Hallo !

Das Thema, ob und wann man eine Aktivschachpartie beenden soll, wird ja heiß diskutiert - dazu hier auch meine Einschätzung. Diese ist bis auf einen
kleinen Punkt mit der Bewertung von Nick völlig identisch - obwohl sie absolut unabhängig voneinander entstand. Anläßlich seines U1400-Turniers hatten
wir einen kurzen Briefwechsel bzw. email-Austausch, aus dem ich hier zitieren möchte:

Nick said: "So for me the rule is simple 3 x repetition is a draw. The computer cannot claim to be better because it won by cheating or not knowing the rules. No excuse".

Dazu mein Standpunkt (hoffentlich mehr oder minder verständlich - in etwas ähnlichem wie englisch...):

"This looks absolutely right at the first sight. I have to confess that at the beginning I wanted to use the repetition rule the same easy way as you do - but I thought twice and changed my mind.

Let's have a close look at it:
If you play 2 matches at the same time, are you always aware of the positions on the boards, especially when the 3fold repetition does not occur 3 times in a row ? I have to confess that I am not and I leave it up to the computer...as the rule states...the player has to claim the draw !

But this is only one aspect: the other and more important is the principle of fairness. Chess Shadow and Mephisto Teufelchen for example are able to recognize a 3fold repetition by itself (not exactly true, but at least if the repetitions occur in a row). They do claim for a draw - game over !

But most of the weak computers don't even know this repetition rule - a big drawback ! IMHO it should also result in a disadvantage - they don't complain for the draw, therefore they don`t deserve the immediate draw, so I play on a few moves.
This does not make a huge difference, sometimes it even averages it out, but sometimes it is to their disadvantage.

Or do you think it is fair to settle all matches as a draw after the third repetition ?
No, it's not justified IMHO, because the weak computers then are handled as if they would master the repetition rule which they obviously don`t.
In this case one would take away an advantage of these computers which had offered some bytes to implement the repetition rule. The other computer knows absolutely nothing about the repetition rule - therefore can it really complain to get punished ? I see no reason why.

There is a real difference between let's say Chess Shadow which "knows" the rule and the other "weaklings" which don't - and this should show up, at least a little bit.
"All computers are equal, but some are more equal than others".

Of course you can play your tourney as you please.

I made up my rules before I startet a similar project and I am not willing to change them. These are the "golden rules" of my private Elo-list:
1. Play as many games as possible
2. Play computers of similar strength against each other (try to avoid results of more than 85% respectively less than 15%)
3. Play an identical number of games against many different computers regarding colour change white/black.
4. Play until mate with "weaklings" (boring, but necessary - 2 queens more don`t always guarantee a victory).
5. Doubles don`t count.
6. 3fold repetition as explained above

I totally agree with you that one should not recommend these computers which don't know the repetition rule to beginners. But be aware of: even Mephisto Rebell doesn't know this rule exactly - it considers only repetitions in a row as a draw.
Probably if we look at the rule even more exactly the result would be more disappointing: if I remember well the rule, then even the status of castling and en passant have to be taken into account. I have never tested this systematically, but I am quite sure that most of the chess computers will fail!

Soweit der Briefwechsel mit Nick - bitte beachten: das oben gesagte bezieht sich auf die "Elo-Winzlinge" U 1400 !

Es ist m.E. völlig in Ordnung, wenn man bei Geräten mit einer Elozahl >1600 nicht bis zum Matt spielt, sondern bei ausreichendem Vorteil die Partie entsprechend abschätzt. Ich glaube, daß man damit in 99% der Fälle richtig liegt. So würde ich auch Nicks Beispielfälle mit Dame gegen König behandeln - bei den U1400 Geräten muss man bis zum bitteren Ende spielen und ggf. mit Remis entweder durch Zugwiederholung oder 50 Züge-Regel werten. Beim Tasc R 40 kann man sich den Rest wirklich schenken...wieder anders beim elostarken Dominator: Experten wissen, daß fast alle Dominator-Versionen die Mattsetzung Dame gegen blanken König nicht beherrschen !

Viele Grüße

Hans-Jürgen