Tables for hit rates of B1/B2 bounds at: http://www.greenbank.org/sob/p-1.html
Caveats:-
1) The percentages displayed are not the chances of finding a factor with a particular B1/B2 combination.
Consider the 44-bit table (which we have sieved in its entirety for the n < 20M range). There are many duplicate factors that may be smoother than the first factor found (the only one listed in results.txt). Also there are many k,n pairs in this range for which we have no factor.
2) The 50bit+ table data is skewed by the fact that we haven't really sieved up there so most of the factors have come from P-1 factoring so they will naturally fit low B1/B2 bounds.
3) The percentages are cumulative.
I'm going to work on automating this using a database and a bunch of scripts.