Re: [gnugo-devel] safety_3_10.1

[Arend Bayer]

Dan wrote:

> I've put up a patch called safety_3_10.1 at devel.html.
>
> This patch implements a new test for when dragons are surrounded.
(...)
>
> The idea is based on my imperfect recollection of and idea of
> Bruce Wilcox.
>
> The convex hull of the neighbor dragons is computed, and a
> slightly larger area is also marked. If every friendly dragon
> that meets the hull is contained within it, the dragon is
> SURROUNDED. If every friendly dragon that meets the hull
> is contained within the enlargement, it is WEAKLY SURROUNDED.
> An examination of 9 test cases identified by myself and Arend
> as ones where GNU is prone to neglect a weak dragon shows that in
> every one of these cases the dragon is surrounded in this sense.

I like the idea, but have some doubts whether it can be made work well.
I found at least 3 kind of situations where it would currently produce
non-sense. I've appended them below. Also, I generally have some
distrust towards valuations that cannot possibly be made continuous...

I will also try to do better tuning for the escape values. There is a
lot of information from the influence computations, and according to
Gunnar, there has not been done much tuning for them. This would have
the added benefit that owl reading could get seriously improved, too.
(Escape value seems to be the least well tuned among the input for
compute_dragon_weakness().)

> I also have the idea that one may be able to make use of
> this idea to identify good defensive moves. This will involve
> caching the polygon when a weak dragon is found. Then any
> move matching an e or E pattern (as discussed in
> http://mail.gnu.org/pipermail/gnugo-devel/2002-October/003003.html)
> is a good defense provided it lies outside the polygon.

Now that part sounds extremely useful.


Btw, the weakness values for all the weak dragons in the positions you
analyzed are mostly higher than 0.7, just one is at 0.62. I think this
is encouraging, and means we should try increasing the strategic effect.
I will post on this in a separate message.

Arend



   A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . . . . . 19
18 . . . . . . . . . . . . . . . . . . . 18
17 . . . . . . . . . . . . . . . . . . . 17
16 . . . + . . . . . + . . . . . + . . . 16
15 . . . . . . . . . . . . . . . . . . . 15
14 . . . . . . . . . . . . . . . . . . . 14
13 . . . . . . . . . . . . . . . . . . . 13
12 . . . X . . . . . . . . . . . X X . . 12
11 . . . . . . . . O . . . . . . . . . . 11     WHITE has captured 0 stones
10 . . . X . . . . . O . . . . . X . . . 10     BLACK has captured 0 stones
 9 . . . . . . . . . . . . . . . . . . . 9
 8 . . . . . . . . . O . . . . . a . . . 8
 7 . . . . . . . . . . . . . . . . . . . 7
 6 . . . X . . . . . . . . . . . X . . . 6
 5 . . . . . . . . . . . . . . . . . . . 5
 4 . . X X . . . . . O . . . . . X X . . 4
 3 . . . . . . . . . . . . . . . . . . . 3
 2 . . . . . . . . . O . . . . . . . . . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

The stone at K2 means that K4 is not surrounded. If you remove it, K4
is surrounded. Now if you add a black stone at _a_, it is only weakly
surrounded (it should be _not_ surrounded according to your
explanations, but running your patch on it gives weakly surrounded).

   A B C D E F G H J K L M N O P Q R S T
19 . . . . . . . . . . . . . . . . . . . 19
18 . . . . . . . . . . . . . . . . . . . 18
17 . . . . . . . . . . . . . . . . . . . 17
16 . . . + . . . . . + . . . . . + . . . 16
15 . . . . . . . . . . . . . . . . . . . 15
14 . . . . . . . . . . . . . . . . . . . 14
13 . . . . . . . . . . . . . . . . . . . 13
12 . . . X . . . . . . . . . . . . . . . 12
11 . . . . . . . . . . . . . . . . O . . 11     WHITE has captured 0 stones
10 . . . X . . . . . + . . . . . + . . . 10     BLACK has captured 0 stones
 9 . . . . . . . . . . . . . . . . . . . 9
 8 . . . . . . . . . . . . . . . . O . . 8
 7 . . . . . . . . . . . . . . . . O . . 7
 6 . . . X . . . . . . . . . X . X . . . 6
 5 . . . . . . . . . . . . . . . . O . . 5
 4 . . X X . . . . . O . . X . . X O . . 4
 3 . . . . . . . . . . . . . . . X O . . 3
 2 . . . . . . . . . . . . . . . X X X . 2
 1 . . . . . . . . . . . . . . . . . . . 1
   A B C D E F G H J K L M N O P Q R S T

K4 is only weakly surrounded because of the white dragon at R5/R7.