- Joined
- 06 Dec 01
- Moves
- 77129
hmm, I need a good explanation for the inflation of this rating system. Are peopel just getting better? Or is it just the fact that more and more peoepl join the site or what? Also, economically speaking, because of the rating inflation, I need someone to posit whether or not the inflation causes a higher equilibrium when you graph the demand of playing on here versus the time spent. If possible, maybe some statistics on what the ratings were back then compared to now, and some extrapolation as to what they will be in the near future, far future.
Originally posted by ZumdahlThis crops up every now and again. I don't think that ratings inflation really exists as such. There's no reason for it to, except that as the number of people on the site expands (assuming that it does) there's more scope for ratings to grow, as the provisional ratings system means that average ratings aren't stuck at 1200. Quite a lot of peoples ratings go up and down quite a lot, so that may also help give the impression of ratings inflation. All that needs to happen is someone to beat 5 players with the same rating as them for their rating to go up by 80 points.
hmm, I need a good explanation for the inflation of this rating system. Are peopel just getting better? Or is it just the fact that more and more peoepl join the site or what? Also, economically speaking, because of the rating inflation, I need someone to posit whether or not the inflation causes a higher equilibrium when you graph the demand of playing on h ...[text shortened]... compared to now, and some extrapolation as to what they will be in the near future, far future.
Originally posted by XanthosNZXanthosNZ is exactly right.
Average rating is remaining fairly constant among active players but spread is increasing.
People sporadically claim of rating inflation, and support it by pointing out how some time ago there was nobody over 2200, or by counting how many players were in each rating band.
I have some data, yes it's less than 5 months worth, but it supports this. About once per week I've recorded the rating that correspond to 50th, 80th, 90th and 95th percentile on the "Player Tables". You can see that over the time period the size of this list (non-provisional players active in past 100 days) has grown steadily from 6154 to 7925 (28 percent growth). The rating corresponding to each percentile however has remained virtually unchanged.
Date NumPlayers 50th 80th 90th 95th
3/20/2005 6154 1291 1478 1591 1703
3/27/2005 6225 1290 1478 1593 1703
4/2/2005 6310 1295 1479 1591 1700
4/15/2005 6526 1289 1477 1594 1699
4/23/2005 6659 1289 1474 1594 1692
5/2/2005 6750 1290 1475 1592 1697
5/7/2005 6825 1290 1476 1593 1697
5/16/2005 6958 1290 1474 1591 1701
5/22/2005 7034 1290 1474 1590 1701
5/30/2005 7150 1289 1474 1589 1702
6/5/2005 7237 1289 1474 1588 1699
6/13/2005 7323 1289 1475 1588 1700
6/22/2005 7445 1289 1475 1585 1699
7/2/2005 7573 1290 1473 1586 1696
7/12/2005 7672 1290 1472 1584 1698
7/29/2005 7845 1289 1474 1586 1700
8/9/2005 7925 1290 1473 1588 1698
I would expect that the extremes will continue to "stretch", but the basic distribution will remain the same. As such a 1700 rating should continue to put you into the top 5 percent, even as ironman's rating grows, and the number of active, non-provisional RHP-ers grow.
edit: is there some magic I could have done to format this table?
Date NumPlayers 50th 80th 90th 95th
03/20/2005 6154 1291 1478 1591 1703
03/27/2005 6225 1290 1478 1593 1703
04/02/2005 6310 1295 1479 1591 1700
04/15/2005 6526 1289 1477 1594 1699
04/23/2005 6659 1289 1474 1594 1692
05/02/2005 6750 1290 1475 1592 1697
05/07/2005 6825 1290 1476 1593 1697
05/16/2005 6958 1290 1474 1591 1701
05/22/2005 7034 1290 1474 1590 1701
05/30/2005 7150 1289 1474 1589 1702
06/05/2005 7237 1289 1474 1588 1699
06/13/2005 7323 1289 1475 1588 1700
06/22/2005 7445 1289 1475 1585 1699
07/02/2005 7573 1290 1473 1586 1696
07/12/2005 7672 1290 1472 1584 1698
07/29/2005 7845 1289 1474 1586 1700
08/09/2005 7925 1290 1473 1588 1698
(Just messing with your format to make line up)
Originally posted by Zumdahli see your point! ratings do build here like cash in the bank. some players only play opponents they know they can beat.yes they would have a high rating by that system . their rating wouldnt amount to much against a strong player. you joined in dec 01 so you may remember xeno. ill say no more.
hmm, I need a good explanation for the inflation of this rating system. Are peopel just getting better? Or is it just the fact that more and more peoepl join the site or what? Also, economically speaking, because of the rating inflation, I need someone to posit whether or not the inflation causes a higher equilibrium when you graph the demand of playing on h ...[text shortened]... compared to now, and some extrapolation as to what they will be in the near future, far future.
Originally posted by xlchainsawWhat?
i see your point! ratings do build here like cash in the bank. some players only play opponents they know they can beat.yes they would have a high rating by that system . their rating wouldnt amount to much against a strong player. you joined in dec 01 so you may remember xeno. ill say no more.
Originally posted by Phlabibit
You guys broke the General forum... everyone has moved over to debates and spirits.
P-
Nothing beats a good arguement while pissed..
Aye.. these are the good moments.. as for Inflation buddy, I keep debating the same thing about my beerbelly..