From 121b24de5b6437c2cacc6b89e1d4ea32745d19f5 Mon Sep 17 00:00:00 2001 From: Petr Baudis Date: Sun, 12 Aug 2012 14:24:52 +0200 Subject: [PATCH] UCT max_maintime_ratio: Raise back to 2.0 Based on empirical observation during a 9x9 tournament. 2.0 is still dramatic decrease from the previous default 3.0 and should avoid most troublesome situations, but my gut feeling is that Pachi is just playing way too fast now. It's not based on hard data, but neither was the reduction to 1.5, IIRC. --- uct/uct.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/uct/uct.c b/uct/uct.c index fcd38c1..3eac436 100644 --- a/uct/uct.c +++ b/uct/uct.c @@ -588,7 +588,8 @@ uct_state_init(char *arg, struct board *b) // TODO: Further tuning and experiments with better time allocation schemes. u->best2_ratio = 2.5; // Higher values of max_maintime_ratio sometimes cause severe time trouble in tournaments - u->max_maintime_ratio = 1.5; + // It might be necessary to reduce it to 1.5 on large board, but more tuning is needed. + u->max_maintime_ratio = 2.0; u->val_scale = 0; u->val_points = 40; u->dynkomi_interval = 1000; -- 2.11.4.GIT