Distributed engine: the slave now receives stats in binary form.
[pachi/derm.git] / TODO
blobae7b58d831d6da6936e656ec0a47221881f632a4
1 Docs:
2 * Manual page - full usage documentation
3 * GTP interface documentation
5 Base:
6 * Further optimize board implementation, oprofile fun
7 * Clean up GTP interface, allow custom GTP commands for modules
8 * Implement parameter setup over GTP
9 * gogui-friendly GTP interface
11 Patterns:
12 * Heavy optimization work
13 * Analyze feasibility of various possible features
14 * Incrementally match all features from fast set
15 * Incrementally maintain probability distribution for ELO playouts
17 Playouts:
18 * Finish the ELO-play policy
19 * Split playout aspects to custom-stackable pieces
20 * Possibly replay top-rated unplayed moves from the tree
21   in the playout?
23 UCT:
24 * Support for prior hinting based by custom externally linked
25   evaluation function
26 * Support for different float types when running with huge playout counts
28 UCB1AMAF:
29 * Fix RAVE for good (still noticeably weaker than fuego?)
30 * Separate constants for prior and amaf elements?