Average Opponent Rating

I have played something close to 800 games in long chess since I joined this website between 3-4 months ago. When I joined I was an absolute beginner and had a rating of about 700, so the majority of my games were playing people with ratings of 600-800. It took me a few months to hit 1000 and the bulk of my games are played under the rating of 1100.

At the moment I am hovering around 1250-1300 and play nobody below 1200 (in long, in blitz I play anyone from 1100+) but my average opponent rating is 1106, even though I have played people above 1200 for more than 30 games.

I suggest the average opponent rating be decided on your last 10-20 rated games, it would give a more accurate portrayal of the level of opponents you are actually playing.

10-20 is probably a bit extreme... but I think it would be nice idea for it to  collapse to the present in some way. 

they could have overal and of last 20

Whoops, I said 800 games, I have actually only played about 450 >.>.

There are players who are rated 300 points above their average opponet.

Why change the number--it reflects exactly what it is supposed to reflect, that is the average of your opponets.

If there was to be a "recent average" for example, I would suggest it be based on no less than 100 games and would be maintained as a different stat.

It can be extremely misleading about the skill of a player.

How so? I thought their rating relfected the skill of the player, not the average rating of your opponents. I know personally, I gauge my moves by THEIR rating, not the rating of people they've played against.

My average opponent is rated much lower than me, I have been playing for a while, and I stayed on one level for a longer time. But the number should still reflect what it tells you.
The ratings of the ongoing games and recent ones are shown on the page as well.

Kupov> It can be extremely misleading about the skill of a player.

I know your skill by looking at your rating, not your average opponent's rating.

(If I was really curious about your strength, I would also check your RD).

 Not so, if for example I am rated 1500 after playing 50 straight games won vs people under the rating of 1100, this only proves that I can regularly beat players under 1100.

The average rating and best win likely say more about your level of play to an observer than your actual rating.

Kupov> I am rated 1500 after playing 50 straight games won vs people under the rating of 1100

Glicko doesn't work that way--that record wouldn't even make you a 1500.

• If you score 100% in 50 games against 1000-players your rating's ~1400.
• If you score 90% in 50 games against 1000-players your rating's ~1300.
• If you score 90% in 50 games against 800-players your rating's ~1200.

(Specifics: player rating:1200, rd:350. Opponent rd:70.)

Kupov this only proves that I can regularly beat players under 1100.

If you could beat 1000-players in 50/50 games, you've proven you're much stronger than 1000 and will be rated accordingly. Glicko guesses 1400... you won't go up much higher unless and until you beat stronger players.

For provisional ratings, I like Glicko more than ICC or USCF. In those systems, you can lose points for beating weaker players! For normal ratings I like them equally.

Not that there aren't rating anomalies. For example, Erik Bloodgood.

Kupov> The average rating and best win likely say more about your level of play to an observer than your actual rating.

To me, "best win" says nothing about a player's strength. If you play stronger players often enough eventually one will time out or blunder.

"Average opponent" does say something, maybe about potential, maybe about how much someone gives back, not so much about current playing strength. I'm curious... how do others evaluate or make use of "average opponent"?

I think any decent club player would tell you that playing lower rated players is bad for their rating.  You basically have to win every game just to stand still.