Is a 100 point difference actually significant?

I don't see how this follows. In the Elo system your K factor is reduced after you have a certain number of games. This allows a new rating to adjust quickly to the correct level. The further reduction of the K factor at the higher levels is not to increase the accuracy of the rating, but to control rating inflation. So you can't assume that a rating is inaccurate just because it isn't GM level.

Ichabod801, you are correct, that's why I put the "accuracy" in between quotation marks. "The further reduction of the K factor at the higher levels is not to increase the accuracy of the rating, but to control rating inflation", that is correct, but I used my own words to say that the inflation itself is the inaccuracy.

That was well explained, I see your point.  As a long time blitz player I wasn't sure, even after reading your post, how much "injected blunders" make a difference though, as your ability to avoid and catch them are a set skill -- although thinking back I do remember my rating can fluctuate even 100 points (and even more for others), which does reflect the volatility of the game where any set measure of skill may never be completely accurate.

I've never realized this about blitz, very interesting.

There is a considerable difference between a 1600 and an 1800 player. Both are patzers, but the 1800 will rarely make the type of elementary blunders the 1600 still makes. All in all, they are both clueless but the 1800 is probably smarter than the 1600.

My rating is around 1550 and I have beaten several 1600's and a few 1700's. Most of the time I've been beaten but I've almost always given the stronger player a good fight. Early this month I beat a person with a 2000 rating, after he made a very bad blunder. That was pure luck.

So would you say that perhaps, if we take an average of blunders/mistakes of each 1200, 1400, 1600, 1800 player, we'd get different numbers and a correlation between their rating and their mistakes? So perhaps getting to 1600 or so, is all about fixing your mistakes and blunders and perhaps improving your tactics?

Take a smart guy and another guy who is also smart but not as smart as the first one, make sure neither one knows anything about chess, and have them practice chess for a year. Assuming that during that year they both played the same amount of chess games against equally strong opponents, and studied equally hard, guess which guy will have the higher rating at the end of the year?

Some people are just better at picturing things in their minds. Not only can they literally visualize the entire board but they are also good at calculating sequences of moves. They can, for example, look at a tense position on the chessboard and effortlessly calculate which side would come out ahead of a major piece exchange.

Presented with the possibility of a piece exchange, an inferior chess player's thought process might be something like: "if I capture the pawn at e5 with my bishop, black will capture my bishop with his rook, which I will then capture with my knight, which black will then capture with either his queen or his other knight, which I will then capture with my queen, but then black will capture my queen with either his queen or his knight, so maybe I shouldn't push this exchange." It's a logical thought process, but a better chess player would make that calculation in an instant and focus on something more important.

I don't believe that every GM has become a GM because they were already extremely smart. I think it's because they improved their calculation skills and memorization skills farther than everyone else. They dedicated a lot more time than the average person to studying and understand chess.

Sure someone born naturally smarter, may have a quicker time getting to a GM level, but I don't think that someone with a relatively average IQ and all, cannot do the same through dedication.

Very enlightening, thanks for the good info!

They became GM's because they were very smart, studied hard, had incredible memories, and improved their calculation skills farther than everyone else.

"They dedicated a lot more time than the average person to studying and understand chess."

True, but the average person could have studied the same amount and still not become a GM.

"Sure someone born naturally smarter, may have a quicker time getting to a GM level, but I don't think that someone with a relatively average IQ and all, cannot do the same through dedication."

Unless the person with a "relatively average IQ" is a savant in some aspect of chess that's not measured in a standard IQ test, he will never become a legitimate GM.

I understand tony, and what do you think of this concept of becoming a GM, do you have to be a savant of some sort or be extremely smart in puzzle solving and such to become better?

I believe its a matter of being able to adapt and change the way you think about chess to become a GM. Sure there are those who may play for decades, and still be around 1200, even with 1000s of games. However, they did not adapt and change the way they calculate, perhaps coaching has a lot to do with becoming a GM too.

I think in our search for the nuances of ratings, what we are actually seeking is the answer to the question - "Can I do that?". First, we have to accept one fact, we MUST accept it. We are NOT someone else, we have our own talents, our own shortcomings, our own capacity for success or failure. An important step in chess improvement is to strive to be the best player you can be, NOT the best player you can be compared to someone else. It is too easy to compare yourself to others, perhaps the newest wonderkid from some remote cave, but what is the point really. You are not them. You are NOT them. If you want to play the game better, do a systematic study. Take an appropriate amount of time to solidify your fundamental skills, and work, work, WORK at polishing your skill to a master level. Stop wasting your time wondering how rating grow, or what they mean, or how they look in a ratings fashion show. The longer you are spending on rating theory, the LESS time you are spending on improving your chess skill. The secret to chess ratings - EFFORT EFFORT EFFORT

I think DeathScepter makes a good point when he says if you want to become better, basically roll up your sleeves and do some systematic learning.  Don't just read 10 pages out of one book one day, then study some line of the Sicilian another.  Many GMs had to work very hard to maximise their ability before earning the title.

But comparison isn't bad.  Part of learning how to improve is measuring yourself against others.  What do they think about some aspect of chess, how have they learned, what did they have to do, what roadblocks did they encounter.  It wont be the same for everyone of course, but it will be similar for most.

There are different ways of studying that I think can improve your gaming significantly. For example, many people might have seen in GM games 3-4 different kinds of cool mates, but how many of them actually remembered that during a game and implemented it?

Same with Tactics Trainer, you see all the tactics there, but how well do you apply those to your game? In TT, you know a good move exists, so you look at all the options, but in a blitz game, you don't notice such good moves at all, because you are not looking hard enough you are trying to play positionally.

In terms of chess understanding how is a 2500 different from a 2200?   The usual.chess games last 40 moves so what.does  the 2500 see in such a short series of moves?

A 2500 would make, in average, better moves and less mistakes than the 2200, he may also understand more about the opening theory and have a wider and deeper repertoire.

not really

at -450 points you still have 5% chance of winning;at minus 500 =4% at minus 750= 0%

The only reason the chess.com rating differentials aren’t as reliable at lower levels (900-1200) is because the participant pool is saturated with people who’ve only put up a handful of games. If you challenge only players with 50+ games under their belt however you’ll probably find the ratings fairly accurate (i.e. 1200 will be very noticeably better than 900)