Docs:
* Manual page - full usage documentation
* GTP interface documentation

Base:
* Further optimize board implementation, oprofile fun
* Clean up GTP interface, allow custom GTP commands for modules
* Implement parameter setup over GTP
* gogui-friendly GTP interface

General improvements:
* Opening book
* Killer moves
* MM local-based patterns
* Balanced local-based patterns?
* Local tree forcing
* Liberty maps
* Reverse status learning
	Run on game corpus. Start at final position, watch development
	of status of all stones. The moment the final status and expected
	status changes, analyze, especially if move choice differs. Use
	learnt status-fixing moves in simulations somehow.

Playouts:
* Split playout aspects to custom-stackable pieces
* Possibly replay top-rated unplayed moves from the tree
  in the playout?

UCB1AMAF:
* Fix RAVE for good (still noticeably weaker than fuego?)
* Separate constants for prior and amaf elements?