Add README.md on ./board

board/README.md

@@ -0,0 +1,16 @@
+Board library
+=================
+
+Default policy
+--------------
+
+In DarkForest, there are 3 default policies that could be chosen, specified by `--playout_policy`:
+
+1. `v2` (default)   
+DarkForest uses its own trained default policy. The training set is Tygem (Thanks Ling Wang for providing this). `./models/playout-model.bin` is the trained model. Check `patternv2.c` for the source code.
+
+2. `simple`  
+DarkForest uses simple default policy (trying to save when atari, trying to put others into atari, 3x3 pattern matching from Pachi, etc).
+
+3. `pachi`   
+DarkForest uses Pachi's default policy. The majority of codes are in `./pachi_tactics`.