A while back we had a brief discussion on deflation of the ratings. I just played a match with an opponent who was 7 points higher(equal) I was 1922 he was 1929. We played 2 games a draw yielded 0 change. The win gave me 10 points. In the past i would have recieved 16 points for this win. So it appears to me if a person performed just as well in a current tournament as he did in an identical tournament say 5-6 years ago, but not well enough to get bonus points, then perhaps he would not gain as many rating points resulting in a deflation? Certainly before the change in the rating system my rating would be 6 points higher after the match.
Also i did not know that it was harder for a 1900 to get points vs a 1900 opponent than for a 1500 to get points against a 1500 (unless a defect with rating calculator). I put in a 1500 winning 2 points out of 2 against an opponent with a 1500 rating. The result a 36 point gain. I did the same with a 1900 vs a 1900 in a two game match( 2 to 0). The result only 21 points gained.
What you’re seeing is the variable K factor that went into effect in 2000. The lower someone’s rating is, the higher the K factor, which affects the maximum number of points that can be earned from the game (or match).
For details I refer you to the mathematical description of the rating system on the ratings page. It is not light reading.
I figured that it was due to a change in K, however I just wasn’t aware that it varied in the current system for players under 2000.
Though thoughts on a player gaining 10 points as opposed to 16? One could, though i wouldn’t care to extrapolate that the 6 less points will also effect the result of the next opponent i would play. I’m not sure if the current system results in any net deflation of the pool. Though in this case i’m certain my rating is deflated by at least 6 (when comparing to the previous system). So i wonder if it is a greater likelihood that an Expert created today is stronger than one created 5 years ago.
Concerns and questions about the rating system formula should be sent to the USCF Ratings Committee. Committee members are listed on the Governance page of the USCF website and the committee chair is usually listed in Chess Life.
The size of the K-factor has nothing to do with inflation or deflation.
Put the shoe on the other foot. Your opponent lost 10 points instead of 16. If you had lost that game, you would have lost 10 points instead of 16.
Higher-rated players tend to change, in playing strength, more slowly than lower-rated players. So a smaller K-factor is appropriate as rating increases. It’s not deflationary.
Many chess players’ egos blind them to the possibility that they are just as likely to lose rating points as to gain them. A player who plays consistently 100 points above his rating will, eventually, gain 100 points, regardless of K-factor. Replace “above” with “below”, and “gain” with “lose”, and the statement remains true.
No doubt, though it still remains that it is harder to raise the rating, though it is possibly true that it makes it harder to lose rating points. So it still may be true that overall ratings do not increase on average at the same rate that they would have on the old system, so it could be deflationary(would need a lot of data to know for sure), one could try to argue also that it is inflationary because people are not losing points as fast as they would under the old system as well(again data is needed). The best that can be done without the necessary data on how the current system has performed in comparison to the old system is to make educated likelihood judgements.
Why is it that once again, you are constantly trying to play mind reader? Nolan the all mighty psychic who knows what others think, their beliefs, religious, and political affiliations, and shoe size. I didn’t say one way or the other what the results would be. I didn’t even say which i thought was more likely, and i definitely didn’t say i wanted it to be one way or the other. If the comittee has done an actual comparisson of how the old rating system actually performed vs how the current system has performed up to recent times, i’d be impressed especially because it would take some time for the population to change significantly enough to show whether a significant inflation or deflation has occurred based upon the current system. I certainly would look at the conclusions of such a study and how those conclusions were drawn if the information was readily and easily available. It is ridiculous for anyone to want to, or not want to agree with conclusions in the absence of evidence. Perhaps you want to, or do not want to agree to things based upon emotions, i don’t know. I do know, that i certainly do not. In any case given the argumentative and irrationally presumptuous nature of your previous responses and behavior i will no longer concern myself with your oppinions.
Nearly every question you have asked here about rating has displayed an attitude which suggests you know more than the USCF about how ratings have been or should be computed.
That just patently is not the case. The PhD mathematicians on the USCF Ratings Committee have spent hundreds of hours analyzing ratings, working with revisions to the formula to predict their impact and then tracking the results of those revisions.
Why not come to the ratings workshop in Phoenix in Arizona and listen to the experts?
Classic tactic when wrong try to redirect to some other angle(Nolan tactic #1)Change the subject(Your above fallacious post, has nothing to do with the last two posts you wrote). It is clear to anyone who reads the above that you are just trying to make an excuse for your behavior and provide a cover up. People can read above and see that absolutely no claim that you are suggesting has been made and no such claim has been made in any post. Your ability to reason is completely limited. When you can’t dispute you move on to tactic #2 let me make up something to attack and then say opponent said it then attack it(your most beloved method). When that doesn’t work Nolan goes to tactic #3 “claim somebody smarter than Nolan” would think nolan is right A fallacious attempt at Toullman logic called appeal to authority. Which is especially ridiculous in this particular post as no claim in either direction regarding the question has been made in regards to there being inflation or deflation in the pool. The fact that you have been caught in the wrong and you are ashamed does not bother me. The old Switcharoo Nolan. I have only responded this last time to you, because apparently as usual you fail to read. So let me point it out to you again. I will not concern myself with your fallacious, absurd irrational presumptuous opinions. Another piece of advice stop assuming what you do not know. Though i am not a PHD i do hold 4 Advanced degrees two of them in biological sciences where i have done EXTENSIVE statistical and analytical work regarding changes in populations. Though again irrelevant in the sense that I have made no claims or done anything that you have made out. Purposeful in that a person does not have to have a PHD or any degree to be correct in what they have stated. Finally you do not even know whether the author of the system, assumes that there will be some inflation or deflation or not. You are just assuming again! Making a system more accurate(a better rating system) is supposed to cause a deflation or an inflation (to at least to some aspect) to the more correct figure! If the numbers were right in the first place there would be no need to change the system. You are the person who wants to come in and make bold claims that a person is wrong or right amazingly when they haven’t even made any claims. I see that you have severe weakness in some aspect of emotion or mentality or both. So again i will respond to you no further and will avert my eyes from anything you write.
[b]Nolan tactical play book and guide to obfuscation:
Change the subject
Put words in opponents mouth and hope people believe he said it
Appeal to a unnamed higher authority and then tell people what the higher authority would think and that the higher authority would agree with Nolan[/b]
I have read his letters to you in this place, and i can not agree with you. He has expressed no attitude at all in this posts, I see only that you have began attacking him reason i know not? When he does posts it almost always seems to be you that begins attacking him even if he was not speaking to you. Further he has said or asked nothing that was unreasonable. I sincerely hope that i am misinterpreting your attitude an purposes, because if the moderator needs to be moderated, this is a bad place.
Here again, so strange as this quote by you does not relate or have any validity to what he said i too would be very upset, and it seems as if it was your purpose, because i can see no point for writing it especially as i see it is not even true. It is very silly to be imagining voice tones in written words as well.
The rating of a person does mean much from one era to the next. As the rating system has been changed and changed again, from one executive board to the next. The ratings of the players from class C down, has had some type of inflation in the past twenty years. It would be harder for the USCF having inflation for ratings above 1800 and more so over 2200. During the 1970’s, nobody could ever be dreaming of players over 2700 USCF. If the K factor was equal for everyone, we could be seeing players in the 2800 and 3000 range.
If the USCF did not have the Swiss System, there would not be a rational reason to have a rating. As a Round Robin, everyone plays everyone, so why does a Round Robin needs a rating? The rating of a player, even in a small event can make sure they would not be able to be paired up with each other in the first round. The rating only helps the director not make the pairings as a random order of players.
There are members out there, that go over old data of ratings. I do not care if the ratings have had inflation or deflation, of the past 30 years or the past few years. I only care of what the rating is now, as the ratings of the players makes the swiss pairings less random. Can say the ratings of the players in 2005 (under 1800) do have some level of inflation to the players in 1975. Also can say the dollar has had a great deal of inflation since 1975.
Hey Td forsythe thanks for your constructive input. Even if we had only round robbins i think the ratings would still be important from a psychological and sporting perspective. Chess in the US only began a really rapid increase after the advent of the system. It gave people something they could qauntify, and say “hey i’m better than mickey bob or harry ”. So you think most class players today are weaker than those in the past? I would have guessed the opposite. 12 years ago, I was talking to two former experts, one had been state champion several times, and neither one of them as amazining as it sounds knew that you could mate with two bishops vs a king. Don’t get me wrong, they were both pretty strong, i’d just guess that since information(books, tapes videoes, instruction) is more readily available, that todays class player might be likelelier to be stronger than one from the 70s. There were other factors though such as the Fischer Boom, and smaller populations and maybe even more isolated populations, perhaps some unsuspected forms of bottlenecking which could potentially have made the average class player stronger though like you suggest. Interesting.
Think of it this way, when starting out in the 1980’s. During the 1980’s, the rating theory was if two players play each other for their very first game, the rating was based on 1000. Today it is based on the players age. Lower for scholastics and higher for adults. In 2005, if two adults play their first games with each other, the based on rating is 1300.
When starting out as a scholastic player in the early 1980’s, the worst rating I ever had as a published rating was 1032. The more I look at it, I should have been around 800 in todays standards. Scholastic ratings are much lower now then they were 20 years ago. Then again adults starting out are a little bit stronger.
The other factor that helps the class player get a higher rating, is the USCF’s idea of having a rating floor. If a player was a low class B player during the Fischer era, with never going over 1800. After 30 years the skill level of the player is not what it was before. Then again the rating floor of the player is 1500, but the player should be at 1300 - 1400. If some plays chess with a person below their rating floor, it will make that class player have a higher rating. If a member at their rating floor of 1500, with the skill level of a class D player. If that 1500 player is active, you will be seeing a lot of class D players with a class C rating.
Just a few ideas why the class players have a little higher rating!
I think this shows that, contrary to popular belief, the ratings system is NOT inflationary, at least in terms of the initial rating for new adult USCF members.
I think the ‘1300’ starting (pre-event) rating part of the formula came in during the 2000 revisions. If anything, that seems to have lent some stability to the initial ratings for adults.
The norm for adult chess players starting out has been in the class D/E range for years. Seeing the data in the early 1990’s with a high class D rating, the numbers just do not look right.
Having a rating is nice, it still does not make you more or less then what you are. Having a rating is a part of me, but I am more then the sum of my parts.
The reduced K is actually inflationary, and a recent tournament of mine explains why. We had a quad last friday where the ratings of the participants were 1600, 1650, 1915 (me), and 2050.
The 1600 won the tournament with 2.5 points
The 1650 had 2 points.
I had 1 point.
The expert had 1 point.
As terrible as our performances were, me and the expert will lose less than 50 points combined, while the 1600 is going to gain 100 points, and the 1650 is going to gain 50 points. The average rating of the participants in that tournament has increased by 25 points. What we have learned here is that the rating system is going to inflate when lower rated players are taking points from higher rated players, and it will deflate when it is the other way around. Since lower rated players tend to improve faster than higher rated players, it seems that the formula would at least reduce the deflationary effect of the learning curve.
In a way it’s not surprising that the average rating goes up when the total score is a half a point more that is possible in 6 games.
But the important thing to remember is that the old rating system was zero sum (for ratings less than 2200 or 2000 or something …). When I played you, the sum of our rating change was zero. If I gained 10 points, you lost 10 points.
All that is completely different. It’s like when the country went off the gold standard - things can float a lot more now. What the rating committee is (I believe) trying to do is allow players to see a rating that reflects their results more quickly. So … if everyone starts out young and bad and gets good quickly then the rating system is inflationary. And then they stop playing and they take all their points with them and that’s deflationary.
It’s a difficult task - there are a couple of assumptions … The K varies by age mainly (not really number of games). An expert’s 2000 rating is usually based on a lot more games and so a bad result is probably just that - one bad result that won’t be repeated. The assumption is that a youngster’s GOOD result isn’t a fluke, but reflects a rapid increase in actual playing strenght. So his rating should be higher. And POOF! It becomes higher.
It does create more work for th rating committee because they have to examine these trends more closely … together with the assumptions in the new rating system - to ensure some measure of stability between generations. But they are capable of doing that …
When I was younger (35 years ago) ratings started off a lot higher. But we started off a lot older too. I’m a little surprised at the statistics Mike quoted, but there are again two possible explanations for it (at least). 1 is that the system is actually quite deflationary (assuming the average strength of new adult players has remained the same). 2 is that many more adult players (like parents of kids) are actually playing tournament chess than before … and they would be weaker.
Every version of the ratings system I recall having seen since 1986 had a bonus formula in it, so it was not a zero-sum game.
The Ratings Committee’s report in the 2005 Delegates Call discusses some of their research, and it does address the issue of whether the current system is inflationary or deflationary.
I think the report may be available in the governance section of the USCF website.
Yes it did, and when the K-factors were different it might not have been either.
However … the way most people would understand the old system - and the way it effected most people - it was zero sum.
And now it is never going to be (almost), and folks should understand that the old simplified way we all understood it does not work. And so the way we all understood inflationary and deflationary is not helpful either.
Thanks for the clarification (and implementing all this stuff!)
(when comparing to the previous system). So i wonder if it is a greater