|
Post by Steve Draper on May 9, 2014 14:15:35 GMT -8
Has any consideration been given to changing (would have to be a version revision like GDL 2, since it impacts semantics) from fixed time per move, to fixed time per role per game (i.e. - the chess clock model)? This would be interesting for several reasons:
1) Latency spikes would be less problematic, since small over-runs on any one move won't have a large impact. Hence less need to have large safety buffers in players for when they return their moves, and less GC-event-type timeouts happening
2) It is another interesting dimension for the players to consider. How long should I take over this move? Should I try to time-pressure my opponent by playing quickly and forcing most of the elapsed time onto his clock?
|
|
wat
New Member
Posts: 32
|
Post by wat on Jul 30, 2014 17:55:15 GMT -8
Has any consideration been given to changing (would have to be a version revision like GDL 2, since it impacts semantics) from fixed time per move, to fixed time per role per game (i.e. - the chess clock model)? This would be interesting for several reasons: 1) Latency spikes would be less problematic, since small over-runs on any one move won't have a large impact. Hence less need to have large safety buffers in players for when they return their moves, and less GC-event-type timeouts happening 2) It is another interesting dimension for the players to consider. How long should I take over this move? Should I try to time-pressure my opponent by playing quickly and forcing most of the elapsed time onto his clock? That would be Temporal GGP: arrogant.stanford.edu/ggp/chapters/chapter_20.html
|
|
|
Post by Andrew Rose on Aug 13, 2014 10:48:18 GMT -8
Sounds like a good addition for the next version of the protocol.
|
|