i like the ratings and think it's a great way to determine the better player. Despite how silly this may sound everybody has bad days and make mistakes.

assume player "A" plays 100 games against player "B" and wins 75% of the time. who's the better player? Say they both entered a tournament and ended up facing each other once, player "B" wins and comes in 1st place while player "A" comes in 2nd. does that mean player "B" is the better player?

if there was a tournament called the palooza. the winner would be the champion of the palooza tournament and dubbed "palooza champion".  The winner of the world championships will be called the "world champion" not because he's the best but because he is the champion of the world championships. It's just the title for that one tournament.

mueller wrote, "in chess, if you have Topalov play some 12 year old from a middle school in louisiana who just learned to play chess a week ago, its not going to be much of a game" This is true. I thought you all, as they say in Louisiana, would appreciate what might seem peculiar about the rating system. The evaluation of your performance is not just based on your results against the opposition, but on who you are. If this 12 year old, or even the unrated Paul Morphy, played in a tournament with nine 2500 players and achieved a plus score he would need a higher score to obtain an IM or GM norm than another player would who had a rating of 2500 and played in an event with nine other 2500 rated players. This is because the group of ten 2500 rated players and the group of nine 2500 players plus the kid (or even Morphy) are treated as two different groups with different scores needed for norms. If a score of 6-4 were enough for a norm in one group, it would not be enough in the other.

I like ratings for "placements". They are hardly infallible and chess has multiple examples of the underdog pulling a win. It's all statistical. They are an indicator of who's stronger after many, many games. I see them more of a stochastic indicator than a "forced win" warning.

If you put a WC contender (Elo 2600) Vs. a national of Fide master (2100), odds are the former is going to win. However, there's a lot of "hard to swallow" opponents whose styles complicate others and make it hard for the highest ranked player to achieve the victories. Furthermore, current tournament rules and rating changes make it even harder to gage opponents. If Anand has a bad day against Roberto Martín del Campo (look him up if you will) and ties, it doesn't mean he's an equal-caliber player. Over the long, long run, chess ratings are designed to either a) reflect the differences in strength of a group of players or b) adjust the relative playing strength of such a group.

This is further complicated by most "modern" competition modes where very rarely there are match games to high scores (say, 8 points to win the match or more). No offense to some "world champions" like Ponomariov, Khalifman or Khasimdzhanov, but the competition system under which they won the FIDE championships is hardly as robust as the long match plays that crowned Botvinnik, Tal, Petrosian, Fischer, Kasparov or Karpov.

I'm all for more tournaments, and for sure the most accurate way to "place" the competitors there are through the use of the ratings (with some allowance for prodigies that are unrated/untitled, like when Kasparov went to Banja Luka and won an extremely strong tournament and emerged with the 3rd highest ELO.

I partly concur with the original post in that ratings are often abused, inflated, and misrepresented. But in order to arrange player's relative strengths, they are for sure useful - more so than the wild two-game-knockout series. Or does anyone else things that Khalifman would win a series of matches against Anand, Kramnik, Topalov and Shirov?

Nicely written SonofPearl, I found it a good read. I certainly agree with most people here about the usefulness of ratings, especially when considering mueller's point - even if ratings are only fairly accurate they provide a good way to make chess competition more efficient.

When considering the match aspect of chess, I agree with your initial point - the best in a series of 24 games can, especially if the gap in scores is large, call him or herself the better player. However, this is the best in the 1v1 match, not necessarily the best of the world. I think ratings prove their usefulness here.

The credibility of the world chess championship is largely the result of the requirements for challengers - they have to effectively beat the highest rated players (or those who played best in a tournaments with the highest rated players).The highest rated players were the best performing players. The challengers, and the challengers who became world champion only became world champion because they defeated the pool of (top rated) best performing players.

I would then say that ratings do increase the credibility of the world championship because they assure us that those two players at the table are at least arguably the best players of the world, because they really did perform well against people that also usually perform well.

I do think that the fact that (live) ratings change quicker than the world championship matches can be played should not have too much effect. Under any circumstance, if Carlsen became the best rated, best performing player, he should win the right to challenge for the next year. There's good charm in the WC title being at least for one year. After all, you can't loose the GM title either :).

Ratings help determine who should play who, not the outcome of the match, though it is often easy to predict.

Per your boxing analogy, if 2 boxers are getting in a fight, one is a professional who has a trainer and is in a high level circuit and the other is a drunk college student who you gave some gloves to, you probably know that fight isn't going to be very challenging for the pro.

similarly, in chess, if you have Topalov play some 12 year old from a middle school in louisiana who just learned to play chess a week ago, its not going to be much of a game. Without ratings, the best players would have to waste a lot of their time playing worse players who aren't any challenge for them, because there would be no way to gauge their comparative strength beforehand. And given the fact there are millions of chess players, someone could have a record of 2000-5-1 (in chess) and still only be rated 1200 or something, if they only played people rated 800ish. And someone could have a record of 500-500-500 and be rated 2645 if they played people of a similar rating to their own. Elo isn't infalliable but it certainly is a lot better than not having a rating system. 

Very interesting post -

Without ratings, we may have more regular chess players.

I like them too, they help me measure how I am progressing.

i like ratings they are a gauge, even if they have a lot of infallibility's