[Bug-gnubg] python analysis question

[Martin Janke]

Hi,

I have problems understanding some of the values found in 

gnubg.match['games'][]['game'][]['analysis'].

Comparing values in gnubg.match with the values displayed in gnubg I get:

nd-cubeful-eq = cubeful MWC for no double
dt-cubeful-eq  = cubeful MWC for double/take

So far so good, but:

nd-match-eq = dt-match-eq = cubeless equity of the position ???

Here's an example:

(No game) first game
(No game) first move
(No game) next
(p1) hint

Cube analysis
0-ply cubeless equity  -0.046438 (Money:  -0.045965)
  0.484137 0.125277 0.003109 - 0.515863 0.136185 0.006439
Cubeful equities:
1. No double            -0.063766
2. Double, pass         +1.000000  ( +1.063766)
3. Double, take         -0.457148  ( -0.393382)
Proper cube action: No double, take (27.0%)

(p1) set output  mwc on
Match evaluations will be shown as match winning chances.
(p1) hint

Cube analysis
0-ply cubeless MWC  49.81315% (Money:  -0.045965)
  0.484137 0.125277 0.003109 - 0.515863 0.136185 0.006439
Cubeful equities:
1. No double            49.74343%
2. Double, pass         54.02356%  (  4.28013%)
3. Double, take         48.16064%  ( -1.58280%)
Proper cube action: No double, take (27.0%)

(p1) >
Python 2.2.3 (#42, May 30 2003, 18:12:08) [MSC 32 bit (Intel)]
>>> m = gnubg.match(verbose = 1, statistics = 1, boards =0)
>>> m['games'][0]['game'][1]['analysis']['nd-cubeful-eq']
0.4974343478679657
>>> m['games'][0]['game'][1]['analysis']['dt-cubeful-eq']
0.48160636425018311
>>> m['games'][0]['game'][1]['analysis']['nd-match-eq']
-0.046437673270702362
>>> m['games'][0]['game'][1]['analysis']['dt-match-eq']
-0.046437673270702362

/Martin