Characterizing Rating Levels

Hey everyone,

I had what I think might be an interesting idea that, should we work it through properly, with necessary help from high rated players, might leave us with an interesting chart (* summons artfizz *). If nothing else, it might also spur some interesting discussions.

Basically my idea is as follows; I thought we could try and describe general playing tendencies according to rating ranges. I’m too much the patzer to come up with some of the specifics, which is why I can only kick it off with some silly examples.

For instance, something rough like;

900; often neglects development, regularly hangs pieces, often pushes a & h pawns and develops rooks early, sometimes misses mates-in-1, etc.

1250; little to no knowledge of opening theory or principles, often makes uncoordinated attacks, often immediately trades pieces should the opportunity arise, sometimes hangs pieces, often sees mates-in-1.

1350; Follows opening principles, often is comfortable with one or two opening systems, usually anticipates forks, understands basic tactical ideas (pins, discoveries)

1450; Preliminary understanding of ideas like maintaining the tension, creating imbalances…

1550 and up; I can kindof imagine but barely, in as much as I've never cracked this... ???

Well, you can see something of what I’m imagining here. Maybe it would work if we made a list of tendencies like “Recognises Mates-in-3” or “Recognises Forks” running along one axis and then tally “Almost always, Usually, Sometimes, Rarely, Never”, I dunno – artfizz and excelguru are the… gurus of that.

If you think that it’s too difficult to assess, or not a worthwhile cause du jour, that’s cool. Maybe it’s not, maybe it’s too difficult to really pin down?

Just an idea.

I think this idea could develop nicely if done properly. The toughest part is to characterize each interval like that, cause part of the reason some people do decent, but not amazing, is because they know and implement one theory better than another.

Get idea I can see where I place and it is embarrassing to say the least. BUt what a great idea I hoep it moves forward.

it's a good idea, though you might want to use larger brackets at first and reduce them as we make progress with the analytical work

I'd suggest using 200 pts. brackets or USCF classification (class F, class E...)

This is just me describing myself ...

900 - Still learning the basics, didn't really understand the rating system or any of the main concepts of chess.  I was quite proud that I knew en passant when none of my friends did and I used it whenever possible.  No understanding of positional play and minimal tactical understanding.  But hey, I could beat my dad so I HAD to be good :).

1300 - I quickly realize that there are a lot of other kids that can beat their dad too.  Meeting up with these much stronger opponents, I begin to notice tactics ... en passant is no longer a novelty :).

1500 - I realize how terrible I actually am.  Much effort finally begins to go into deeper tactics, and I very rarely missed a tactic (for or against...) however, I knew very little of openings and often found myself 'losing for no reason' against class A players simply because I didn't understand how midgames take you to the endgame.

1600 - "Oh, that's why I lost" ... My tactical eye, now very well trained has left me making very few mistakes... I begin to develop favourite openings ... my positional understanding grows very quickly.

took me less than 2 months to go from 1500's to ...

1900 - "Oh well that can't be right" ... my intuition (which I still believe I've always had) is now being used for good ... instead of speculative sacrifices I begin to look at my position more critically (and my opponents position more agressively)... becoming accustomed to winning most of my games, I begin to hate losing and start studying more heavily.

Looking back to when I was a 1300 player ... I actually was pretty good... but I always did something that just made no sense ... why... I have no clue. 

I will say that the difference between a 1600 and a 1700 player is HUGE from my perspective ... 1600's are still trying weak 'traps' and missing the occasional tactic ... 1700's don't always find the right 'idea' but they always have one and they are very proficient at making it work ...

Basically... the better you get, the more you start to notice the subtle things and with that, your understanding of the game deepens.

From what I've been told, the difference between a 2000 player and a 2200 (NM), is largely openings... a 2000 knows the openings.. a 2200 knows WHY that opening is played... he see's the ideas behind it.

But again... I was only a 2000 player briefly and I"m no 2200 :).

2 months to get from 1500 to 1900?That s huge accomplishment!

It surprised me too ... only AFTER I got into the 1900's did I begin studying ... what happened to cause the jump is still a bit of a mystery :P ...

Ratings on Chess.com tend to run around 300 points lower than USCF ratings, and perhaps 200 points lower than FIDE ratings.

Players who can attain and maintain a rating on chess.com of 1950 or higher are approaching master strength, assuming they are not cheating with chess engines.

Usually, ratings should be fairly consistent across the speed of play. Thus quick = blitz = long to a first approximation.  If you find a huge differential in long vs blitz or quick, this is something to investigate.  For example if a player is 1800 quick, but can't maintian 1650 or higher in longer play, its' a sign of a tendency to be careless.

Now there are special cases of rating differential that are just plain suspicious. If a player has a 1200 rating in quick, but a 1900 rating in long, that smells like engine use.

I'm starting to think this is more difficult to do than I initially thought, but in the interest of advancing it once more step I piled ih8sens' and my rough sketches for criterion, removing all of the rating levels.

I do realize it really boils down to "the better you are, the more you notice and avoid, and the less you blunder", but I'd hoped that by creating more of a "psychological profile" of a player at X rating that it would benefit lower rated players in some way.

_____________________________________

often neglects development
regularly hangs pieces
often pushes a & h pawns and develops rooks early
sometimes misses mates-in-1
little to no knowledge of opening theory or principles
often makes uncoordinated attacks
often immediately trades pieces should the opportunity arise
sometimes hangs pieces, often sees mates-in-1.
follows opening principles
often is comfortable with one or two opening systems,
usually anticipates forks,
understands basic tactical ideas (pins, discoveries)
didn't really understand the rating system or any of the main concepts of chess. 
I knew en passant
No understanding of positional play and minimal tactical understanding.
begin to notice tactics
I very rarely missed a tactic (for or against...)
I knew very little of openings and often found myself 'losing for no reason' against class A players simply because I didn't understand how midgames take you to the endgame.
"Oh, that's why I lost" ... My tactical eye, now very well trained has left me making very few mistakes...
I begin to develop favourite openings ...
my positional understanding grows very quickly.
 "Oh well that can't be right" ... intuition now being used for good ... instead of speculative sacrifices I begin to look at my position more critically (and my opponents position more agressively)...
becoming accustomed to winning most of my games
 I actually was pretty good... but I always did something that just made no sense ... why... I have no clue.
 difference between a 1600 and a 1700 player is HUGE from my perspective ... 1600's are still trying weak 'traps' and missing the occasional tactic ... 1700's don't always find the right 'idea' but they always have one and they are very proficient at making it work ...
 From what I've been told, the difference between a 2000 player and a 2200 (NM), is largely openings... a 2000 knows the openings.. a 2200 knows WHY that opening is played... he see's the ideas behind it.
 

 1900 separates boys from men :) Hope i can achieve that too..but it s coming hard for me.

Don't forget ratings are a measure of consistency not ability. This means that, in any particular game a player may exceed or fall far below his/her actual ability. Sometimes you can get into a position where you feel comfortable and play much better than you usually do, it all depends. As Korchnoi said, "To win, you need a little luck."

Are you talking about ratings on chess.com or about real ratings? As for the ratings on chess.com I am trully amazed how the GM's can play in blitz games like kindergarden kids compared to the level of (quite) many players on chess.com

Ahhh good point, chawil. That does throw a wrench into the old gears, doesn't it? Because the measures I was thinking about are rough estimates of sometimes, often, etc, and as Immortalgamer showed us with his blindness thread, even someone as skilled as he drops his queen every once and again.

Ah well.

This is a very interesting idea, but I'm thinking that the work involved in creating an accurate "psycological profile" would be tremendous. That doesn't mean it's not worthwhile, mind you. The Panama Canal was a tremendous undertaking, too.

Brainstorming here... The results would need to be objective, so personal opinion would have to be left out as much as possible. How to do that, exactly, is a topic of debate (engine analysis?). A "sample group" of players would have to be chosen for each rating group. To achieve statistically significant results, the sample group would need to consist of (at least) hundreds of players. Let's assume 200 players per rating group.

Then you would need to "objectively analyze" (there's that debate topic again) a statistically significant number of games from each player in the group. This could mean hundreds of games per player. The more the better, obviously, but for the sake of argument let's say 200 games per player x 200 players = 40,000 games to be analyzed... objectively. Wow. Unfortunately, I have plans this weekend. How about next weekend? LOL

As you can see, this is quickly getting into the neighborhood of a federally-funded research study or perhaps a graduate (institutionally-funded) research study. It might actually prove very beneficial for the purpose of helping individuals, coaches and teams to fine-tune their studying efforts. A very interesting idea indeed.

Now, where did I put Matthew Lesko's book about getting government grants?? I know it's around here somewhere...

I do think this is an interesting idea. It may give people insight into there own game. ... even me, I think I would like to see where I stand :-)

Even Kramnik " sometimes misses mates-in-1"

Alright, well, that idea is a bust then.

                * crumbles paper and tosses it over to the trashbin *

                             * heads back to drawing board *

I have more good ideas where that one came from! I'm an Idea-machine!

/grumble grumble if artfizz had just gotten here in time we'd have a chart...

900: This player pretty much a patzer at chess.

1250: This player is pretty much a patzer at chess.

1400: This player is pretty much a patzer at chess.

1650: This player is pretty much a patzer at chess.

1900: This player is pretty much a patzer at chess.

2150: This player is pretty much a patzer at chess.

2400: This player is pretty much a patzer at chess.

2650: This player, whle predominantly a patzer, often plays a nice game.

2900: This player, while usually a patzer, often plays creatively and brilliantly.

3100: This player will be considered a patzer by the next version of the software.

The point? We're all patzers, just on different levels.

I disagree with the statement that the difference between 2000 and 2200 is openings. Especially in correspondence (turn-based), where you are free to look up openings to your heart's content.

I've played 2000 players who are excellent at a single aspect of the game, for example at attacks. I would prefer to categorize players in terms of their understanding on various dimensions. I recently beat a 2150 USCF player pretty easily just because the player insisted on sacrificing and attacking, in a position which quite obviously (even my opponent admitted this after the game) was not that type of position. But - one can get pretty highly rated being a single-dimensional player. My favorite example is Emory Tate. He can produce some excellent chess, but he'll always have a ceiling unless he develops his play along another dimension. Put another way, if you told him before the game that he couldn't sacrifice his way to victory, his performance rating would go down. That is the problem as I see it with single-dimensional players.

I'm looking at a chart right now but ...

Your idea has a lot of potential - however, simply turning into a table won't do it justice. What you seem to be on the track of is an audit tool. As you suggested early on, the factors need to be separated out and given a scale e.g.