[EricK]

Given an unobstructed auction, how often do top pairs reach the correct contract?

By "unobstructed" I do not mean that they would know beforehand that the auction is going to be unobstructed. I am just asking you to consider those auctions which turn out to be unobstructed.

By "correct contract" I mean the contract they would choose if they could see both hands bearing in mind the state of the match etc.

I doubt anybody has exact figures for this, but what are your gut feelings?

[pclayton]

Pretty often.

Even in hands that are frequently set-up, like Challenge the Champs in the BW, the contestants frequently score 75% and higher.

[kenberg]

Here is a way you could get to a variation (but not an extreme variation) on this question: Look at team matches where, at each table, there has been no interference. See how many time the two tables end at the same contract. Of course sometimes 4M and 3N are both fine contracts and you may want to, in an imp game, to call them equal. But you could determine, using a computer rather than heavy labor, how often one side lands in 4M, the other in a part score or 6M. Presumable whenever this happens, one pair would regard the bidding result as satisfactory, the other pair would not (just looking at the two hands, not the result). Similarly with other variations where one side is clearly in a substantially different contract.


I realize that I have changed you question a bit, but it seems to still be in the same spirit and it would be easy for those who have the correct software to get the right data. Of course "top pairs" needs some definition but looking at the late rounds of vugraph events would seem to be reasonable.

I agree with Phil that the answer is "quite often". But watching vugraph, it's clear the answer is also quite a distance from "always". It would also be interesting to know what sort of variations are the most common.

[jtfanclub]

EricK, on Mar 10 2007, 12:21 PM, said:

Given an unobstructed auction, how often do top pairs reach the correct contract?

By "correct contract" I mean the contract they would choose if they could see both hands bearing in mind the state of the match etc.

About 50%, I think.

I'm a strong believer in the Principle of Fast Arrival- your results will be better if you get to a good contract quickly than going for the best contract slowly. In match points, if you miss a 52% game by not trying for it, or get to a 48% game with a wild leap, that's usually well worth it just beause you gave the defense less information how to beat it. I think that on a large number of boards there are a lot of 'good' contracts, and even the experts don't try to differentiate them.

If a have mentor who's been working on this. For one simple example, in match points the auction goes:

1NT 2♥
2♠ 3NT

Are you better off in spades or in NT? Well, it turns out that there's something like 7 major and a number of minor factors that determine this. And he has a far more complex Keri-type auction to determine these factors so that yes, you know things like responder's trump strength, controls, partner's length across your doubleton, etc. etc. I think even for experts over half the time on this auction it ends up just being a coin flip. Meanwhile, my mentor's auction will get you the right answer something like 90% of the time. Unfortunately, it reveals so much to the opponents that it allows them to play it double-dummy after the opening lead. The loss from this is probably greater than 'heads I pass'.

It amazes me to look at the travelers for high level events. You see very, very few boards with the same score going all the way down. Sometimes it's overtricks, but more often it's people with different contracts. They aren't all right.

[the hog]

A rough guess, Eric? I would say that in an uncontested auction they reach the right spot about 80%+ of the time.

[Fluffy]

top pairs often do not very good at part score level.

Strong ♣ systems do 95% at slam, 85% at game level

Nat systems do 90% at game, 85% at slam and a bit better on part scores than the strong club ones.