Where to get the biggest free database of human pgn games?

[Uri Blass]

Thanks in advance.

Uri

[velmarin]

Probe it:

+- 5.000.000 games in pgn,

http://icofy-blog.de/icofy-base/

[Norm Pollock]

I like the ICOfY massive database by Ingo Schwarz. I intend to make a more efficient, modified version, with credit given to Ingo for the foundation. I intend to "clean it up" and "filter it slightly".

But it will take a few days. I will post back when finished.

-Norm

[Vinvin]

Nice ! Thanks !

I like the ICOfY massive database by Ingo Schwarz. I intend to make a more efficient, modified version, with credit given to Ingo for the foundation. I intend to "clean it up" and "filter it slightly".

But it will take a few days. I will post back when finished.

-Norm

[Ozymandias]

You plan on upgrading your own db, then? It was already quite good, very clean.

[Norm Pollock]

You plan on upgrading your own db, then? It was already quite good, very clean.

I am not changing my own db. This is a new project. It has 10x has many games (4M vs 400K). It is for users who want an extremely large collection of games and are willing to sacrifice some quality.

I will make it as clean as I can, correcting errors, removing copies of games with the same moves, removing excess tags, sorting by ECO code, splitting into manageable files. But with 4M games there are things I'm NOT going to do: blunder check 1st 25 moves, remove near-duplicates, remove short-timed games.

[Ozymandias]

I look forward to it.

[Norm Pollock]

https://dl.dropbox.com/u/66249444/ecoAB.7z
https://dl.dropbox.com/u/66249444/ecoCDE.7z

These 2 files unpack into 14 pgn files based on ECO codes. Games are sorted by ECO code. Altogether over 4M games. No duplicates, no FRC, no result errors when there is a mate, ECOs recomputed using SCID.

[Dan Honeycutt]

https://dl.dropbox.com/u/66249444/ecoAB.7z
https://dl.dropbox.com/u/66249444/ecoCDE.7z

These 2 files unpack into 14 pgn files based on ECO codes. Games are sorted by ECO code. Altogether over 4M games. No duplicates, no FRC, no result errors when there is a mate, ECOs recomputed using SCID.

Thanks, Norm, I was in need of some pgn files to test a utility I'm working on. Could you post a link to the 400k collection you mentioned a couple posts up?

Thanks
Dan H.

[Norm Pollock]

Thanks, Norm, I was in need of some pgn files to test a utility I'm working on. Could you post a link to the 400k collection you mentioned a couple posts up?

Thanks
Dan H.

That would be the 4 "gm" pgn files at

http://www.hoflink.com/~npollock/chess.html