You can download the entire archive of all completed games as of 11/29/2003. The file is 30Mb compressed, and over 120Mb uncompressed, so you'll probably want to make sure you really want it before clicking the URL. It's in GZip format, which WinZip, and almost any other compression program, can handle.
http://www.net-chess.com/archive.pgn.gz
Later I'll divide it up by year/month you you can get all of the new games each month without downloading the whole thing again.
Complete archive download - as of 11/29/2003
-
tellymetwise
- Posts: 360
- Joined: Mon Sep 22, 2003 3:48 pm
I've downloaded the Archive and after cleaning it up a bit (placing Event above Site, Writing WhiteELO into WhiteElo and such) and discovered two thing about ratings ...
Between oktober 2002 and latest 2003 there has been an inflation of 300 points average with some as high as 600 or more, depending their endrating, or the beginning of the rise. (some start half spring to rise).
Between the end of 2001 and oktober 2003 the inflation was about a 100 average, again, with some going 300 points up.
With both I've ignored games where people jump 1000 points up within 2 months time, as they tend to have few games here or drop back down 500 points after as much months.
Maybe an idea, scratch any timeout from the Archive which doesn't count towards the rating. It saves quite some space (35 Mb against 85 Mb with "finished" games.)
Between oktober 2002 and latest 2003 there has been an inflation of 300 points average with some as high as 600 or more, depending their endrating, or the beginning of the rise. (some start half spring to rise).
Between the end of 2001 and oktober 2003 the inflation was about a 100 average, again, with some going 300 points up.
With both I've ignored games where people jump 1000 points up within 2 months time, as they tend to have few games here or drop back down 500 points after as much months.
Maybe an idea, scratch any timeout from the Archive which doesn't count towards the rating. It saves quite some space (35 Mb against 85 Mb with "finished" games.)
-
talfischerman
- Posts: 2
- Joined: Tue Dec 16, 2003 3:47 pm
Does ratings even matter?
You are getting slower more preserved games here at NET-CHESS.COM. These games are blunder free 
(well for the most part lol) The fun of the game should over ride these inflated ratings. If you play on the internet you should know that ratings are always going to be inflated and not perfect. I feel the only "real ratings" are over the board ratings because you can see your opponent. You might have more blunders but thats just the name of the game also. Have fun and enjoy yourself and dont worry about the ratings because they arent real. HAPPY HOLIDAYS.
(well for the most part lol) The fun of the game should over ride these inflated ratings. If you play on the internet you should know that ratings are always going to be inflated and not perfect. I feel the only "real ratings" are over the board ratings because you can see your opponent. You might have more blunders but thats just the name of the game also. Have fun and enjoy yourself and dont worry about the ratings because they arent real. HAPPY HOLIDAYS.-
tellymetwise
- Posts: 360
- Joined: Mon Sep 22, 2003 3:48 pm
Re: Does ratings even matter?
Dont mistake observations with presumed feelings.
a general raise of ratings will always be observed over time, when people start to get better. Especially when you look at players who've been playing here for some time.
if a general increase is too high it can only mean a couple of things.
- A couple of start ratings of 9999 or higher generated an 200-600 increase above normal.
- The floor generated a somewhat less influence on inflation, but in combination with the above it became more widespread.
- The relative rating between players is stable and functional.
- There are not a lot of people cheating, or cheating over a long time. If arlowc still hangs around, he might visit greg's provided "friends page".
As far as I know, Greg lowered the floor by a hundred, and placed a top on the starting rating.
The factor is about 0.75 times your current rating to get your actual OTB rating.
But the actual reason why I downloaded the database, was in search for puzzles.
a general raise of ratings will always be observed over time, when people start to get better. Especially when you look at players who've been playing here for some time.
if a general increase is too high it can only mean a couple of things.
- A couple of start ratings of 9999 or higher generated an 200-600 increase above normal.
- The floor generated a somewhat less influence on inflation, but in combination with the above it became more widespread.
- The relative rating between players is stable and functional.
- There are not a lot of people cheating, or cheating over a long time. If arlowc still hangs around, he might visit greg's provided "friends page".
As far as I know, Greg lowered the floor by a hundred, and placed a top on the starting rating.
The factor is about 0.75 times your current rating to get your actual OTB rating.
But the actual reason why I downloaded the database, was in search for puzzles.