Re: [gnugo-devel] strategic effect

[Tristan Cazenave]

> > I am sure that GNU Go will improve a lot once readconnect works reliably.
> > OTOH I wonder a bit whether it is really useful to have the reading
> > analysis split up into one more module. How do you imagine the interaction
> > between the owl and the readconnect codes? After all, it is rather the
> > rule than an exception these two issues are interdependent. (Make an eye
> > by first threatening to cut s.th. off, connect by threatening to capture
> > cutting stones, make live with invasion stones by various attempts to
> > connect to the outside, reduce eye space while threatening to cut, etc. 
> > etc.)
> 
> The connection analysis done by readconnect is of a tactical nature.
> Just like the tactical reading tries to capture stones by reducing
> their liberties to zero, the purpose of readconnect is to try to
> connect strings by playing stones to make them actually belong to the
> same string. Both of these will be kind of stupid and shortsighted
> (tactical reading gives up if the string gets five liberties,
> readconnect gives up if it can't find a move getting closer to its
> goal).
> 
> The strategic understanding of life and death is up to the owl reading
> and the semeai reading, which use tactical reading and (in the future)
> connection reading as tools to decide which moves are worthwhile to
> try.
> 

I think it is better to say readconnect stops when the shortest path
between
the two strings is greater than a threshold (it is not completely true
but
closer to truth).

I have put up generic gradual games definitions in English at :

http://www.ai.univ-paris8.fr/~cazenave/GradualGames.h
http://www.ai.univ-paris8.fr/~cazenave/GradualGames.c

Next thing is making it work with readconnect.c
Handling of non transitivity has some links with atari_atari,
if generic things are done in atari_atari, maybe it can be
reused to detect non-transitivity.

Here are the current results of Golois on connect.tst

1 passed
2 failed: Correct '0', got '1 J17'
3 PASSED
4 failed: Correct '0', got '1 Q4'
5 passed
6 passed
7 failed: Correct '0', got '1 H8'
8 PASSED
9 passed
10 failed: Correct '1 A6', got '1 B7'
11 failed: Correct '1 B7', got '0'
12 failed: Correct '0', got '1 H8'
13 passed
14 passed
15 failed: Correct '1 B7', got '0'
16 failed: Correct '1 D5', got '1 E9'
17 passed
18 PASSED
19 failed: Correct '1 P13', got '1 O12'
20 passed
21 passed
22 failed: Correct '1 O12', got '1 Q10'
23 FAILED: Correct '1 M7', got '1 N5'
24 passed
25 failed: Correct '1 G1', got '0'
26 failed: Correct '1 F1', got '1 D3'
27 passed
28 passed
29 failed: Correct '0', got '1 P9'
30 failed: Correct '0', got '1 J11'
31 failed: Correct '1 L10', got '1 M11'
32 failed: Correct '0', got '1 Q13'
33 failed: Correct '1 F14', got '1 J10'
34 passed
35 failed: Correct '0', got '1 M14'
36 failed: Correct '0', got '1 C15'
37 failed: Correct '0', got '1 E3'
38 PASSED
39 PASSED
40 PASSED
41 passed
42 passed
43 passed
44 failed: Correct '1 H18', got '1 G17'
45 passed
46 failed: Correct '1 B4', got '1 A7'
47 failed: Correct '0', got '1 J3'
48 passed
49 passed
50 passed
51 PASSED
52 failed: Correct '1 C13', got '1 B15'
53 PASSED
54 PASSED
55 PASSED
56 failed: Correct '0', got '1 Q4'
57 failed: Correct '1 D5', got '1 C6'
58 failed: Correct '1 B12', got '1 B13'
59 passed
60 passed
61 PASSED
62 passed
63 passed
64 passed
65 PASSED
66 failed: Correct '1 K1', got '1 L3'
67 passed
68 passed
69 failed: Correct '1 (E11|F11)', got '0'
70 failed: Correct '0', got '1 O3'
71 failed: Correct '1 P4', got '1 Q3'
72 failed: Correct '1 O11', got '1 O12'
73 failed: Correct '0', got '1 N14'
74 passed
75 passed
76 PASSED
77 failed: Correct '1 J2', got '1 J3'
78 failed: Correct '0', got '1 C2'
79 passed
80 passed
81 failed: Correct '0', got '1 T10'
Summary: 44/81 passes. 13 unexpected passes, 1 unexpected failure