I think this might explain it: twin fractions are varying through the course of the runs due to variability of crystal illumination. So I compared first 90 deg to second, got the results below. I wonder whether this happens often in twinning cases in which the crystal is not fully immersed (my case, unfortunately (home source, big crystal)).
JPK
======================================================
#### 0-90 degrees:
Twin fraction estimates excluding operators
Twin fraction estimate from L-test: 0.23
Twin fraction estimate from moments: 0.25
Twin fraction estimates by operator
---------------------------------------------------------------------------------------
| operator | L-test | H-test | Murray | ML Britton |
---------------------------------------------------------------------------------------
| -h-k, k, -l | Yes | 0.32 | 0.3 | 0.3 ( N/A ) |
---------------------------------------------------------------------------------------
TWINNING SUMMARY
Twinning fraction from H-test: 0.32
L-statistic from L-Test: 0.40
======================================================
#### 90-180 degrees:
Twin fraction estimates excluding operators
Twin fraction estimate from L-test: 0.25
Twin fraction estimate from moments: 0.29
Twin fraction estimates by operator
---------------------------------------------------------------------------------------
| operator | L-test | H-test | Murray | ML Britton |
---------------------------------------------------------------------------------------
| -h-k, k, -l | Yes | 0.42 | 0.39 | 0.45 ( N/A ) |
---------------------------------------------------------------------------------------
TWINNING SUMMARY
Twinning fraction from H-test: 0.42
L-statistic from L-Test: 0.40
|