Join Bridge Winners
Data regarding notrump evaluation

I've located some old data posted by Klaas Naaijkens on rec.games.bridge. Linking did not work, but you can google: rec.games.bridge 2NT invite klaas . You have to read down the thread a bit to makes sense of what he posted. His data is consistent with others I've seen, but I've also seen data with significantly different results. I used his data in a complex spreadsheet to simulate matchpointing and convinced myself that the potential benefit of inviting game opposite a 15-17 or other three point range was very small, the optimum inviting strategy was "invite 8.5 or 9, accept 15.5 or better" and this was little different than "pass 8.5 or less, bid game 9.0 or more." Half a point is roughly the value of an extra ten or a five card suit. Klaas only reported whole number values, so I interpolated to half-points.

A key finding for me was that one point adds about .18 or .19 to the probability of making 2NT or 3NT, when your combined values are between about 23 and 27. While an extra point in this range usually translates into an extra .3 tricks, not every trick contributes to making the contract, some represent extra overtricks or fewer undertricks. It is well known that real-world declarers tend to make 3NT more often than double-dummy, I usually figure 50% double-dummy is about 55%, or perhaps 60% for a declarer very skilled relative to the defense. But perhaps those are too conservative -- my partners who love to stick me in 22 point games seem to think so :)

I'll pull some specific results out when I have more time.

31 Comments
Getting Comments... loading...
.

Bottom Home Top