Nunn2 Test 20 or 25 Positions ?

[brianr]

IIRC, this topic has come up before, but I could not find anything after quite a bit of online searching, so I'm hoping someone here can chime in.

For the Nunn2 test set, are there 20 or 25 positions?

I have files with both.
Most internet references are to 20 positions.
Some papers and ChessBase references are to a 25 position version.

Thanks,
Brian

[Eelco de Groot]

IIRC, this topic has come up before, but I could not find anything after quite a bit of online searching, so I'm hoping someone here can chime in.

For the Nunn2 test set, are there 20 or 25 positions?

I have files with both.
Most internet references are to 20 positions.
Some papers and ChessBase references are to a 25 position version.

Thanks,
Brian

Hello Brian,

I did a quick search on the CSS forum, and Ernest Bonnem said there a while ago that John Nunn apparently did not authorize either the 20 or 25 position Nunn 2 tests himself and since they appeared first on the Chessbase site I presume they are probably made by somebody there without mentioning they were not by John Nunn himself. Not that I think it is so very important which kind of positions you would put in a testset, as long as you put in enough different positions and reverse colours each time, you will introduce some bias or tuning for positions regardless what positions you use. So do not always use the same positions for your tests, that is the most important thing to consider regarding Nunn-type testing.

Geschrieben von Ernest Bonnem (fr) am 17. November 2006 17:45:22:

Als Antwort auf: Re: Wieder mal falsches Nunn2 geschrieben von Wolfgang Battig am 17. November 2006 16:14:06:

>>was ist denn so falsch an diesem Test, der mit der chessbase GUI kommt?
>>Kann ich Deinen Test einfach aus der PGN raus laden ins engine match?
>>Beste Grüße
>>Clemens Keck
>"falsch" ist daran gar nichts, es ist nur nicht der originale Nunn2-Test, denn der hatte nur 20 Stellungen. Woher die 5 anderen Stellungen kommen ist mir nicht bekannt.
>Gruß
>Wolfgang

Nunn selbst wurde gefragt und hat gesagt, daß er nicht weiß woher diese 25 Stellungen (nicht nur 5!) kommen.

ChessBase hat natürlich nie ein Wort davon gesagt...

Regards, Eelco

[Albert Silver]

IIRC, this topic has come up before, but I could not find anything after quite a bit of online searching, so I'm hoping someone here can chime in.

For the Nunn2 test set, are there 20 or 25 positions?

I have files with both.
Most internet references are to 20 positions.
Some papers and ChessBase references are to a 25 position version.

Thanks,
Brian

20.

I asked Nunn and he replied himself in CSS.

Albert

[ernest]

Ernest Bonnem said ...

The genuine and recognized by Nunn himself Nunn2 test set, is 20 positions.

Why Chessbase calls some 25-position test set, shipped with Chessbase engines, as "Nunn2" is obviously a misnomer ...

There also is a genuine Nunn (also called Nunn1) test set, which is 10 positions.

Both genuine Nunn and Nunn2 test sets can be found in Kurt Utzinger's site

http://www.utzingerk.com/test.htm

[brianr]

Thanks for the definative answers.
Brian