PV depth drop + LMR re-search deeper

[jhonnold]

Bench: 3556179

This patch merges two concepts.

• When raising alpha on a PV node, but not producing a fail high, lower depth by 1
• When LMR fails by a significant margin, the research will be extended

These patches both did well at STC, but struggled LTC (both looked yellow) and thus were combined to make 1 known good patch.

ELO | 4.14 +- 2.96 (95%) SPRT | 6.0+0.06s Threads=1 Hash=8MB LLR | 2.95 (-2.94, 2.94) [0.00, 3.00] GAMES | N: 26032 W: 6558 L: 6248 D: 13226 http://chess.grantnet.us/test/33812/

ELO | 3.65 +- 2.70 (95%) SPRT | 6.0+0.06s Threads=1 Hash=8MB LLR | 2.96 (-2.94, 2.94) [0.00, 3.00] GAMES | N: 30880 W: 7657 L: 7333 D: 15890 http://chess.grantnet.us/test/33796/

ELO | 3.39 +- 2.55 (95%) SPRT | 30.0+0.30s Threads=1 Hash=64MB LLR | 2.96 (-2.94, 2.94) [0.00, 3.00] GAMES | N: 34016 W: 8274 L: 7942 D: 17800 http://chess.grantnet.us/test/33838/