

In very long matches however, it is only the IMPs that count and the 6-1 ratio is exactly right. So that means that you optimize your pay-off by making all doubles that appear to succeed more than 6 out of 7 times. There might be psychological reasons for making only 11-1 doubles, but no logical reasons.

They might make BECAUSE you doubled, and would have gone set had you not Xed. You might beat them one when they would go set 2 or 3 had you not Xed. Your teammates might get to a better spot and go plus causing your X to gain only 1 imp instead of 2 or 3. They might even go plus 800 causing your X to gain nothing. These all seem like logical arguments why 6 to 1 is not good enough odds to double them off for a 1 trick set.

 

If you told me that they would make 1 time out of 8 and 7 times out of 8 they would go set 1 trick, I would definitely not double them.

[han]

As I said, the discussion is flawed.