Added code to MoveCollection and Playmaker to defer or skip bad captures.
Added a SEE testsuite to verify correctness.

…n. (Good, bad, equal)

Experimented with many different ways of using SEE (and BLIND) to order moves better and do some pruning and reductions.
The committed version is 0.5.8c which uses SEE to identify "bad" captures. Bad captures are only played after killers and not at all in QSearch.
Using SEE for reductions and futility pruning didn't work out. Comparing static eval on non quiet positions with alpha or beta for futility and reverse futility pruning and reductions also didn't help in actual play even if the results on wac.epd were promising. Quieting the positions first and including the dynamic component is too expensive. So the overall benefit of adding SEE over 0.5.7 (master-branch) was only +15 to +30 ELO in 5s + 0.5s TC against Zahka300, AbsoluteZero and Bit-Genie_5. Not "minimal" enough for the little gain as it stands now.

MinimalChess 0.5.8g vs zahak300: 1412 - 1405 - 1183  [0.501] 4000 which is 17 ELO over 0.5.8c

…ave raised alpha) and false negatives (missed futiles) and optimized the futility pruning. 0.5.8h using futility pruning with a new Eval.Estimate(move) and depth <= 3. Over 50 ELO gain against BitGenie5 compared to 0.5.7c!

Score of MinimalChess 0.5.8h vs Bit-Genie_5: 1801 - 1301 - 898  [0.563] 4000

added a faster variant for GetLeastValuableAttacker that does find queens during the rook/bishop scan.
added PST-value based SEE variants.