Chess System Tal II & Engine Style

[mclane]

yes. there was a method to make CSTAL more stupid for beginners.
"dumb down" if you want.

[fern]

Well, T, you had a hand in CSTAL. You should be the man to develope it furthermore, alone or associated with some other guy.
As Greeks used to say.

Hic Rodas, Hic Salta....


Fernando

[Patatov]

Do you know if we can connect CST to FICS or any other free server? I entered the IP address but it doesn't work... I didn't see a setting where you put your ID/password... But is what I'm trying to do possible?
I'm just trying to evaluate the strength of the program with other known engines using our current hardware.

Thanks

[mclane]

yes this was possible.

it should still work today.
there could be that the operating system has something against you doing it.
e.g. CSTAL was designed to run in win95/98.

[mclane]

he left me alone and retired. his last and famous words were: you signed that paper!

[cornetmike]

Caution!!!

http://www.aifactory.co.uk/AIF_Games_Chess_Tal.htm

This version of CSTal II has NO opening book of any kind. I know. I bought this one.

CSTal II on "10 Pro Board Games" and "Tournament Chess" both have some kind of opening book, imo fairly large and adequate. Both of these are available at www.amazon.com for very cheap. I own both of these, too.

regards
Mikey

p.s. Will, the Thai seller of CSTal couldn't find his ass with both hands. Of course the version he's selling has all 50 levels. He/she obviously doesn't know the product.

[Eraserheads]

Unfortunately if you let CSTAL play against today's beasts, hoping see witness the famous CSTAL style, you'd only see a sad spectacle. CSTAL would get so badly beaten that she would be unrecognizable. Better let her play vs old or weaker engines.

[wbseries]

p.s. Will, the Thai seller of CSTal couldn't find his ass with both hands. Of course the version he's selling has all 50 levels. He/she obviously doesn't know the product.

I found that angelfire Thai dealer page on the back pages of a search result. Reading that page description and knowing that something had been mentioned on the Oxford Softworks web page about a 2.02 version. I took the 2.02 version to mean that 2.02 was some sort of patch that had been released. I was obviously wrong and I have been corrected, and I was posting the question to be corrected if I was wrong in my thinking about it.

I was only using the angelfire page as an example for the question I had (considering their page description of the program). If I were to seek an Oxford Softworks version of the program, I would want to insure I was getting a legitimate disc and not a burned copy.

[cornetmike]

Unfortunately if you let CSTAL play against today's beasts, hoping see witness the famous CSTAL style, you'd only see a sad spectacle. CSTAL would get so badly beaten that she would be unrecognizable. Better let her play vs old or weaker engines.

Ahh...but CSTal was designed to play HUMANS, not other chess engines! I experience CSTal's famous style all the time when it beats me!

regards
Mikey

[Eraserheads]

Hi all,

Hope this will make you drool more for the vintage CSTAL!

======================================

Chess System Tal
Chris Whittington

CSTal is designed to play in the romantic and dangerous style of Michael Tal, famous for his daring and aggressive style of play.

Programmer Chris Whittington has developed a radically different approach to chess programming, concentrating on speculative chess knowledge within the evaluation function; and the use of forward pruning techniques which rely on this evaluation function knowledge.

One effect of using a high knowledge-based approach is that CSTal operates at a nodes per second rate much less than programs with simple evaluation functions. The risks and benefits of this strategy are obvious; on the one side CSTal is able to steer games towards tactical king-attack complexities, and to execute stunning sacrifices. On the other side the disparity in effective search depth means that state of the art search programs will have the advantage if the position does not contain factors where CSTal's knowledge is able to give it the edge.

CSTal's computer-computer games are often very exciting and double-edged, with the result in doubt until the end. It is capable of causing serious upsets to top programs, but also of being seriously upset itself.

In a materialistic world, in the materialistic world of computer chess, Chess System Tal offers the alternative pathway of idealism.


=-============================


Subject: Advice from a CSTal customer to CSTal customers
From: Dirk Frickenschmidt
E-mail: PROTECTED@jimknopf.wupper.de
Message Number: 27266
Date: September 20, 1998 at 07:08:26

When I read the present CSTal debate I get the impression that as customer of
the program you will probably need some humor not to be irritated.
So here some advice to other customers:

1. Enjoy your program.

It contains some really interesting chess programming ideas and is well able to
play beautiful chess.
Although it may be somewhat behind in overall playing strength compared to the
top programs, it is well able to beat any of them in single games.
And its unique playing style is worth it alone.

2. Do not worry about some minor problems concerning program versions

Up to now there are three more or less public versions of the program.

- a) One on CD is available all the time. But it was updated and is no longer
regarded as the best version by the programmer since quite a while.

- b) One called 284 was offered as Paris update version on the Oxford softworks
server and was recommended by the programmer over a longer period. But Thorsten
Czub tells us that it is a "broken" version. It has disappeared from the server
now suddenly, though texts and link still tell the customers that it is
available and should be used as update.
Unfortunately you can't get this officially recommended update any longer

- c) One called 289 consists of two versions, "CSTal White" and "CSTal Black".
CSTal Black was built because CSTal had an unusually big negative score with
black against other programs, with the very speculative parameters it used (more
successfully) with white.
Now don't get mistaken from this background and from the names of the versions.
As we now learned, this does not mean that you have to use different program
versions playing with white and black (which most customers just like me would
not appreciate because it would simply be inconvenient to switch manually
between them all the time).
Instead we now learn that CSTal Black should be used for games with computers,
while CSTal White suits games with humans. Perhaps proper names for both
versions could make that clearer even to the customers.

So CSTal Black would be the choice recommended by Thorsten Czub for games
against computers. He is probably right, being the most experienced tester of
the program over many years.

But unfortunately you can't get this version either, neither by public download
nor by buying an update disk or CD.

It once was on the Oxford softworks server, but isn't any longer. Some people
received it by email, others downloaded it in the short period when it was
available, but probably most of CSTal owners don't have it and don't know they
should use it for computer games. Thorsten Czub tells us the reason for taking
it away were some from his and Chris Whittington's view inappropriate questions
concerning the inconvenient switching between CStal white and black games
supposed by one or two customer(s):
>We put 289 from the server because harald faber complained about it in rgcc.

d) One newer version, CSTal Win95 is announced to be released soon (on the
Oxford Softworks server). But Thorsten Czub tells us, that it may never be
released, again due to some from his view inappropriate critisism from some
people concerning the way CSTal versions are published.
The ongoing existence of customers of a program even after having bought it can
be really annoying! ))


3. Some customers are probably quite irritated by tone and content of Thorsten
Czubs way of promoting the program and giving advice to its customers.

In any case don't worry. He is not yet CSTal, even if it might occur like that
sometimes. He has contributed a lot to testing and developing the program,
but he is not the author and probably will not write any updates, although he
preferably says "we" and "our program" when speaking of Chris Whittington's
program. So it may be prudent to assume that neither the program nor Oxford
softworks nor the programmer are responsible for everything Thorsten says, or
agree to it in every case.

So if you for example you might read from Thorsten Czub

- a) to a customer of CSTal:
>we don't want to sell the program to idiots. Why should we. WE choose which
>people we want to adress. People like you, i don't think we want to adress.

then don't wonder about this new form of customer support.

Don't worry: you don't have to prove to Thorsten or Oxford Softworks that your
intelligence is sufficient for getting CSTal or an upgrade.
Be sure: *nobody* is able to get a reasonable public udate right now.
So don't feel too stupid even after being adressed by Thorsten.

- b) Or if you are surprised to read as customer advice:
>Its the best you delete cstal from your harddisc.
>The best for us. The best for you.

Then don't be shocked. You signed no contract when buying CSTal saying you have
to delete CSTal from your harddisk as soon as the programmer or one of his
testers doesn't like your questions concerning the program any longer. You are
allowed to leave it on the harddisk and be happy to have at least an outdated
version of the program.
This is more than nothing, is it?

Of course you will receive some heavy critisism if you use this version for
computer tournaments. Why weren't you smart enough to get it during the short
period it was available? So better play silent with your old version and don't
publish games of it anywhere, but on the other hand you shouldn't dare to
complain about not getting a new version from a public server or not even on a
buyable disk or CD: complaints would only make things worse for you and all the
other customers as well...

You have to understand one thing really now, namely that nasty complaints have
to be punished:
>as i said. we deleted it. Now you get nothing. IMO exactly what you deserve.

But as I said, don't worry about these minor drawbacks. You still have the
outdated version, and this is more than nothing for your money, is it?

- c) Also don't worry if your personal customer punishment by Thorsten Czub is
getting a bit irrational, like it happened versus Moritz. When someone asked:

>>I understood that Moritz got the better version by email but does not want to
>>test a version that is not now on the server

Thorsten answered.
>now he cannot test anything because there is NOTHING on the server anymore.
>AETSCH !

You have to understand this sentence properly: the expression "Aetsch" is a
German children's expression, meaning something like "boo, now you got it". This
may be a signal to understand the whole sentence in a proper way.

Anything else would be difficult, because if Moritz just is testing the old
server update *and* the new one (sent by Thorsten to him by email) as well, he
is obviously using both and cannot be punished by taking anything from the
server.

So the sentence
>now he cannot test anything because there is NOTHING on the server anymore.
may have some weird kind of logic, but then rather like else children use it,
claiming they cannot be seen anymore soon as they have their hands on their
eyes.

But anyway you as customer shouldn't worry, your program CSTal, however unusual,
is *not* based on this kind of logic

d) Even if you as an inappropriate CSTal customer read from Thorsten about
another inappropriate CSTal customer:
>Maybe i should better send him a nice virus. He would deserve it...

Don't worry again. Even if you deserve it (and you certainly deserve it after
any form of CSTal critisism) Thorsten will probably not send you a virus.
This means some relief, does it?

e) As a customer you should first of all memorize the main rule and always keep
it in mind:
> Pah - chess system tal was never made for the customers.

This would not only explain some things concerning the user interface.
Mainly it reminds you: in a way you are a happy man still having got it!
Because, as you learn:
>we have made cstal mainly for us.
>And than we shared it with others.
>Because we do the same in normal life,
>sharing the things we love/like together.

I am always happy remembering I was one of those who were allowed to share this
program for some more bucks than the now official price (Oxford softworks page)
of 29.99 English pound, around 50 $ or some more than 80 German Mark, with
Thorsten Czub and Chris Whittington. This idelistic way of not keeping the
program for themselves and not dwelling alone in its richness, but "sharing the
things we love/like" with us less idealistic mortals, are exactly the roots of
our post-modern, human forms of capitalism.
We don't just spend money for products we want to use. Oh no!
Next time I buy some bread and butter I will remember that this is rather a way
of "sharing beloved things together"

***

What can we learn from all this?

1. As I said above: enjoy CSTal in whatever version you have.

2. Don't take it too serious how you are treated as CStal customer. Not taking
it too serious is the only way of not feeling disappointed or even offended.

3. Enjoy computer chess: fine with CSTal, if unavailable without it...
There are still two or three other interesting programs. Or even more?

4. Even after severe punishment threats there suddenly might occur a new
version, either a server update or a new selling version sharing more of the
"things they love to share with others" at a fair prize. And even the critics
might get it. You never know, because they even happen to surprise themselves
sometimes.
So last not least finally here we got something sympathetic, do we?

Kind regards
from Dirk


================================================

CHESS SYSTEM TAL Chris Whittington

Chess System Tal II is a knowledge based program. It packs many chess ideas into its evaluation function. It helps to give it some time to work up its ideas. The longer the time control the better.

CSTal II is not designed to play fast chess against other programs. It is designed to play chess against other chess players (eg humans). You can watch CSTal play, see its ideas, observe the positional and sacrifical algorithms and learn new concepts from it. Playing endless automated games against other materialistic programs is not recommended for the sake of your own sanity and the development of your own brain. Please get a tank of fish instead.

The opening book is on the original release CD. If you don't want to copy 520 Mbytes onto your hard drive, then please make sure the CD is in the drive, and the opening book dialog is set up to point at it (this is done by default, but is worth checking on start-up). If the CD is not in place, CSTal will then use a much smaller gambit book, but this is less desirable.

End game table bases are automatically installed if you chose the 'full install' option. All the 4-man bases are supplied, and you can download any or all of the 5-man table bases from the Internet. If you are short of space or download time, then we recommend KRPKR and KQPKQ as the two most useful 5-man table bases.


==============================================

Thorsten, I just had to post your wonderful articles here:



Complete Chess System 2 - TAL
=============================

Classical paradigm
==================
When should we expect a major breakthrough in science ?
When will a lone developer 'step through the looking-glass' ?
Who will this developer be ?

The answer to the above two questions is of course whenever the old, classical programmers say 'we've reached perfection, there is no way to improve'; when the old paradigm says 'there is only one way'; when all the developers produce roughly equal results.

This is the situation we have today with chess programs. The classical paradigm is represented by Fritz3: fast and simple evaluation, pre-processing of the position before the search; and all strength, all hopes, in the search - nodes per second and search efficiency are the buzzwords.

For a classical program, to keep the search fast, the evaluation at each node must, of necessity, be brief. This evaluation is usually no more than a weighting given for each piece on each square (for example a knight might be worth 3.3 pawns on centre squares and 2.9 pawns on edge squares) and evaluation of the pawn structure for doubled pawns, passed pawns etc..
The classical pre-processing function looks for themes in the position and adjusts the square weightings accordingly - for example, if a knight is attacking a square next to the king, then increase the weighting for all the squares that the queen could cooperate with the knight in making a king attack, increase the knight weighting to keep it on the original square, increase other cooperating piece weightings and so on.

There is no doubt that this approach works but it cannot be the way forward. Preprocess- ing knowledge becomes more stupid with increasing search depth, as positions deep in the search tree becomes more removed from the assumptions of the original position,
the square weighting adjustments become more irrelevant (why weight the squares for the queen after the cooperating knight has been removed from the board ?- but the classical paradigm doesn't understand that !).

I call this type of search Artificial Stupidity (AS). Since all the current programs operate in this way, ELO grading lists and inter-program tournaments are no more than a reflection of the partially-sighted playing the blind, whose AS algorithm is most efficient, but it is not
chess.

They don't even know that they don't know
=========================================
Classic programs have static knowledge only, dynamic knowledge is beyond
the fast and simple evaluation function.

Statics:
- Material
- Structure
- Chronic weaknesses
- and more

Dynamics:
- Lead in development
- More active piece placement
- A specific and cooperative concentration of pieces
in a certain sector of the board.
- and more

Static features tend to be stable, they remain with time. Dynamic features can be dissipated with time. Static features are easy to calculate, classical programs include them. Dynamic features are difficult to calculate, they rely on interaction between the pieces, 'looking-glass' programs will begin to include them. And it is the lack
of the difficult dynamic feature calculation that marks the classical programs with so many bad games and bad moves - the types of games that allow GM's to laugh at chess pro- grams.

As GM John Nunn says 'the top programs occasionally win games against grandmasters, but they habitually lose games against ordinary club players, often making the most appalling anti-positional moves in the process.' What else does can he expect ? The old classical program finds a 24 move deep check thread, gets to the end of the thread, finds
it is not yet mate, and all it can do is add up the material, evaluate the pawn structure and return a score that shows absolutely no concept of the position !

To play chess without knowledge of chess is not to play chess, strong players will always beat such programs with superior knowledge.
The classical program play chess as if it were the First World War in the trenches, no concept of mobility, no concept of cooperation of forces, no concept of knocking the enemy off balance with well timed blows; just material and pawn structure - if it plays boring chess, that's why - if it blunders against club players, that's why. It understands nothing of consequence.

The philosophers of classical search claim that search finds everything and knows everything - they give as an example the knight fork: Without search the program knows that it is good to capture the queen with the knight. With three ply the search knows that it is good to knight fork the king and the queen. With five ply the search knows it is good to
play the knight to a position where it can threaten a fork and so on.
But the point must surely be that the search only has this knowledge within the tree. At the leaf nodes it has no such knowledge. An intelligent program can calculate as part of its evaluation function whether a knight fork is available; thus the intelligent program has
this knowledge distributed evenly over the entire search tree. In this way intelligence can replace search.
It is important here to distinguish between combinational knowledge and dynamic knowledge. In our example of the knight fork above, the classical program only has this 'knowledge' if the situation arises in tactics - the classical program only generates this knowledge as part of a combination to win the queen. If this win of the queen does not emerge from the search, then the knowledge does not exist !
The situation is perhaps clearer (and more serious) in the case of a king attack. If the classical program can find mate or win of material by some line attacking the king, in such case it has knowledge of the king attack; but if, at the search horizon, it has a strong attack, but not yet any material won, or king mated, it does not know this is a good line !
The 'looking-glass' program can calculate the attack strength FROM ITS EVALUATION FUNCTION. So, without actually finding mate or material win, the looking-glass program has the dynamic knowledge of the attack.
The classical program has combinational knowledge only by resolution of material within the search horizon. The looking-glass program has dynamic knowledge from its evaluation function. The looking-glass program is a planner, the classical program is a finder. The looking-glass program is pro-active, it makes plans to exploit the position; the classical program is re-active, it waits for a mistake by its opponent and then exploits it.

Dynamic knowledge v. Combinational knowledge
============================================

Oxford Softworks CCS2-v9.0
White: CCS2 486/33
Black: Genius2 486/33
Venue: 1 minute per move
Comment: 1-0

1. e4 e6
2. d4 d5 1
3. Nc3 Nf6 3
4. Bg5 Be7 5
5. e5 Nfd7 8
6. h4 Bxg5
7. hxg5 { CCS2's opening book ends }
.... Qxg5
8. Nf3 Qd8 { Genius2's opening book ends }
9. Bd3 h6
10. Qd2 { CCS2's dynamic knowledge - preventing O-O because
of the threat of Rxh6 }
.... c5
11. Nb5 O-O { Catastrophic - any reasonable club player can
see this move is a disaster, but Genius2 has no
dynamic knowledge, there is no immediate mate so Genius2
thinks all is ok ! }
12. Rxh6 { CCS2 needs only a few seconds thought to find this move }

bR bN bB bQ -- bR bK --
bP bP -- bN -- bP bP --
-- -- -- -- bP -- -- wR
-- wN bP bP wP -- -- --
-- -- -- wP -- -- -- --
-- -- -- wB -- wN -- --
wP wP wP wQ -- wP wP --
wR -- -- -- wK -- -- --

.... a6 { Incredibly, Genius2 thinks the position is even ! }
13. Bh7+ Kh8
14. Rh5 axb5 { Genius2 still thinks this game is drawn ! }
15. Ke2 { CCS2 finds the killer move .... }
.... Nf6 { Genius2 begins to see the trouble now ... }
16. exf6 Qxf6
17. Rah1 g6
18. Bxg6+ Kg8
19. Rh8+ Qxh8
20. Rxh8+ Kg7
21. Rh7+ Kxg6
22. Qh6+ Kf5 { and mate in 2 more moves. Genius2, the classical
program, soundly defeated by dynamic knowledge.
CCS2 didn't know its attack would win material or
deliver mate, it just knew, dynamically, the the
attack was strong and worth the sacrifice of material. }

This game clearly shows the development and strength of the 'looking-glass'
paradigm. Genius2, a classical program, seemed to have no idea of what
was going on. CCS2 had dynamic knowledge of the strength of its attack from
move 12 on, CCS2 knew from its evaluation function; Genius2 only began to
see the trouble on move 15, seven half-moves later, Genius2's knowledge
was combinational, only 'known' when the search found it.

Who will be the developer ?
===========================
To answer our third question - 'who will be the developer ?', it is necessary to look at the personality of the classical programmers and their hangers-on. These programmers are characterised by a failure to show their emotions (do they ever smile), fear (just watch them operating at tournaments), refusal to discuss how their programs work (just try talking to them) , aversion to taking risks. It has always surprised me that the 'top' programmers are not good chess players. The hangers-on only make a little money, they jealously support their chosen proteges, and viciously attack their opponents. The hangers-on know little, pretend to know much and are governed by fear and greed.
Overall the impression is of a static, non-risk taking, hostile, World War I environment. The new paradigm will come from an unexpected quarter. From a developer with extrovert personality, accustomed to taking risks, a developer with chess knowledge, probably someone unpopular with the classical paradigm supporters, certainly unpopular
with the hangers-on and computer chess entourage. This developer will have been and certainly will be furiously attacked by the classicists.


Search - the lazy programmer's way to avoid evaluating a position.
==================================================================
The new paradigm differs from the classical by one simple conceptual switch.
The classical paradigm makes fast and simple evaluation at each node and generates intelligence from the search tree. The classical programmer looks for ways to make his search more efficient and his evaluation function simpler and faster. The 'looking-glass' paradigm makes slow and complex evaluations at each node and prefers to prune the search tree by use of this evaluation function. In this model search is to be avoided
unless absolutely necessary. Thus the search tree is not central to the new paradigm, rather the search tree is used to find details overlooked, or mistakes made, by the evaluation function. The 'looking-glass' paradigm has the components of human thought - detailed, intuitive evaluation, with search carried out to ensure that the program is not
falling into any traps. I estimate that the difference in nodes per second between and extreme classical program and a 'looking-glass' program will be of the order of 20-30 times, sufficient to give the classical program an extra two plies of search (albeit with reduced knowledge at the nodes). Thus the increased knowledge of the 'looking-glass' program has to compensate for this apparently reduced search depth. The looking-glass strategy necessitates much programming effort, and requires the programmer to have an exceptionally good knowledge of chess strategy and tactics. When such a program is first
being developed it will constantly be outplayed by classical programs, for classical programs see everything within their horizon and the newly developing 'looking-glass' program cannot yet hope to know sufficient tactical and positional themes to compete, but our experience shows that once breakthrough (a knowledge o f sufficient chess themes to compensate for reduced search depth) occurs the looking-glass program begins to
consistently outplay the classical programs. Further advantages emerge from the high level of chess knowledge in the evaluation function - better move selection and move sorting, resulting in more efficient search - more possibilities of accurate forward pruning, resulting in smaller search trees. With increases in tree size (from faster hardware), these advantages are geometric.

B-Search or A-B-Search? - NO! Evaluation based or search based!
===============================================================
The classicists maintain the computer chess dichotomy of B-search (which I understand means pruning occurs at all levels of the tree) or A-B Search (which apparently means that part of the search is full width).
The looking-glass programmer condemns this dichotomy as meaningless.
The new paradigm makes the issue clear: chess programs either have simple evaluation and generate intelligence through search, or have complex evaluations and use limited search as a backup to cover oversights and mistakes. All chess programs prune in one way or another, but looking-glass programs, with complex evaluation, are able to prune more.

Of course, the issue is not so black and white. There is a grey scale between the extreme looking-glass (human play style) and extreme classical style. At the classical end of the scale the B or A-B dichotomy tries to position the program on the scale, but basically classicists believe in search. At the looking-glass end of the scale the issue is how much does the evaluation function allow us to prune or extend - how many risks can we take based on our evaluation function ? Basically looking-glass programmers believe in evaluation.

Von Manstein
============
If, as is said, chess is war, then there must be lessons to be learnt from military history. I have already alluded to the static, boring First World War style of the classical programs (and their programmers !). The opposite style can be found in several histories, Rommel in North Africa, Alexander the Great against Darius, Von Manstein in Russia. Alexander, despite being outnumbered many times, concentrated the powerful mobile part of his army, attacked the stronger Persians, cut through and went straight for Darius himself. The bulk of Darius's army was not engaged, but the battle was decisively won - a classic king attack. Von Manstein (and Rommel) both understood that the power of the outnumbered German army lay in superior staff work, concentration of forces, striking blows to knock the enemy off balance. The looking-glass chess program must contain knowledge of these dynamic elements; and it is only the looking-glass program that has the knowledge and evaluation time available to calculate such ephemerals.

Tal function
============
To find a chess player who understood the king attack, the concentration of forces, the striking of blows to unbalance the opponent, one need look no further than Michael Tal, Russian grandmaster, and player of such romantic and swashbuckling style that his games continue to thrill all lovers of chess. For the developers of the Complete Chess System 2 it was an emotional, and unexpected, experience to find their program playing, sacrificing, in the style of Tal. Opposing programs, well respected, began to fall like dominoes, they appeared to have absolutely no understanding of CCS2's style. We
were almost able to guarantee exciting games against all our opponents.

We believe that the progress we have made with our program, the looking-glass algorithm which we have developed gives us the justification to call our program the Complete Chess System 2 - TAL.




This is G o o g l e's cache of http://www.thorstenczub.de/aegon96.html as retrieved on 7 Jul 2007 20:59:19 GMT.
G o o g l e's cache is the snapshot that we took of the page as we crawled the web.
The page may have changed since that time. Click here for the current page without highlighting.
This cached page may reference images which are no longer available. Click here for the cached text only.
To link to or bookmark this page, use the following url: http://www.google.com/search?q=cache:bO ... =firefox-a

Google is neither affiliated with the authors of this page nor responsible for its content.
These search terms have been highlighted: cstal style
Aegon 96

Like every year, the Aegon-tounament is THE HIGHLIGHT for testing computer-
programs. Brilliant organized from Cock de Gorter, it is a pleasure for all
participants and tourists.
One can see and speak with all the programmers, comment on the games, discuss.
The programmers let their programs play against each other, long time before
the rounds, just for fun. There is an exchange of ideas between most of the
amateur-programmers, and sometimes a professional gives a hint: how to do
this and that.
This year every program got a fast Hewlett Packard Pentium 166Mhz with
16MB Ram.
So - everybody has this minimum hardware. Because programs are not playing
against each other, every has the allowance to use a faster, or his own
machine instead the given.
Of course the professionals have done this. Nimzo run on a 180Mhz Pentium,
Rebel maybe on a Pentium-Pro. Of course Quest used a faster machine and and and.

Computer-chess fans can meet the great programmers behind their products.
See the shy and ascetic Frans Morsch, a kind of his program, totally different
appearance than the advertising of Fritz. It is well known that the one reason
why he is playing with quest, and not with "FRITZ" is: he identifies with
totally different values. Could anyone ever call chess-base and Fritz as
modest and ascetic ? So Frans Morsch is, and therefore he is playing with
Quest.
See Ed Schr"der operating Rebel, always open for a talk, discuss with Dieter
Steinwender, publisher of Computer-Schach & Spiele about censorship in the
one and only monopol-magazin in germany. Dieter is one of the human players
again, and tries to fight his so called Karate-chess against the computers.
It was a surprise to see the new Uniacke-family. All the years before I have
seen Mark with his father beside. And suddenly, I see him with a nice son,
a nice girl, as a father himself. Of course the young boy (3 ages old?!)
played chess against his famous programmer-father. "A young capablanca" said
Mark, when I asked him about his son. "He beated Mephisto III on Level 0".
Unbelievable. So the future is planned so far. The son will maybe learn
from his father and develop later the Hiarcs-family. I hope to see the
Uniacke family in next tournaments again.

It is a pleasure to see people like Karpov, Bronstein, Seirawan and Sofia Polgar
playing chess. They don't have any "problems" with the light, the noise,
or any disease when they lose. I think Kasparov could learn how to behave
from those giants of chess. They are natural. They don't give you the feeling:
We are something better then the rest of the chess players in this room, or
even: in the world.

They are open for everything. They take it easy. Laugh. And chess gets its
best PR-Work, because they are there.

Have a talk with other VIP of the computer-chess scene, ask Jan Louwmann
about latest development on Kallisto. When there is a reason why the nether-
lands is so good in computer-chess, it has something to do with the
programmers... and Jan Louwmann. He is the pioneer who held 1986 in cologne
the hair-dryer, while Rebel on a accelator-card heated too much, able to
win the big championship, but too hot, a pity, too hot. This was the year
Hegener+Glaser and Schr"der met the first time to later release the Mephisto
Rebel 5.0 module. And Jan was there. One main reason. Ask him about all the
storys behind the scene. He will tell you.
Every night after the rounds Yasser Seirawan made blitz-games against M-Chess
Pro. And we all sat arround and discussed. Cock de Gorter is not only the
organizer of this nice tournament, he is not only the writer of several
opening-books, he is the capacity in computer-chess in the netherlands.
I don't know how late it was each day, 2h, 3h or even later ? But Yasser
was fighting and Mchess showed that is is a better blitz-player than
people say.
Meet Jeroen Noomen, ask him about Rebel, about his work with Schr"der
and his opinion about different programs. Speak with Mr.Gambit Soft, the
nice and friendly Berthold Seifriz with his family. Ask Matthias Feist,
the Fritz-programmer about the latest patch-file for Fritz4.
Keep the fingers crossed for Ananse, the program of the friendly suisse
programmer, this year followed by his nice and beauty daughter.
Talk with Chrilly Donninger, the austrian programmer of NIMZO, who will release
his latest version in a few days. This version of Nimzo will be his big
start as a professional programmer.
Ask Johann de Koning when the new ChessMaster 5000 will come on the market,
and which engine is used.
And discuss with all the other programmers about the future of computerchess.
Discuss why they all say: Mchess is not strong, it is the opening-book that
is tricky. Ask about these prejudices. They will say this about all the programs
that come higher in the list. Every programmer will learn, that there is
a kind of lobby, a kind of (what we call in germany) mafia, that controls
the market. Whenever a newcomer or any other outlaw trys to break through
these arrangements, he will be mobbed by all the other guys, parts of the mafia.
So it was with Hiarcs, never tested on fast machines in the swedish-rating
list, although he had won the championship 1993. They did not test it for
1 year. So it is now with MChess, suddenly Mchess is not strong, why ?
Because the others are afraid it could take away a part from the market.
Capitalism shows it's dark side.

For a long time now, I follow and comment on the work of Chris Whittington.
It was the ATARI-time, when I found out, that the program Chess Player
2150 for ATARI ST had something special that I wanted to take note of.
I bought every following program and some day contacted Chris.
From that time, we worked together. Kilometres of fax-paper, telephone
costs I could buy cars from, and some elo-points and bug-fixing is
maybe the result. And... (i hope) a working friendship (if you talk
with somebody from England more often than you see your girl-friend,
he will automatically be a good friend in a few years).

For 3 years Chris Whittington starts with his latest efforts on the Aegon
tounament. 3 years ago we numbered the versions with Aegon 1994.
We found out that not the aegon-version was good, but the version number
2 after the tounament. In 1995 we started with version 74. And also this version
was buggy. You always find the faults in a tournament. If there is a bug
in the program, the program will play it out in an important tournament.
We proofed this at the championship 1995 in Paderborn, where Chess System
Tal made big Tal-games, beat Genius with a nice sacrifice...

[Event "13th World Microcomputer Chess Championships"]
[Site "Paderborn (Germany)"]
[Date "Oct 11 1995"]
[Round "4"]
[White "Chess Genius"]
[Black "Chess System Tal"]
[Result "0-1"]

1. c4 e5 2. Nc3 Nf6 3. Nf3 Nc6 4. g3 d5 5. cd5 Nd5
6. Bg2 Nb6 7. 0-0 Be7 8. a3 0-0 9. b4 Re8 10. Rb1 Bf8
11. d3 a5 12. b5 Nd4 13. Nd2 Rb8 14. e3 Ne6 15. Nf3 Ng5
16. Ng5 Qg5 17. Ne4 Qd8 18. Qc2 Bg4 19. f4 ef4 20. Rf4 Be6
21. Rf2 f5 22. Nc3 Bf7 23. Ne2 Qd7 24. Qc3 Nd5 25. Bd5? Qd5
26. Nd4 Re5 27. Qc7 Ra8 28. Qc3? Rae8 29. Nf3? Re3!! 30. Be3 Re3
31. Nd4 g6 32. Re2 Bc5! 33. Re3 Bd4 34. Qc8 Kg7 35. Rbe1 f4
36. gf4 Qf3 37. Qc7 Be3 38. Re3 Qe3 39. Kg2 Qe2 40. Kg1 Qd3
41. Qb7 Qa3 42. b6 Qc5 43. Kf1 a4 44. Qc7 Qc7 45. bc7 Be6
46. h4 a3 0-1

..., and showed
different bugs

[Event "13th World Microcomputer Chess Championships"]
[Site "Paderborn (Germany)"]
[Date "Oct 15 1995"]
[Round "11"]
[White "Chess System Tal"]
[Black "XXXX"]
[Result "0-1"]

1. Nf3 d6 2. d4 g6 3. e4 Bg7 4. Bc4 e6 5. Nc3 Nc6
6. 0-0 Bd7 7. Bg5 Nge7 8. Bh4 h6 9. Rc1 g5 10. Bg3 g4
11. Nd2 Bd4 12. Qg4 Ne5 13. Qh5 N7g6 14. Ne2 Bb6 15. c3 Qg5
16. Qg5 hg5 17. Rfe1 0-0-0 18. Nd4 c5 19. N4f3 f6 20. Ne5 Ne5
21. Rcd1 Kb8 22. h3 Ka8 23. Re3 Rh7 24. Be5 de5 25. Nb3 Rdh8
26. Red3 Bc8 27. a4 a5 28. Kf1 Kb8 29. f3 Rh4 30. Bb5 f5
31. Nd2 Bc7 32. Kf2 g4 33. fg4 fg4 34. Rh1 b6 35. Kg1 Bb7
36. Bc4 R4h6 37. h4 Rh4 38. Rh4 Rh4 39. Be6 Ba6 40. c4 Bc8
41. Bc8 Kc8 42. Nf1 Bd8 43. Kf2 Rh1 44. Rd5 Kc7 45. b3 Bh4
46. g3 Rf1 47. Kf1 Bg3 48. Kg2 Bf4 49. Rd1 g3 50. Rh1 Kd7
51. Rh7 Kd6 52. Rg7 Kc6 53. Rf7 Kd6 54. Rf6 Kc7 55. Rg6 Kb7
56. Kf3 Kc7 57. Kg4 Kd7 58. Kf5 Ke7 59. Rb6 Be3 60. Rg6 Bf4
61. Rg7 Kf8 62. Kf6 Ke8 63. Ke6 Kd8 64. Kd5 Ke8 65. Kc5 Kf8
66. Rg4 Ke7 67. Kd5 Kd7 68. c5 Kc8 69. Kd6 Bd2 70. Rg8 Kb7
71. c6 Kb6 72. Rb8 Ka7 73. c7 Bb4 74. Kc6 g2 75. c8=N???? 0-1

but also many highlights in between, sacrifices against almost
any opponent

[Event "13th World Microcomputer Chess Championships"]
[Site "Paderborn (Germany)"]
[Date "Oct 8 1995"]
[Round "1"]
[White "Chess System Tal"]
[Black "Alpha I"]
[Result "1-0"]

1. d4 d5 2. c4 dc4 3. Nf3 c6 4. e3 b5 5. a4 e6
6. b3 Bb4 7. Bd2 Bd2 8. Nfd2 cb3 9. ab5 b2 10. Ra2 Ne7
11. Nc3 Nd5 12. Nd5 cd5 13. Qc2 Nd7 14. Bd3 Bb7 15. 0-0 Qb6
16. Nf3 f5 17. Qb2 0-0 18. Qa3 Rfc8 19. Rfa1 h6 20. Qe7 Nf6
21. Ne5 Rc7 22. Qb4 Ng4 23. Ng4 fg4 24. Qb1 Rc3 25. Be2 g3
26. hg3 Rcc8 27. Qg6 Rf8 28. Bd3 Rf6 29. Qh7 Kf7 30. Ra4 Bc8
31. f4 h5 32. Qh5 Ke7 33. Qh8 Bb7 34. Qg7 Rf7 35. Qh6 Kd8
36. Bg6 Rc7 37. Ra5 Rac8 38. Qf8 Kd7 39. Qf7 Kd6 40. Qf6 Rc1
41. Kf2 Ra1 42. Qe5 Kd7 43. Qg7 Kd8 44. Ra1 Qc7 45. Qf8 Kd7
46. Qf7 Kd6 47. Qf6 Qb6 48. Rh1 Qb5 49. Rh7 Qd7

[Event "13th World Microcomputer Chess Championships"]
[Site "Paderborn (Germany)"]
[Date "Oct 10 1995"]
[Round "3"]
[White "Chess System Tal"]
[Black "Hiarcs X"]
[Result "1:0 (Time)"]

1. Nf3 d5 2. g3 Nf6 3. Bg2 c6 4. 0-0 Bg4 5. Ne5 Bf5
6. d4 e6 7. Nc3 Nbd7 8. Bg5 Be7 9. Nd7 Qd7 10. Qc1 h6
11. Bf4 0-0 12. Rd1 Rad8 13. h3 Nh5 14. Be3 Bg6 15. Bd2 Rfe8
16. g4 Nf6 17. Bf4 Bd6 18. Be5 Qe7 19. Bd6 Qd6 20. Qd2 e5
21. Rac1 a5 22. de5 Qe5 23. Qd4 Qd6 24. e3 b5 25. Ne2 c5
26. Qf4 Qf4 27. Nf4 Be4 28. Nd3 Rc8 29. Be4 de4 30. Nf4 Rcd8
31. Ne2 Kh7 32. Kh2 g6 33. Nc3 b4 34. Nb5 Rd1 35. Rd1 a4
36. Nd6 Re7 37. Nc4 Kg7 38. Kg3 h5 39. g5 Nd7 40. Rd6 Nf8
41. h4 Ne6 42. Nd2 Nc7 43. Rc6 Ne6 44. a3 Kf8 45. Rc8 Re8
46. Re8 Ke8 47. Ne4 Kd7 48. Kf3 Kc6 49. c3 ba3 50. ba3 c4
51. Nd2 Kd5 52. e4 Kc5 53. Ke3 Nc7 54. f3 Nb5 55. Nb1 Kd6
56. f4 Ke6 57. Ke2 Kd7 58. Kf2 Ke7 59. Kf3 Kd6 60. Kg3 Ke6
61. Kg2 Kd7 62. Kf1 Kd6 63. Kg1 Ke6 64. Kh2 Nd6 65. e5???? Nf5
66. Kh3 Kd5 67. Nd2 Ne3 68. Kh2 Nd1 69. Nb1 Ke4 70. Nd2 Kd3
1:0 (Time)

This game is also well known because of its end-result. I was
shocked by the fact that the Hiarcs-operator did not understand
my kicks under the table, that he should push the reset button
or unplug the power-supply by mistake, because of hiarcs-time
problems. Because David Levy sat next to me, I was unable to do it
myself. I was in a total mess. I worked with both programmers,
the game was a horrible opening for Tal, a horrible middle-game
from Hiarcs, a horrible-endgame-lack of Tal and a horrible-time
problem of hiarcs [ok, the operator did not take it serious enough]
and I was unable to react. So I wispered in german (other german
whitnesses are known): Drücke den Reset-Knopf!!
but my opponent operator just friendly laughed at me. He thought
I am joking! It ruined half or more of my friendship with Mark Uniacke.
How can I ever forget this situation.


[Event "13th World Microcomputer Chess Championships"]
[Site "Paderborn (Germany)"]
[Date "Oct 12 1995"]
[Round "6"]
[White "Zeus"]
[Black "Chess System Tal"]
[Result "1/2-1/2"]

1. e4 e5 2. Nf3 Nc6 3. Bc4 Bc5 4. c3 Nf6 5. d4 ed4
6. cd4 Bb4 7. Bd2 Bd2 8. Nbd2 d5 9. ed5 Nd5 10. Qb3 Na5
11. Qa4 Nc6 12. Bb5 Bd7 13. 0-0 0-0 14. Rfe1 Nf4! 15. Re4 Qf6!
16. Ne5 Ne5! 17. de5 Nh3!! 18. gh3 Qg5 19. Kh1 Bb5 20. Qb5 Qd2
21. Re2 Qd4 22. Rc1 Qf4 23. Rc3 Rad8 24. Qb7 Rd1 25. Kg2 Qg5
26. Rg3 Qc1 27. Qa7 Rg1 28. Kf3 Qc6 29. Kf4 Qh6 30. Ke4 f5
31. Kd5 Rd8 32. Kc5 Qh4 33. f4 Qf4 34. Qc7 Qd4 35. Kc6 Rg3
36. hg3 Qc4 37. Kb6 Qe6 38. Kb5 Rc8 39. Qc8 Qc8 40. e6 Qa8
41. Rd2 Qb8 42. Kc6 Qc8 43. Kb5 Qe6 44. a3 Qe8 45. Kc4 h6
46. Rd3 g6 47. b4 Kf7 48. Rc3 Qe6 49. Kc5 Qe5 50. Kc4 Ke6
51. b5 Qe2 52. Kb4 Qb2 53. Rb3 Qd4 54. Ka5 Qd8 55. b6 Qd5
56. Rb5 Qa8 57. Kb4 Qe4 58. Kb3 Qf3 59. Kb4 Qe4 60. Kb3 Qb1
61. Ka4 Qc2 62. Kb4 Qe4 1/2-1/2

Also the above game has its story. After a marvellous attack
against a too materialistic zeus [winner of the spanish championship!]
CSTal punished me buy giving: what else can it be: endless check-repetition

[Event "13th World Microcomputer Chess Championships"]
[Site "Paderborn (Germany)"]
[Date "Oct 13 1995"]
[Round "8"]
[White "Chess System Tal"]
[Black "Junior"]
[Result "1-0"]

1. d4 Nf6 2. c4 g6 3. Nc3 d5 4. cd5 Nd5 5. e4 Nc3
6. bc3 Bg7 7. Bc4 0-0 8. Ne2 c5 9. 0-0 Nc6 10. Be3 Qc7
11. Rc1 b6 12. dc5 bc5 13. Nf4 e6 14. Nd3 Ne5 15. Bf4 Bb7
16. Qe2 Rab8 17. Rfd1 g5 18. Ne5 Be5 19. Bg5 Bh2 20. Kf1 Be5
21. Bh6 Rfd8 22. f4 Rd1 23. Qd1 Bf6 24. e5 Be5 25. Qg4 Kh8
26. fe5 Rg8 27. Qf4 Rg6 28. Rc2 Qd8 29. Rd2 Bd5 30. Bb3 f5
31. Bd5 ed5 32. Rb2 Rg4 33. Qf5 Rg6 34. Be3 d4 35. cd4 a5
36. d5 a4 37. Rb7 Qe8 1-0




All in all, it was a good effort [and we will try to play in paderborn
1996 again.
For a few years we are now working on a version that follows the Complete
Chess System. I was very disatisfied with the evaluation-function of
the Complete Chess System. And so Chris was. In many many games I played
with the Complete Chess System I came to the conclusion that it's evaluation
-function is more than strange. A program can only plan for positions, when
it evaluates them accurate. Or almost that. So we tried to change it.
That took now many years.
But Chris is a strong chess-player and invents many ideas, also he visualises
many new screens and graphs for me, to see what the program does without
the need of a debugger.

By reading and studying the book from Mikhail Tal and Iakov Damsky:
"Attack with Mikhail Tal" many new ideas for knowledge implementation
were born. This book is pubkished by Cadogan-Chess, and I can advise
anybody to read it. It is full of intersting test-positions for chess
programs. Chris was inspired by the book and so he invented the Tal-
Function and we got a program that played really unbelievable games.
The book is our Talisman, and we use Tals hypnotic-picture to freeze
the opponent.

For each evaluation-function he made me a slider in the program,
for each extension he made a switch that I can adjust and try out
anyfunction I want to. This way we made good progress. When I have seen
a pattern, why CSTal is doing this or that, Chris can follow very fast.
Then he implements a new algorithm and the bad play will never repeat.
But this has also the problem that it need much fine-tuning. The strength
of CSTal is not the search but the evluation in the tree.
So the whole program gets changed and out of balance after implementing
new evaluation-functions.

This year we started with version 165. So within 3 years from Aegon94
to aegon96 he made (and I had to test) 165 big versions (we don't
count the little changes in between!). Now you can imgaine the big work
and also this creative man some austrians called in the PC-SCHACH 3/94
page 49 (in association to Asterix and the romans):
"Is all hope gone, is the chess-playing part of the earth in the hand
of the roman-empire ? No - not the whole on - somewhere in britannia
there is this small village Oxonia, whose inhabitants still fight bitter
resitance. There chieftain Chris Whittingtonix did produce programs
with the Chess Champion 2175 and the Complete Chess System, that were
different because of their "human" playing-style. Different from the
general-pulp of the A/B-strategists, but tactically to harmless to break-
through in a significant way. But now we have heard through the oxonic
barde/singer Thorsten Czubix that now the breakthrough has reached
and that further Complete Chess System 2 is on the best way to develop
as a top-program. The message I listen carefully , but shall I believe it ?"
This was 2 years ago, before we changed the name of the program into Chess
System Tal because Chris devoloped the Tal-Function in addition to the
difficult work on evaluation-function.

Last year Chess System Tal had to play against Sofia Polgar. It was a
pleasure for me, because I was operating Tal, and I always followed
Sofia Polgars way. I was very nervous and concentrated so much on
not making faults that after a while she asks me if it is boring
for me, to operate. HA! It was not boring, but I was so much trying
not to show my nervousness, that she thought I am bored by everything.
We came into a little talk, I explained her about the reason
why this program is called Chess System Tal, I told her about the
2 different kinds of programs, and that this is like in life:
we have the materialists and the idealists. She said: oh - this opening
cannot be called a Tal-opening. But this was false. Within a few moves
CSTAl was able to develop a Tal-Position. The game went up and down,
and in the end she won - of course. But for me it was a pleasure.
I hoped to continue the little conversation, but luck was not on my side.

Therefore Cstal had to play against the nice female GM Peng Zhaoqin
(ELO 2355), who played the first time in Aegon and has not so much
computer-experience.
One thing happened before the game, Karpov (who was there the
first round) and many other players stood arround the Chess-System
Tal board (no special reason for that, but because this, I recognized it)
and showed his opening structure against the computers. In my opinion
many players tried in the following games all the same ideas, with different
openings, but the same structure. Also Peng Zhaoqin tried this structure.
The Pentium 166 was 3 times faster then my 486-DX100 I use for testing.
It was 2 times faster than a pentium90 and Chess System Tal computed
arround 3900-4600 nodes per second on this fast machine. Not much compared
to the positions other programs try out.
I think this game shows that CSTal evaluates not the real position, but more
the chance to wi the game. It plays speculative. It will not always search
for the best move in the position, often it plays the best moves for its
plan. The plans were found by evaluating things that exist only in the
phantasy of CSTal. The program is dreaming!!
Peng Zhaoqin shows that she plays accurate chess, good in tactics, without
blundering. The games were played 1h30m blitz with 20 seconds Fischer-bonus.

[Event "AEGON Mankind vs Machines"]
[Site "The Hague NED"]
[Date "1996.04.10"]
[Round "1"]
[White "CHESS SYSTEM TAL"]
[Black "Peng Zhaoqin"]
[Result "1/2-1/2"]

1. e4 e6 2. d4 d5 3. Nc3 dxe4 4. Nxe4 Bd7 5. Nf3 Bc6 6. Bd3 Nd7 7. O-O Ngf6 8.
Ng3 Be7 9. c4 b6 10. Bg5 [she expected Bf4] O-O 11. Re1 [CSTals evaluation
function liked this position much and scored +2,09 pawns] Bb7 12. Bc2 [CSTal
+ 1.73] h6 13. Bf4 c5 [now the bishop stands where she wanted it to stand,
but therefore she has that weakness on h6. A pretty pawn to sacrifice
on for CSTal.] 14. d5! [very late, and only in permanent brain, CSTal found
d5. Before it would have played dxc with +1,58 pawns. Because Peng thought
that much about c5, CSTal found d5. Thanks!] exd5?! 15. Nf5 [this is a
nice place for a knight. Of course CSTal expected Nxc5] Re8 16. cxd5 Bxd5
[She did never considered about Nxd5. Of course our program would have played
Bxh6 almost instantly with +1,96 pawns. Bxh6! Bf6 Rxe8+ Qxe8 Nd6 Qb8 Qd3
Qxd6 Qh7+ Kf8 Ng5 says the main-line! If gxh instead of Bf6 CSTal says
+2,43 and would play Qd3 and feels good. Peng Zhaoqin has seen that,
was afraid of it and choose the safer Bxd5.]
17. Bc7 [very late, in the last second, CSTal saw this move.] Bxf3
18. Qxf3 Qxc7 19. Rxe7 Rad8 20. Rae1 [still one pawn back, CSTal has
pressure enough for the pawn and scores still +1,45.] Rxe7 21. Rxe7 Re8
22. Rxe8+ Nxe8 [I told before, that a program will find a bug, when
there is one. Later on we found a bug in the draw-repetition code. I will ex-
plain later, in the game of the second round. But of course this bug occurs
also here.] 23. Ne7+ Kf8 24. Ng6+ Kg8 25. Ne7+ Kf8 26.Ng6+ Kg8 27. Ne7+
1/2-1/2

Later on we discussed several different moves from her, as alternatives.
She said: "I have one pawn more, but your pieces are so activ, what
can I do else ?" and tried out some things. But always CSTal found a
way to make it hopeless. So - in the end she was also lucky about the
computer "offer" of draw. She said: also in a game against a human, she
would have made this game into a draw very likely.
I told her that it is very complicate to find out with our program, in
later analysis, "what would have been the best / most accurate move in
this position". Still she looked surprised: "But it is a computer! It
should find the accurate move!" Again I tried to persuade her, that
this program is different. It plays chess like you have to play poker,
I mean, it plays not like Tarrasch, it plays like Tal.
If you want to know what is accurate in a position, ask the fast programs.
But is the best move in a position the best move in a game ?
The best move within a range of 10 plies from a certain position is not
the same best move within a range of knowledge seen for the whole game.
But if you want to know the exact move in this position, as if the rest
of the game never happens, you must use another program, not ours.
Our program is bluffing like a human does.

[Event "AEGON Mankind vs Machines"]
[Site "The Hague NED"]
[Date "1996.04.11"]
[Round "2"]
[White "Richard Oranje"]
[Black "CHESS SYSTEM TAL"]
[Result "1/2-1/2"]

1. Nc3 d5 2. e4 d4 3. Nce2 Nf6 [this human-player was a Van-Geet fan.] 4. e5
[maybe Ng3 would have been better?!] Ng4 [CSTal feels +1,67 and has
Ng4 f4 d3 cxd Qxd3 Qb3 Qxb3 axb in the main-line] 5. f4 d3 6. cxd3 Nc6
7. Nf3 e6 8. a3? [h3 would have been better] Be7 9. Qc2? [again h3 or d4 or
b4 and giving back the pawn would have been better.] Qd5!! [Most programs
want to play O-O, but this is too passive. On O-O white could play
d4 Nh6 b4 Nf5 Bb2 Bd7 Rc1 Rc8 and white comes free. Qd5 with the idea to
control c5-f2 important-diogonal is much more interesting and earns +1,67.
Maybe O-O is the best move for the position. But Qd5 is the best move, if
you have a plan.] 10. h3 Qxf3! [for machines this move is nothing special.
In the later bulletin it was written about the game: "Chess System Tal attracted
a lot of attention when it sacrified its Queen on move 10. A rarity even in
human chess"... One has to be fair: almost every computer will find Qxf3.
CSTal says 2,92 expected the better hxg Qxg4 g3 Bd7 Bg2 O-O-O d4 h6. Now the
bug came again: When the opponent is NOT defending with the best move, when
opponent is better in material -because we sacrified, it thinks the opponent
will force a repetition draw and instead of senseful lines in the tree it sees
only rubbish repetition lines. Although alpha-beta can handle this misbehaviour,
it waists computing time and together with some other strangenesses of the Tal
-Function, it results here again in one problem: Although we stand better
- like in round 1 - it plays the draw-line!] 11. gxf3 Bh4+ [Evaluation increases
to +10,97 pawns! Many people circled arround the board. A very strange
position. Tal would have liked it.] 12. Kd1 Nf2+ 13. Ke1 Nxd3+ 14. Kd1
Nf2+ 15. Ke1 Nxh1+ 16. Kd1 Nf2+ 17. Ke1 Nxh3+ 18. Kd1 [After eating all the
material, the bug came again! Try it out with Genius or Mchess. Also our
fixed version without the bug would have played better, black can easily
continue by playing Bd7 and developing. He should win then.] Nf2+ 19. Ke1 Nd3+ 20.
Kd1 Ncb4 21. axb4 Nf2+ 22. Ke1 Nd3+ 23. Kd1 Nf2+ 1/2-1/2

So again we missed the point or the better play in a better position because
of the same bug.
In the following games, I don't need to show you, they are not that good,
we unbelievably played 3 more draws.
Rumors circled. Jokes were made, called us Chess System Draw.
I was very disapointed. I was that disapointed in the 5th round, that
I was unable to do anything. I commit suicide in some blitz-games the operators
did during the rounds. I thought: better to lose the last round, than again
a draw.

Meanwhile I let CSTal playing against all the concurrent programs, the
programmers or operators joked the "Chess System Draw" joke.

Here are some fantastic games we played just for fun, between the programs:


[Event "1"]
[Site "1h30'blitz 20"Fisher"]
[Date "????.??.??"]
[White "Breaktrough2"]
[Black "CSystem Tal"]
[Result "0-1"]

1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 d6 6. Bg5 e6 7. Qd2 a6 8.
O-O-O h6 9. Be3 Nxd4 10. Bxd4 b5 11. Qe3 Bb7 12. Bxf6 gxf6 13. Qd4 Rc8 14. a4
Bc6 15. axb5 axb5 16. Nxb5 Ra8 17. Qd3 Ra1+ 18. Kd2 Ra4 19. Nd4 Bd7 20. Ke2 Qa5
21. b3 Ra2 22. Rd2 d5 23. exd5 Qxd5 24. c4 Qe5+ 25. Qe3 Rxd2+ 26. Kxd2 Qa5+ 27.
Kd3 Bc5 28. Qg3 e5 29. Qg7 Ke7 30. Qxh8 Qa1 31. Nf3 Bf5+ 32. Ke2 Bc2 33. Qxf6+
Kxf6 34. Kd2 Qd1+ 35. Kc3 Qc1 36. Ne1 Be4+ 0-1

[Event "2"]
[Site "1h30'blitz 20"Fisher"]
[Date "????.??.??"]
[White "CSTal"]
[Black "Breakthrough2"]
[Result "1-0"]

1. c4 b6 2. d4 Bb7 3. Nc3 Nf6 4. Nf3 Bxf3 5. gxf3 a6 6. Bh3 d5 7. Qa4+ c6 8.
Bf4 dxc4 9. Qxc4 b5 10. Qd3 Nd5 11. Nxd5 cxd5 12. a4 Nc6 13. axb5 Nb4 14. Qb3
e5 15. Bd2 Qh4 16. Bg4 h5 17. Rxa6 Nxa6 18. Qxd5 Nb8 19. Qxa8 Bd6 20. dxe5 1-0

It is a pity that Dieter Steinwender played his Karate-Attack against Shredder,
and won with it! I awaited the round Steinwender had to play against us, but
never happened. What a pity. Steinwender played on rank behind CSTal.
Sorry Dieter. Maybe next year you can try.

[Event ""]
[Site "30'blitz"]
[Date "????.??.??"]
[Round "1"]
[White "Shredder"]
[Black "CSystem Tal"]
[Result "1-0"]

1. d4 Nf6 2. Nf3 g6 3. c4 Bg7 4. Nc3 O-O 5. e4 d6 6. Be2 e5 7. Be3 h6 8. O-O
Ng4 9. Bc1 Nc6 10. d5 Ne7 11. h3 Nf6 12. Be3 Bd7 13. c5 Qb8 14. Qd3 Rd8 15.
Rad1 Qc8 16. Nd2 Bxh3 17. cxd6 cxd6 18. gxh3 Qxh3 19. Nc4 Ng4 20. Bf4 Nxf2 21.
Rxf2 Qxd3 22. Bxd3 exf4 23. Rxf4 Bd4+ 24. Kg2 Bxc3 25. bxc3 b5 26. Na5 Rac8 27.
Bxb5 a6 1-0

[Event "?"]
[Site "30'blitz"]
[Date "????.??.??"]
[Round "2"]
[White "CSTal"]
[Black "Shredder"]
[Result "1-0"]

1. d4 Nf6 2. c4 e6 3. Nc3 d5 4. Nf3 c6 5. e3 Nbd7 6. Bd3 dxc4 7. Bxc4 b5 8. Bd3
Bb7 9. O-O Rc8 10. a3 Be7 11. b4 O-O 12. Ne2 a5 13. Qb3 Qc7 14. Bd2 a4 15. Qa2
e5 16. Bf5 exd4 17. Nexd4 g6 18. Bh3 Qd6 19. Rac1 Rcd8 20. Rfd1 Nb6 21. Be1 Qb8
22. Nxc6 Bxc6 23. Rxd8 Qxd8 24. Rxc6 Nc4 25. Nd4 Qd5 26. g3 Qh5 27. Bf1 Qd1 28.
Bxc4 bxc4 29. Qd2 Qxd2 30. Bxd2 Ne4 31. Be1 Nd6 32. Ra6 Rc8 33. Kf1 f5 34. Ke2
Ne4 35. f3 Nf6 36. Rxa4 Kf7 37. b5 Bd6 38. Ra6 Bc5 39. Ne6 Be7 40. Nf4 g5 41.
Ng2 Rc5 42. a4 c3 43. Rc6 Nd5 44. Rxc5 Bxc5 45. Kd3 Bb4 46. h3 Ke6 47. h4 gxh4
48. gxh4 Ke5 49. Bg3+ Kf6 50. Nf4 Nb6 51. e4 Ba5 52. Nd5+ 1-0

[Event "?"]
[Site "30'blitz"]
[Date "????.??.??"]
[Round "3"]
[White "Shredder"]
[Black "CSTal"]
[Result "1-0"]

1. d4 Nf6 2. Bg5 d5 3. Bxf6 exf6 4. e3 c6 5. Bd3 Qb6 6. b3 Qa5+ 7. c3 Bd6 8.
Ne2 Bg4 9. f3 Be6 10. O-O Nd7 11. Nd2 O-O-O 12. c4 Bb4 13. Nb1 dxc4 14. bxc4
Ne5 15. Qc2 Nxd3 16. Qxd3 Qg5 17. Nbc3 Rhe8 18. f4 Qa5 19. f5 Bd7 20. e4 c5 21.
d5 Qa6 22. Kh1 Re5 23. a3 Ba5 24. Rab1 Be8 25. Nb5 Bb6 26. Ng3 Bc7 27. Nxc7
Kxc7 28. Rb2 Kc8 29. Rfb1 Rd7 30. Nh5 g5 31. Rb5 Rd6 32. Rxb7 Qxb7 33. Rxb7
Kxb7 34. h3 Bd7 1-0


A famous opponent had to play here. Although I told Peter Schreiner, the
friendly M-Chess-Operator and good adviser of Marty Hirsch, that we have
60% against Genius and Mchess, and his chances against us were rare, he
did'nt believe it. Here it goes. It is a lucky draw.

[Event "1"]
[Site "30'blitz"]
[Date "????.??.??"]
[White "Mchess Pro Aegon"]
[Black "CSystem Tal"]
[Result "1/2-1/2"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Bc4 e6 7. Bb3 b5 8. O-O
Be7 9. f4 O-O 10. e5 dxe5 11. fxe5 Nfd7 12. Qh5 Nc6 13. Nxc6 Qb6+ 14. Be3 Qxc6
15. Rf2 Bb7 16. Bd4 Rad8 17. Qg4 Bc5 18. Re1 a5 19. a4 Bxd4 20. Qxd4 Nc5 21.
Qe3 b4 22. Nb5 Ne4 23. Rfe2 Qc5 24. c3 Qxe5 25. cxb4 axb4 26. Bc2 f5 27. Bxe4
fxe4 28. b3 Rf5 29. Rd2 Rxd2 30. Qxd2 Qc5+ 31. Qe3 Qxe3+ 32. Rxe3 Rd5 33. Nc7
Rd1+ 34. Kf2 Kf7 35. Ke2 Ra1 36. Nb5 Ra2+ 37. Ke1 Bd5 38. Nd6+ Kg6 39. Nxe4
Rxg2 40. Nc5 Rxh2 41. Nxe6 Kf7 42. Nf4 Ba8 43. Nd3 Rh4 44. Re5 Rd4 45. Ke2 Kf6
46. Re8 Bc6 47. Rf8+ Ke6 48. Rb8 Bd5 1/2-1/2 Draw by agreement.

Also the program Isichess, that has made much progress and had to play
good against strong grandmasters this year, had to feel Chess System Tals
sacrifice mood:

Let me ask one question: Which program would have played 26.Bxg5 !!

[Event "1"]
[Site "15'blitz"]
[Date "????.??.??"]
[White "CSystem Tal"]
[Black "Isichess Aegon96"]
[Result "1/2-1/2"]

1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 dxc4 5. a4 Bf5 6. e3 e6 7. Bxc4 Bb4 8. O-O
O-O 9. Qe2 Nbd7 10. e4 Bg6 11. Bd3 Bh5 12. Bf4 Re8 13. e5 Nd5 14. Nxd5 cxd5 15.
h3 a6 16. Rfc1 f5 17. Qd1 h6 18. Be2 Qb6 19. Rc2 Red8 20. Rac1 Rab8 21. Kh1 Qa5
22. b3 Ba3 23. Ra1 Bb4 24. Qc1 Bxf3 25. Bxf3 g5 26. Bxg5!! hxg5 27. Qxg5+ Kh7 28.
g4 f4 29. Be2 Nf8 30. Bd3+ Kh8 31. Rca2 Bc3 32. Rc1 Rd7 33. Rac2 Bxd4 34. Qxf4
Qb4 35. Kh2 Kg8 36. Rc7 Rbd8 37. Qg5+ Kh8 38. Qh5+ Kg7 39. Qg5+ Kh8 40. Kg2 Qa3
41. Qf4 Ba7 42. Qf6+ Kg8 43. R1c3 Bd4 44. Qg5+ Kh8 45. Rc2 1/2-1/2

[Event "2"]
[Site "15' blitz"]
[Date "????.??.??"]
[White "Isichess"]
[Black "CSTal"]
[Result "1/2-1/2"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nf3 Qc7 8. a4
Be7 9. a5 O-O 10. Be2 Nbd7 11. O-O b5 12. axb6 Nxb6 13. Nh4 Be6 14. Bxa6 Nc4
15. Bxc4 Qxc4 16. Bg5 Rxa1 17. Qxa1 h6 18. Bxf6 Bxf6 19. Nf3 Rb8 20. Rd1 Qb4
21. b3 Bd8 22. Rd3 Rc8 23. Nd2 Bb6 24. Qa4 Qc5 25. Rf3 Qd4 26. Ndb1 Qxa4 27.
bxa4 Bd4 28. Kf1 Bd7 29. Ke1 Rc5 30. Ne2 Rxc2 31. Na3 Ra2 32. Nxd4 exd4 33. Nb5
Ra1+ 34. Kd2 Ra2+ 35. Ke1 Ra1+ 1/2-1/2

Also some unbelievable 5'blitz games against Diogenes, the program of my
eastern-germany Genosse J"rg Burwitz shows highly interesting chess:

[Event "1"]
[Site "5'blitz"]
[Date "????.??.??"]
[White "Diogenes"]
[Black "CSystem Tal"]
[Result "0-1"]

1. e4 e5 2. Bc4 Nf6 3. d3 Nc6 4. Nf3 Bc5 5. O-O d6 6. Nbd2 Bg4 7. h3 Bh5 8. Nb3
Bb6 9. Re1 O-O 10. Bd2 Re8 11. Qe2 a5 12. Be3 Bxe3 13. Qxe3 Nb4 14. Rec1 d5! 15.
exd5 e4!! 16. dxe4 Rxe4 17. Qc5 Bxf3 18. gxf3 b6 19. Qb5 Rf4 20. Rd1 Nd7 21. c3
Ne5 22. Be2 Nc2 23. Rac1 a4! 24. Rxc2 axb3 25. Qxb3 Qg5+ 26. Kh2 Qh4 27. Kg2 Rd8
28. d6 Rf6 29. Rh1 Qg5+ 30. Kf1 Rfxd6 31. h4 Qf5 32. c4 Rd1+! 33. Kg2 Qg6+ 34.
Kh2 R1d4 35. Kh3 Rxh4+! 36. Kxh4 Qg2 37. Qe3 g5+ 38. Kh5 [Qxg5? Ng6+] Qxh1+ 39. Kxg5 h6+ 40.
Kf4 Qh4+ 41. Kxe5 f6+ 42. Ke6 Rd6+ 43. Ke7 f5+ 44. Qg5+ Qxg5+ 45. Ke8 Rd8+ 0-1
The white King is mated on e8 !!

[Event "2"]
[Site "5'blitz"]
[Date "????.??.??"]
[White "Diogenes"]
[Black "CSystem Tal"]
[Result "0-1"]

1. e4 e5 2. Nf3 Nc6 3. d4 exd4 4. Nxd4 Bc5 5. Be3 Qf6 6. c3 Nge7 7. Nc2 Bb6 8.
Bxb6 axb6 9. Nd2 d5 10. exd5 Nxd5 11. Bc4 Be6 12. O-O O-O-O! 13. Re1 Nf4! 14. Ne3
Ne5! 15. Bxe6+ Nxe6 16. Qa4 Nc5!! 17. Qa7 Rxd2!! 18. Qa8+ Kd7 19. Qxh8 [the
copy of the immortal game ?!] Qxf2+ 20. Kh1 Ned3 21. Rf1 Qxe3 22. Rxf7+ Kc6 23. Raf1 Nf2+ 24. Kg1 Ncd3 25. Qf8 Ng4+ 26. Kh1
Ndf2+ 27. Kg1 Ne4+ 0-1

[Event "3"]
[Site "5'blitz"]
[Date "????.??.??"]
[White "Diogenes"]
[Black "CSystem Tal"]
[Result "0-1"]

1. d4 Nf6 2. c4 e6 3. Nc3 Bb4 4. Bg5 h6 5. Bh4 c5 6. e3 Bxc3+ 7. bxc3 d6 8.
dxc5 g5 9. Bg3 dxc5 10. Qxd8+ Kxd8 11. Rd1+ Nbd7 12. Bd3 b6 13. Ne2 Bb7 14. O-O
Ke7 15. Rfe1 e5 16. Nc1 Ne4 17. Bxe4 Bxe4 18. Nd3 f6 19. h3 Rhd8 20. Rd2 Nf8
21. Red1 Rac8 22. f4 exf4 23. exf4 Ne6 24. fxg5 Nxg5 25. Nf4 Rxd2 26. Rxd2 Rg8
27. Kh2 Rd8 28. Nd5+ Kf7 29. a4 Kg7 30. a5 bxa5 31. Bc7 Rd7 32. Bxa5 Kf7 33.
Rd1 f5 34. Rf1 h5 35. h4 Ne6 36. Rf2 Kg6 37. Rd2 Kh6 38. Re2 Rg7 39. g3 Kg6 40.
Rd2 Rb7 41. Nc7 Nxc7 42. Rd6+ Kf7 43. Bxc7 Rb1 44. g4 f4 45. Rd7+ Ke6 46. Bxf4
Kxd7 47. gxh5 Rb3 0-1

Also I analysed one interesting game that happened between Seirawan and Comet.
I guess this is an interesting game, it shows us a nice position:

[Event "AEGON Mankind vs Machines"]
[Site "The Hague NED"]
[Date "1996.04.11"]
[Round "2"]
[White "Yasser Seirawan"]
[Black "COMET"]
[Result "1-0"]
[WhiteTitle "GM"]
[WhiteElo "2630"]
[WhiteCountry "USA"]

1. d4 Nf6 2. c4 c5 3. d5 e6 4. Nc3 exd5 5. cxd5 d6 6. e4 g6 7. f4 Bg7 8. Bb5+
Nfd7 9. Nf3 O-O 10. O-O f5? [a6 or Qa5] 11. exf5 [CSTal says instantly +2,17
for that move.] Rxf5 12. Bd3 [CSTal +2,23] Rf8 13. Ng5! Nb6

DIAGRAM

it is interesting to analyse, that many programs cannot find this move, for
a human it is a very obvious move.

14. Nxh7!! CSTal finds the key-move on ply before, arround 1' on my 486-100
when computing over a better defense than Nb6. CSTal finds Nxh7 in the above
position after a few seconds on my machine with +2,85.
I guess Genius4 makes a good effort on this position. Also Fritz4.01
(patch-version) shall be better in this position than Fritz4.00.
But many programs don't understand this easy move and play Ne6, Re1 or Bc2
etc.
Kxh7 15. Qh5+ Kg8 16. Bxg6 Bd4+ 17. Kh1 Qd7 18. f5 Na6 19. Bh6 1-0


After these games, the programmers of the other programs did'nt joke anymore.
They were, like me, impressed by the sacrificing play of Chess System Tal,
that stopped the rumor we can only play draws.


Henk Arnoldus must have a friendship to David Bronstein. Whenever
he had some time, he walked arround to Bronsteins board, and also vice
versa, Bronstein came and took a look on our game.
It is a strange feeling when such a good player like Bronstein
considers about the game you operate.

[Event "AEGON Mankind vs Machines"]
[Site "The Hague NED"]
[Date "1996.04.17"]
[Round "6"]
[White "Henk Arnoldus"]
[Black "CHESS SYSTEM TAL"]
[Result "0-1"]

1. Nf3 Nf6 2. d4 g6 3. g3 Bg7 4. Bg2 O-O 5. O-O d6 6. c3 Nbd7 7. Nbd2 Nb6 8.
Re1 Bg4 9. h3 Bd7 10. e4 Qc8 11. Kh2 c5 [Puh! I thought we could never
open the game. The Aegon-players are all really old foxes. They try to
hold everything closed and overprotected, they design their own anti-
computer-strategies and are maybe 200 ELO points stronger then their
elo shows it.] 12. e5 Nfd5 13. dxc5 dxc5 14. Qe2 Rd8 15. Nb3 c4 16. Nbd4 Be8
17. e6 [Whi