The worse thing that happened to Chess

[CRoberson]

Was the invention and use of the first Glickman rating system. After its invention, freechess.org adopted it. In the beginning, I argued with them that
it is a bad system, but they didn't listen. Given the open source release of their server software, the plague that is the first Glickman rating system is spreading like a virus to chess.com and others.

The problem with it is that it does not assume practice makes perfect. It only assumes that practice makes you consistent and only consistent which is clearly
wrong. If you like at the USCF rating graphs for most people, you'll see that people that improve, have oscillating ratings but the oscillation reduces. At the
time it reduces an epiphany happens and their rating jumps. Just after the rating
jump, the oscillation goes back to what it was before, thus showing Glickman's idea that with more and more practice you become more consistent and that
rating system is adjusted so that minimal rating increase is possible.

Even Glickman has announce that his first system has the flaw of omitting the practice makes perfect concept. So, he has created new systems that
incorporate practice makes perfect.

So, I request a petition be made and sent to Chess.com, freechess.org and others to drop that system. It seems freechess.org doesn't have anybody
that can code the change, but it isn't that hard.

I have seen many people quit playing online due to that system. How many people have started playing chess online instead of at a club and then quit due to that system?

[AdminX]

Hi Charles,

Question: Do you know what system Fide Online uses? I ask because I believe it was just Premium Chess before the partnership with FIDE. I don't know if they ever changed the rating system they have in place.

http://www.premiumchess.net/

[reflectionofpower]

Was the invention and use of the first Glickman rating system. After its invention, freechess.org adopted it. In the beginning, I argued with them that
it is a bad system, but they didn't listen.
The problem with it is that it does not assume practice makes perfect. It only assumes that practice makes you consistent and only consistent which is clearly
wrong.
I have seen many people quit playing online due to that system. How many people have started playing chess online instead of at a club and then quit due to that system?

I have noticed I due best if I keep it 30-35 games a day, anything over that and I get sloppy and I start playing like a monkey brain and I am not noticing trends,misapplied concepts, anything less than that I do even better as a rule. The RD concept is theoretical and as such does not come out on paper/chess board.

[Sean Evans]

Hi Charles,

Question: Do you know what system Fide Online uses? I ask because I believe it was just Premium Chess before the partnership with FIDE. I don't know if they ever changed the rating system they have in place.

http://www.premiumchess.net/

Premium chess now has 800 players online and Chess.com has 15,000 players online, so where will it be easier to get a time control and opponent rating faster? The big issue with Chess.com is their crappy Internet GUI, they need an app I can download to my PC with more controls similar to Playchess.com

Thanks,

Sean

[reflectionofpower]

Hi Charles,

Question: Do you know what system Fide Online uses? I ask because I believe it was just Premium Chess before the partnership with FIDE. I don't know if they ever changed the rating system they have in place.

http://www.premiumchess.net/

Premium chess now has 800 players online and Chess.com has 15,000 players online, so where will it be easier to get a time control and opponent rating faster? The big issue with Chess.com is their crappy Internet GUI, they need an app I can download to my PC with more controls similar to Playchess.com

Thanks,

Sean

I bet if opponents could have a webcam game with that avatar model of yours the numbers would be reversed

[AdminX]

Hi Charles,

Question: Do you know what system Fide Online uses? I ask because I believe it was just Premium Chess before the partnership with FIDE. I don't know if they ever changed the rating system they have in place.

http://www.premiumchess.net/

Premium chess now has 800 players online and Chess.com has 15,000 players online, so where will it be easier to get a time control and opponent rating faster? The big issue with Chess.com is their crappy Internet GUI, they need an app I can download to my PC with more controls similar to Playchess.com

Thanks,

Sean

I am not a fan of any site in particular, fact is I play on many.

[reflectionofpower]

Hi Charles,

Question: Do you know what system Fide Online uses? I ask because I believe it was just Premium Chess before the partnership with FIDE. I don't know if they ever changed the rating system they have in place.

http://www.premiumchess.net/

Premium chess now has 800 players online and Chess.com has 15,000 players online, so where will it be easier to get a time control and opponent rating faster? The big issue with Chess.com is their crappy Internet GUI, they need an app I can download to my PC with more controls similar to Playchess.com

Thanks,

Sean

I am not a fan of any site in particular, fact is I play on many.

ICC in my opinion is top notch and has been around a long time and it's access to features through BlitzIn is awesome. FICS is good to but GUI's are lacking.

[bob]

Was the invention and use of the first Glickman rating system. After its invention, freechess.org adopted it. In the beginning, I argued with them that
it is a bad system, but they didn't listen. Given the open source release of their server software, the plague that is the first Glickman rating system is spreading like a virus to chess.com and others.

The problem with it is that it does not assume practice makes perfect. It only assumes that practice makes you consistent and only consistent which is clearly
wrong. If you like at the USCF rating graphs for most people, you'll see that people that improve, have oscillating ratings but the oscillation reduces. At the
time it reduces an epiphany happens and their rating jumps. Just after the rating
jump, the oscillation goes back to what it was before, thus showing Glickman's idea that with more and more practice you become more consistent and that
rating system is adjusted so that minimal rating increase is possible.

Even Glickman has announce that his first system has the flaw of omitting the practice makes perfect concept. So, he has created new systems that
incorporate practice makes perfect.

So, I request a petition be made and sent to Chess.com, freechess.org and others to drop that system. It seems freechess.org doesn't have anybody
that can code the change, but it isn't that hard.

I have seen many people quit playing online due to that system. How many people have started playing chess online instead of at a club and then quit due to that system?

I don't think the Glicko system is that bad. The normal Elo system also has its warts. For one example, it assumes a modest number of games over time and therefore responds too quickly to performance changes. In the days where one might play 5-10 games in a month, it was not bad. In the days where one can play 20-30 games in one day, it is too reactive. Nobody's "real rating" will change 100 points in a day, yet that is not uncommon using the usual Elo formula.

A computer's rating doesn't change at all unless you change the program, yet computer ratings fluctuate wildly on ICC just like humans.

Neither the old or new Glicko nor current USCF is all that accurate, particularly for server ratings. I'd hope you would not think "practice makes perfect" has anything to do with computers, as one counter-point.

[CRoberson]

Actually, I think it is provably worse than the Elo system. In the Elo system, if you lose to an opponent then win 3 and you are supposed to score 75% with
them, then your rating is unchanged. Not so in Glickman's first system. If you lose the first game, you lose rating points and the RD value drops after that game and
after each following game which reduces the points you win back which means you will not have the same rating that you started with; you will have a new
rating less than your starting rating. Thus, the first Glickman system is deflationary. Multiply that effect by large numbers of people and the whole rating pool is deflated.

Thus, the first Glickman system is not good at all.

[OneTrickPony]

I was arguing about it with chesstempo author/owner and it's main reason I don't use the site too much anymore.
Glicko ratings are just not fun (even improved one). Great thing about ELO is that it fluctuates and changes quickly when you start playing better. Glicko ratings are great for entities like computers who have the same strength all the time but terrible and demotivating for humans who have better days, worse days, sometimes improve, sometimes play in bad state of mind etc.
Also time constraints (the more time you don't play the more RD increases) are completely arbitrary. If you play like 30 games one day and you don't do well because you were drunk, tired or just wanted to let the steam off that will hunt in Glicko system for long long time which again makes it not fun.