[gnugo-devel] strategic effect

[Arend Bayer]

Trevor's patch that removed the strategic effect from the move valuation
caused a remarkable number of PASSes in the regression test suites. I
looked at them in some detail. In case others are interested, below is
is a list with short explanations of problems that were uncovered.

Apart from some pseudo-PASSes (where gnugo's wrong tenuki was honored
by a #? [!XX] result line in the .tst-file), there were some common themes:

1. At the moment, the surrounding moyo size is computed as the size of the
moyo to which the origin of the dragon belongs. This is wrong, as a
dragon can belong to more than one moyo (e.g. via a diagonal connection).
This can cause a wrong status evaluation. (arend:2, dniwog:7, global:22)

2. Sometimes the strategic effect overlaps with other effects. E.g. a move
gets both a owl-followup value as it threatens to attack a dragon, and a
strategical attack value for the same opponent's dragon. Has happened
with owl-followup and defend_both.

3. Sometimes a move should get a strategic bonus and does not, and so
an inferior alternative that (correctly) gets the strategic bonus is
preferred. Happens with combination attacks (strategy:17); owl-attacks
that capture a dragon in between two friendly dragons do not get a
strategic connect bonus. (global:22, ego:8)

4. There remain a few where the strategic bonus is simply too high.
(Maybe strategy:1,strategy2:65, viking:2). Capping the size of the
strategic effect might help here. OTOH, after seeing these results
in detail I do no longer believe that this capping is very important;
it is probably less urgent than trying to find an educated guess of
how much a strategic attack/defense hurts/helps the respective dragon.
Also, the guess about how safe a dragon is should be improved. (E. g.
dragons with a 5 pt eye space in a row need less strategic defense than
a dragon with 20 pts moyo but no 100% eye space. Doing this intelligently
should automatically turn off annoying strategic effects in the endgame.)


My next e-mail contains a patch to fix the first problem.


-Arend
  


Unexpected PASSes:
arend:2 PASSED  -- 2 moyos neighbouring one dragon
buzco:8 PASSED  -- pseudo-PASS
dniwog:7 PASSED -- 2 moyos neighbouring one dragon
ego:4 PASSED -- difficult to solve this one
ego:8 PASSED -- correct move does not get as much strategic effect as
it deserves; it seems connecting yields more strategic effect than connecting
by capturing the lonely G11 stone in between
global:22 PASSED -- various effects: again, big neighbouring moyo is not
counted towards moyo size as origin is not part of it; OTOH dragon can be cut
at G11; finally F2 (capturing E2) should give the same huge strategic
effect as L18 does
neurogo:6 PASSED -- tenuki not so horrible; I only dislike the strategic
effect of P4 on Q4; to me it seems that this might already be counted in
the followup value
nicklas1:502 PASSED -- dragon B9 deserves better status than weakly alive
nicklas2:2102 PASSED -- minor problem;
nicklas5:1202 PASSED -- pseudo-PASS
strategy2:53 PASSED -- owl-failure uncovered
strategy2:65 PASSED -- could be solved by capping strategic effect
strategy4:207 PASSED -- pseudo-PASS
strategy:1 PASSED -- this is mainly an owl-failure (GNU Go claims that
Q15 owl-attacks Q14), but the 8.4 pts strategic effect on O16 may indeed
be a bit "voodoo".
strategy:17 PASSED -- combination attack forgets strategical effect for
adjacent worm
strategy:18 PASSED -- main problem is territorial overvaluation of E5
trevor:160 PASSED -- ?
trevor:310 PASSED -- ?
trevor:410 PASSED -- B4 effect on B2 seems to be counted at least twice
(territory(?) + defend both + strategic effect)
viking:2 PASSED -- indeed 14.85 pts strategic effect Q8 may be too much.
Capping would help here.

Index: engine/move_reasons.c
===================================================================
RCS file: /cvsroot/gnugo/gnugo/engine/move_reasons.c,v
retrieving revision 1.39
diff -u -b -B -r1.39 move_reasons.c
--- engine/move_reasons.c       2001/11/17 19:47:56     1.39
+++ engine/move_reasons.c       2001/11/20 19:19:07
@@ -3334,9 +3334,12 @@
       continue;
     }
     
-    TRACE("  %1m: %f - strategic effect on %1m\n",
-         pos, dragon_value[k], dragons[k]);
-    tot_value += dragon_value[k];
+    /***
+    *tm - removed this "voodoo" effect 3.1.15
+    *TRACE("  %1m: %f - strategic effect on %1m\n",
+    *    pos, dragon_value[k], dragons[k]);
+    *tot_value += dragon_value[k];
+    */
   }
 
   move[pos].strategical_value = tot_value;