[helene_t]

I had stratified by hearts vs spades in the 5M hands in the previous version but without reporting which was which the table became a bit useless. It was 5-card spades that was not statistically significant by the way.

Fixed now.

[helene_t]

Just thought of something:

Those who open a normal 1nt shape are compared to others who play a different nt range, or who chose to up/downgrade.

Those who open an offshape 1nt ate partially compared to others who chose not to open offshape but play the same range.

Not sure if this can explain why offshape 1nt works better than normal 1nt. Anyway, the confidence intervals are fairly wide so maybe we shouldn't read too much into it.

[awm]

helene_t, on 2025-March-11, 10:52, said:

Just thought of something:

Those who open a normal 1nt shape are compared to others who play a different nt range, or who chose to up/downgrade.

Those who open an offshape 1nt ate partially compared to others who chose not to open offshape but play the same range.

Not sure if this can explain why offshape 1nt works better than normal 1nt. Anyway, the confidence intervals are fairly wide so maybe we shouldn't read too much into it.

There is also some selection bias -- a lot of people don't open 1nt on ALL semi-balanced hands in range and you only see the ones where they chose that option (more likely to have strong doubletons, rebid problems, etc). Still, it does seem to suggest that the semi-balanced 1NT openers are a reasonable strategy.

Do you have a breakdown of rebid problem vs. no rebid problem (rebid problem is basically hands with 4♥ and 5 minor or 4♦ and 5 clubs)?

[helene_t]

clubs diamonds hearts spades N E s se lower upper
1: 5 2 2 4 963 0.41457899 5.370615 0.1730656 0.07537035 0.7537876
2: 5 4 2 2 1127 0.21734066 5.242778 0.1561707 -0.08875389 0.5234352
3: 6 2 3 2 553 0.09591486 5.799369 0.2466142 -0.38744906 0.5792788
4: 2 6 2 3 634 0.29215053 5.438111 0.2159750 -0.13116040 0.7154615
5: 6 2 2 3 578 0.56311738 5.766904 0.2398716 0.09296905 1.0332657
6: 4 5 2 2 1034 0.19552244 5.604403 0.1742886 -0.14608330 0.5371282
7: 2 5 4 2 1243 0.35905021 5.439165 0.1542754 0.05667045 0.6614300
8: 6 3 2 2 565 0.52709898 5.715363 0.2404471 0.05582261 0.9983753
9: 5 2 4 2 1180 0.44651054 4.987283 0.1451854 0.16194710 0.7310740
10: 2 6 3 2 524 0.89385179 5.605173 0.2448631 0.41392005 1.3737835
11: 2 5 2 4 1032 0.24508357 5.552211 0.1728328 -0.09366866 0.5838358
12: 3 6 2 2 538 0.71983996 5.771305 0.2488186 0.23215542 1.2075245

Including all hands (also those where multiple tables opened 1NT, but I suppose that's won't happen so often with semibalanced hands). I don't distinguish between weak and strong NT here but even then the confidence intervals are too wide for comparison, I would say.

We do see that rebid problems correlate with prevalence. 4225 and hands with a 6-card minor are the least common, while it is 4H5m and 5c4d hands that are the most prevalent.