endgame databases and misc
Pc Solutions
>>(proving game 1800 player GIVING odds of rook to Fritz3follows, but first:)
>>The truth is, computer chess has accomplished almost NOTHING inthe last
>>20 years or so. The first incidental win of a computer againsta top
>>Grandmaster in a quick game was around 1975 (Larsen). Sincethen, all
>>efforts to score more than accidental wins in non-serious quickgames have
>>failed. The only thing that's better now is their publicity.
>>There is, until now, NO computer program that knows all therules, as
>>demonstrated by some of my latest anti-computer problems. Chessprogramming
>>however, is only 50 years old, and the basic rules of chess areVERY
>>complicated. :-) Maybe in another 500 years they know the basicrules.
>>Every A300-player can still win against computers usinganti-computer
>>techniques like in the game below.
hy...@willis.cis.uab.edu (Robert Hyatt) (BLITZ) writes:
>I'll take this wager... you pick the time, the place, and the1200
>player... you pick the time control, etc. I'll even spot your1200
>player a piece... let me know when... :^)
(My guess is you will hear no more..keep us informed)
With ref to above, and also the "Should computers use end/opening
databases" debate, read this, it should interest you..
I wrote a program called DYNAMO, which can play 10x10(International)
and 8x8 (Spanish/"POOL"/Russian) draughts - in fact any variant,since it
can set the board size flexibly from 4x4 up to 10x10. A shareware version
is widely available on many shareware CDs, etc, although I sell the full
feature version commercially. I will cover game-details later - but in
essence the game resembles "standard" draughts in the opening, although
men can capture backwards (but only move forwards), and kings can sweep
the diagonals like bishops, leaping and capturing opponents as they do so,
and even going on to make further jumps in different directions. Another
rule states that you must take the line that captures the most pieces.
The point I am getting at is that the game is not trivial to play, or
program - indeed it is probably more widely played than traditional
checkers (I say traditional - but I understand Spanish Pool 8x8 predates
all other variants). It is a very exciting and dynamic game, and there
are many players across the US, South America and Russia - with many
tournaments and clubs organising activities.
OK, so I wrote DYNAMO originally to play the 10x10 game, then the late
and great Derek Oldbury (former World Draughts Champion) mentioned
these other Draughts variants to me, so I adapted it to also play
with these rules. I picked up a 2nd hand book on the game, (Vladimir
Kaplans "Tournament Checkers" - an excellent book) and learnt
enough about it to get it to play a reasonable game - and there I had
it - the first program in the world to play American "Pool" Checkers!
Amongst my customers for this program was Valtas Hockenbury, an
enthusiast who also happened to have a Pentium-90 - so he decided to
challenge some Pool players. Amazingly, it did very well - it beat
one local player, who was regarded as a reasonable player, by 11 wins/
& 1 draw in 12 games! So Valtas cast further afield, and played some
informal telephone games against Calvin Munroe, the current US champ
- and the score was 1 win, 2 draws and 1 loss! It also managed a
couple of draws in 2 games against Vladimir Kaplan, a very highly
regarded former champion and author in the game.
Now OK, some interesting points. First, these games WERE very informal
- it is hardly the same as match conditions, when something important
is at stake, and you pull out your interesting opening 'cooks'.Even so,
analysis with the players showed that DYNAMO had played very credibly
- finding many 'only' moves in critical lines. However,
and this is the point I wanted to make - this early version of DYNAMO
was able to perform this well with an opening book of... about 100 moves!
Also - no endgame database! Now what we have here is something that
cannot be replicated in the computer chess world - a unique
experiment - because I had, at time of writing that version of
the program, NEVER played Pool checkers in my life
against a human, and had little access to books on the game,
or informed opinion on what strategies work, etc. I just tried
to figure it all out from general principles, based on what I knew
from programming english checkers, plus a little bit of knowledge
from 10x10 checkers (not very much).
Similarly on the human side, we have a population with NO
exposure to computer programs - no chance to check out my program
and suss out the weaknesses - everything, on the face of it, on
pretty even terms.
Now of course, this situation may not last - if top players
perceive my program as a threat they will start cooking up lines,
so it then becomes necessary for me to do the same. Now can you
really turn round and say I cannot use any databases in my
program - if the humans start using theirs?
The above poses the interest "what if" question - If you had a
Pentium-120 loaded with Chess Genius-3 and went back in time to
try it out on an unexposed population of Grand+masters, how well
would it do? In chess, the players have "grown up" with the computers.
With ref to the discussion at the top of this article - another
Pool player contacted me, saying that my program was poor because
it did not solve some well-known endgames. Like many endgames,
they required long intricate play, the result of much human analysis
- very difficult to compute in a few secs from general principles.
I wrote back, saying it had got good results against strong players -
and he wrote back challenging me to prove it, saying he did not
believe it could. I then contacted him, telling him exactly WHO
it had played, and challenged him to a match - he declined..
of course, he only had been using a 386sx25, which didn't help..
As a result, I have put some more work into the opening/endgame
databases, and I am investigating the possibilities of generating
exhaustive Chinook-style databases for Pool..
Here is a "fact sheet" I distribute on Pool - I have left the
blatant plug in at the end. You will find the older Shareware version
on many ShareWare CD's and FTP sites (Try Ziff-Davis's) -
and I will be uploading a new shareware demo version, with many
improvements - Windows style Windows, Picture buttons, Tape controls,
spin buttons, etc. somewhere soon - will post something on
this at a later date. Get it, and use it to play & solve the games below..
.. (from A.MILLETT, author of DYNAMO)
POOL DRAUGHTS
-------------
POOL DRAUGHTS is a game of special interest to players of the ENGLISH
game because it has the powerful kings of the International game, and
yet has the 8x8 board of the English game, and as such serves as a good
intro to 10x10 draughts for normal Checker players.
The POOL game is popular in the USA, and gets its curious name from
POLISH draughts - POOL is a corruption of this! In addition to moving
forward as in Checkers, men can also jump BACKWARDS. The man is crowned
when it reaches the top of the board, but the KING behaves in a different
fashion. Kings can sweep ANY distance along a diagonal, rather like BISHOPS
in chess. Futhermore they can leap an opponent, land on ANY space beyond,
and then make further leaps - in a different direction if necessary. The
king is NOT allowed to pass a chance to jump - For instance, if a White
king is on square A1, and black men on C3 and B6, White MUST jump A1xD4xA7
- he can not simply jump A1xH8 ignoring the man on B6.
A man is not crowned until the END of a multijump sequence -ie. a man
can leap into and then backwards out of the KING ROW without being crowned
- it is only crowned if it finally comes to rest on the king row.
To understand the nature of POOL draughts, you need to look at the
nature of the pieces. Notice how powerful the kings are in the game
- as a very rough guide, you can say that your first king is worth 3 men,
although this can vary widely, depending on how your opponent is
positioned. For instance, if both sides have a king, subsequent kings are
worth a lot less. It is also important to note that even if one side
is reduced to only 1 king, the opponent may need up to 4 kings
to trap it and force a win - although sometimes he can do it with 3,
if the lone king is on the A1..H8 diagonal, 4 are needed.
For this reason, it is generally considered important to get the
first king - then, unless you opponent is able to quickly recapture
it or break through himself, he will lose.
The board is numbered in the same way as the ENGLISH game - however the
strategy is rather different. Again, limited space prevents further analysis
of this fine game, but you will find DYNAMO a mean opponent! Here
are a number of demonstration games, showing interesting tactical traps..
(If you are a chess player - just put pawns/pieces on the 1st 12 black
squares for each side. Black on sqrs A1-H3, Black moves 1st)
1>C3-B4 D6-E5 B4-A5 B6-C5 B2-C3? A natural looking move, but a loser.
C5-D4! E3-C5 C7-B6 A5-C7 C8-B2 A1-C3 E5-F4 G3-E5 F6-B2 and White gets
a king and wins.
2>G3-F4 F6-E5 F2-G3 E7-F6 A3-B4 B6-C5? G3-H4! E5-G3 H2-F4 C5-A3 F4-G5 H6-F4
E3-C5 and Black is a piece up.
3>C3-D4 D6-E5 B2-C3 E7-D6 G3-H4 D6-C5 A1-B2? Beware of moving back men -
you may offer your opponent a king! C5-B4! A3-C5 E5-F4 E3-E7 F8-B4
C3-A5 H6-G5 H4-F6 G7-A1 and White wins.
4>C3-D4 D6-C5 B2-C3 F6-G5 G3-F4 C7-D6 D4-E5 B8-C7? A3-B4! C5-A3 E3-D4 G5-C5
C3-B4 D6-F4 B4-B8 and Black wins.
5>E3-D4, D6-E5,A3-B4,H6-G5,B4-C5,G5-F4,B2-A3? This move loses to this
tactical exchange : F6-G5!,D4xH4,B6xB2,G3xE5, E7-F6,A1xC3,F6xB2, and now
Black must give a piece to stop the king D2-C3,B2xD4 and White, a piece
up, should win.
6>C3-B4,B6-C5,8-C3,F6-E5,E3-F4,G7-F6,6-E3,H8-G7,G1-F2? This fatally
weakens Blacks back rank, allowing a break-through thus:-C5-D4!,E3xC5,
F6-G5,D2-E3,E5-D4,C3xE5,E7-F6,C5xE7,F6xD4,E3xC5,G5xG1 & Whites king wins.
7>C3-D4 D6-C5 B2-C3 E7-D6 G3-F4 F6-G5 F2-G3 G5-H4 G1-F2 F8-E7 A1-B2 G7-F6?
A fatal mistake.. A3-B4! C5-A3 F4-E5 D6-F4 E3-G5 H6-F4 G3-G7 H8-F6 D4-C5
B6-D4 C3-G7 and Black wins.
8>E3-D4 D6-E5 A3-B4 E7-D6 B2-A3 B6-C5 D4-B6 A7-C5 D2-E3 C7-B6 B4-A5?
F6-B2 C7-E5 E7-F6 A1-C3 F6-B2 and White wins. Black does have other options,
but all lead to defeat - try them against DYNAMO!
9>C3-D4 B6-A5 D4-C5 D6-B4 A3-C5 C7-B6 E3-D4 C8-C7? The evacuation of the
back square C8 is a loser here - whatever path White takes, Blacks deadly
tactical threats win the day.. G3-F4! C7-D6 B2-A3 D6-B4 A3-C5 B8-C7
C1-B2 C7-D6 B2-A3 D6-B4 A3-C5 F6-G5 D2-C3 G5-E3 C3-B4 A5-E5 F2-C8
and Black wins.
10>C3-D4 F6-E5 D4-F6 E7-G5 B2-C3 B6-C5 C3-B4 G5-H4 G3-F4 G7-F6 A1-B2 F6-G5
B2-C3 H8-G7? This loses to this really neat combination.. F4-E5! D6-F4
B4-D6 C7-E5 E3-D4 G7-F6 C3-B4 E5-A5 F2-G3 H4-F2 E1-G7 and Black wins.
When pulling a shot, you should beware that your opponent does not have
something up his sleeve in reply, or you could be a victim of the BOOMERANG
shot...
White: A3,E5,D6,B6,E7,A7,F8 Black: E1,C1,F2,D2,C3,H4,H6
F2-E3! A cunning move, that lures White into playing what he thinks is a
winning shot.. E5-F4? E3-G5 E7-F6 G5-C5 B6-B2 and White think she has a
king, and is home and dry, but.. H6-G7!! F8-H6 H4-G5 H6-F4 D2-E3 F4-D2
E1-A1 and Black has the 'move' against Whites men .. A7-B6 A1-B2 B6-C5
B2-C3 and Black wins.
Along similar lines, here is another super shot..
White: G5,C5,H6,F6,D6,G7,C7,A7,F8,C8,B8
Black: G1,E1,C1,A1,F2,E3,C3,A3,F4,B4,A5
C1-D2! Again, the bait is dangled.. C5-D4? E3-E7 G5-C1 E7-G5 H6-F4 However
Whites king is short-lived.. A5-B6!! C7-A5 C3-D4 A5-E5 E1-D2 C1-E3
F2-H8 and White is lost.
By the way, DYNAMO can see & analyze all the above shots, (usually in 1 or 2
seconds!) although he will sometimes play a different line to themain lines
given above - ie sometimes it is better to throw a piece than concede the
king. The above games and problems are taken from an excellent book
'Tournament Checkers', by Former US and USSR Champion Vladimir Kaplan,
(Written in good English I should add! Available from him at
2306 Ocean Av,Apt D2,Brooklyn,NY 11229. (718).336.3132.)
DYNAMO PRO runs on any PC-Compatible computer, (486 ver option supplied,
for DOS and WINDOWS) and costs 29 pounds (or $49 USA) and is available
from:- PC SOLUTIONS, PO BOX 954, BOURNEMOUTH, ENGLAND, BH7 6YJ
+---------------------------------------------------------------------+
| Standard disclaimer: The views of this user are strictly their own |
| ===> Sent via CC-NET BBS [centron.com] +44-1473-273246 <=== |
+---------------------------------------------------------------------+