I was looking at the fancy new feature on "my home" page, and noticed that my average opponent rating is more than 200 points higher than my current rating, and 70 points higher than my highest rating (90 days).
This peaked my interest a bit, and I've been checking out some other players and noticed some are the other direction. They have ratings significantly higher than their average opponent rating. kinda makes ya think.
Originally posted by duecerI suppose.
I was looking at the fancy new feature on "my home" page, and noticed that my average opponent rating is more than 200 points higher than my current rating, and 70 points higher than my highest rating (90 days).
This peaked my interest a bit, and I've been checking out some other players and noticed some are the other direction. They have ratings significantly higher than their average opponent rating. kinda makes ya think.
But don't read too much into it. I for example have entered a lot of tournaments and been matched with lower rated opposition. I play a trickle of higher ranked opponents too. In the end, the rating averages out. You keep playing players ranked well above you, and presumably losing most of them, meaning your lower comparative rating is probably accurate.
I might play lots of players ranked below me, but as long as I beat them always, it means my higher rating is accurate.
As soon as you win some of yours against better players, and I lose some of mine against lesser player, our rating averages to reflect it.
You can't argue with a rating, even if you manipulate your choice of opposition. Because if you keep playing way above your ability, you will never get any rating points, and if I keep playing way below mine, I wont get any rating points. For our ratings to change, we will have to pick closer opponents, and then our ratings will be calibrated.
😉
I'm not sure if it necessarily tells much, as open tournaments (and even banded ones) always give you loads of lower rated opponents. you need to select opponents very carefully to get a high opponent rating.
and still, for improving players the averages always drags behind. the average of my current opponents is 2045, yet the profile stats are 1935 (90d), 1842 (1y) and 1589 (5y). on the surface the profile stats might look like I've played mostly against lower rated opponents, but in reality I've deliberately sought to play stronger opponents for several years already.
I am certainly not saying the gradings are wrong, but I think i have acheived my grade without actaully beating anyone of my own grade or above.
Given that my record is losing against people of my own grade, it suggests I am better at putting away the lower graded players.
I do like a good end game which is where i win many of my matches but suspect that I fail in the opening and middlegames too many times against my own grade level to allow better ending to do more than salvage the odd draw.
Originally posted by PolicestateI'm not sure about that. right now you're close to 1900. what's the chance that you'll win less than 95 games in 100 games against beginners(1200-1400)?
You can't argue with a rating, even if you manipulate your choice of opposition. Because if you keep playing way above your ability, you will never get any rating points, and if I keep playing way below mine, I wont get any rating points. For our ratings to change, we will have to pick closer opponents, and then our ratings will be calibrated.
😉
will those 5 games you've drawn or lost pull your rating down enough so that your rating won't skyrocket up to 2100?
I'm not arguing, just curious.
Originally posted by philidor positionIt would be a stretch to win 95/100 games from players in the 1200-1400 range for a 1900 player... And besides, they would only get 1 or 2 points from winning, but lose say 20 points for losing. So even if they win 90% of the games they still lose 200 but gain 180 (I'm making up the numbers but this is the concept).
I'm not sure about that. right now you're close to 1900. what's the chance that you'll win less than 95 games in 100 games against beginners(1200-1400)?
will those 5 games you've drawn or lost pull your rating down enough so that your rating won't skyrocket up to 2100?
I'm not arguing, just curious.
Also many of the 1200-1400 players would be underrated. So they are not getting the benefit of winning against others who are overrated (relative to their rating range).
If I'm 1900 the last thing I'd want to do is play 100 players in the 1200-1400 range, it would do nothing for the 1900 player.
Originally posted by philidor positionWith my current rating of 1886, I would only score zero rating points for beating anyone below 1166. I would gain a single rating point for beating anyone rated 1167 - 1362, and two points for a win against a 1363 - 1457.
I'm not sure about that. right now you're close to 1900. what's the chance that you'll win less than 95 games in 100 games against beginners(1200-1400)?
will those 5 games you've drawn or lost pull your rating down enough so that your rating won't skyrocket up to 2100?
I'm not arguing, just curious.
Losing to anyone <1166 loses me 32 rating points. I lose 31 in the next category, and 30 for the 1363 - 1457 rating (all based on a recent thread on rating changes).
So if I won 95 games against a mix of 1200 - 1400 opponents, I would put on between 95 and 190 rating points. Losing the other 5 against opponents of this standard, I would lose between 155 and 160 points.
So in most cases I would have a lower rating by the end in the example you give, or a slightly inflated one if all those 95 opponents were over 1363.
This makes the point that you can't reach 2100 by playing people far below your playing strength alone. The rating formula is designed for this reason.
Ed
Originally posted by SpacetimeSorry didn't read this post - same point.
It would be a stretch to win 95/100 games from players in the 1200-1400 range for a 1900 player... And besides, they would only get 1 or 2 points from winning, but lose say 20 points for losing. So even if they win 90% of the games they still lose 200 but gain 180 (I'm making up the numbers but this is the concept).
Also many of the 1200-1400 players w ...[text shortened]... nt to do is play 100 players in the 1200-1400 range, it would do nothing for the 1900 player.
Sure you can play people 200 pts better than you, but you probably don't have a good winnning ratio as compared to others. I think of it as if you can beat the majority of 1600 players, but lose a good amount of games to 1800 players then you are 1700. Maybe if you played opponents that weren't so high your rating would be higher.
After 1100 games your still at 1200? That doesn't sound right. You should make sure to learn from each game rather than play hundreds of games and make the same mistakes.
Subsequently you should play 1300 players for a while. When you win more than 70% of those games then play 1400 players until once again you can beat them 75% of the time. Playing people 200 pts above you doesn't necessarily teach you anything. I can play grandmasters all day, but I might not get any better because they are just that much better than I am.
Originally posted by PolicestateOK, thanks spacetime and policestate for making it clear to me. I didn't know the math.
With my current rating of 1886, I would only score zero rating points for beating anyone below 1166. I would gain a single rating point for beating anyone rated 1167 - 1362, and two points for a win against a 1363 - 1457.
Losing to anyone <1166 loses me 32 rating points. I lose 31 in the next category, and 30 for the 1363 - 1457 rating (all based on a r ...[text shortened]... le far below your playing strength alone. The rating formula is designed for this reason.
Ed
Interesting discussion. I use a player's rating, his opponent's average rating, and the recent slope of the green rating graph in assessing the true strength of an opponent. An unusual number of green graphs appear as if the player has jumped off a cliff, which means he has either been very busy, ill, manipulating the rating system or whatever, and timed out or given up in a lot of games. I recognize that if the player's rating is 200 points above that of his average opponent, he plays lower rated players on average, and leads me to believe that his rating may inflate his real ability. I would find it useful if the average rating of opponents was broken down into the average rating of those that he won, and those that he lost against
Originally posted by wormwoodnot if you just don't understand what is going on. Why did the GM play what they did? I think it is better to play opponents that are slightly better, you will loose most, but you will find special joy in those you win, all the while you are learning quick. I hate to play those my level, I loose more to players my level than to slightly better players somehow. They just don't understand chess, and that makes me loose interest for those games, I wish they would be good enough for me to defeat them. Weird huh?
you absolutely would. the only problem playing against superior opponents is getting beaten a lot. but if you can take it, you'll also learn the fastest.
Originally posted by orion25the problem is your lack of self-discipline, not their lack of skill. the same applies if you piss away the chance of learning from a master.
not if you just don't understand what is going on. Why did the GM play what they did? I think it is better to play opponents that are slightly better, you will loose most, but you will find special joy in those you win, all the while you are learning quick. I hate to play those my level, I loose more to players my level than to slightly better players someho e interest for those games, I wish they would be good enough for me to defeat them. Weird huh?
doing badly against lower rated opponents also suggest your technique is probably lacking. meaning you can't do on autopilot the things required to win a won game on your level.
but, of course, if getting beaten a lot tends to put you off from chess, it's extremely important to avoid putting yourself in that position. because once you lose motivation, there's absolutely no chance of doing the work required for improvement.
Originally posted by kmac27I think it helps, i.e. play a 2500, then play a 2300, then a 2100, then a 1900, then a 1700 etc. etc.. You just "feel" the difference - the lower rated the player is, the more leeway they WILL give you. I.e. after the first two games, the 2100 will already seem pretty weak. 😀
Sure you can play people 200 pts better than you, but you probably don't have a good winnning ratio as compared to others. I think of it as if you can beat the majority of 1600 players, but lose a good amount of games to 1800 players then you are 1700. Maybe if you played opponents that weren't so high your rating would be higher.
After 1100 games your sti ...[text shortened]... rs all day, but I might not get any better because they are just that much better than I am.
It can be compared to weight lifting, if you move heavy objects first, and then less heavy - but still heavy - objects, you will notice the moderately heavy objects seem pretty light.