OM GOLEM database with 30 mio games

[Glarean]

Does anyone here have the OM GOLEM database on their computer?
https://www.openingmaster.com/index.php/chess-databases
If yes: How is its quality (duplicates, bye games, game lengths, player names etc.) ?

Greetings: Glarean

.

[Werewolf]

I do. I can't comment on quality like duplicates, I've not tested that.

But it often finds games in positions where Megabase 2021 has nothing - so it's definitely bigger. But be warned the quality of the games themselves is questionable, blitz games of weak players count.

[Glarean]

Thanks for the info. That's what I thought too. But it probably also contains millions of blitz games from strong players, right? Next question: Are you using Chessbase? Or Scid? Is the handling with 30 million games a problem?

[KLc]

I don't have the database, partially because I was put off a bit by postings of the "owner/producer" of the database on a chess.com thread. Anyways, quality is not easy to achieve. The ChessBase Big/Mega Database (8+ million games) still contains a lot of garbage (I estimate ~5%) but at least player names, time controls, etc. seem to be fine mostly and to date I haven't found a better database in this respect. The free Caissabase (4+ million games) is OK but certainly contains more garbage and it's more difficult to filter because time controls etc. aren't set properly. I don't believe it's possible to ensure high quality with 30+ million games. Working with the Mega Database in ChessBase is OK but it's certainly not fast (I believe SCID is actually faster). Working with 30+ million games in ChessBase is probably possible but you'll need a good coffee machine.

[Ozymandias]

Curious, the guy who wasn't willing to invest 300€/month on a DB maintainer is the only one complaining about a free DB maybe using copyrighted material.

As for the DB itself, I bought it years ago and it was full of duplicates and bad formatting. Not to mention most of the games in the Golem are blitz human internet games. If it had 30 million Playchess engine room games...

[carldaman]

Right, he emphasizes quantity over quality - but, it may be useful if looking for some quality games that are not found in the 'better' databases. Still not sure if it's worth what he's asking. [I am referring to his other bases, not GOLEM necessarily.]

[Ozymandias]

I can't speak for the current situation, but when I compared them, there was very little value to choosing one over the other. Just go with whatever strikes a good deal with you and stick to it.

Computer games are freely available, correspondence games are freely available, playchess and infinity chess games are freely available... Human OTB chess games are freely available (partly). If you want to spend money on a regular basis for the extra human games you can find in paid DBs, by all means go for it.

[Werewolf]

Thanks for the info. That's what I thought too. But it probably also contains millions of blitz games from strong players, right? Next question: Are you using Chessbase? Or Scid? Is the handling with 30 million games a problem?

It's OK ish on the speed side. You'll need a fast SSD, convert it to CB format and create a booster (I'm using CB16)

[Dann Corbit]

Scid can never hold that many games.
There are a number of limits that are hardwired even into the database format itself, so a simple code change cannot fix it.

[jdart]

It's unfortunate, but there are a lot of garbage games out there. At the least, people have mangled and abused the Event and Site fields, if not corrupted the actual game. Then the bad games are copied and pasted into other databases and the rot spreads. ChessBase is pretty good about cleaning up what gets into the database, and TWIC has been publishing high-quality games for years. Other than that, there are very few guarantees about quality, in my experience.