I guess this new parameter is related to the number of contextual subsegment matches below which they are ignored in AA, and I would like to know more about it, in particular, the meaning of "contextual."
Does "contextual" mean:
1. the same statistical approach is also applied to the previous and next segments of the TM? or
2. the "context" field in the TM is taken into consideration when making guesses?
Context subsegment TM matches should be pretty accurate as CafeTran compares the current subsegment context (in the current segment) to the TM subsegment context of the hit. If the contexts match, the probability of the hit being accurate fragment match is very high. Hence the default lower number of such hits to be included in auto-assembling.