My point remains the same. You cannot produce any conclusions on performance from the 3 victories. You are assuming a playing strength of 2450 to produce a performance of 2450. It's circular.IM Jack Rudd wrote:If the answers to "would a 2450 strength player (playing exactly according to his strength) have made this or a lower score"? and "was your performance of at least 2450 strength?" are different, this suggests something may be awry with your concepts.Richard Bates wrote:Norms are not about determing how good one is. They are about determing how well one has performed over a set number of games. Your suggestion is twisting the current system to change a norm from the latter (establishing the level of a performance) to the former (asking "how would a 2450 strength player have performed"?)IM Jack Rudd wrote:
And his performance over nine games? If it's lower than 2450, then the underlying model strikes me as very dubious. "We came up with an estimate of your strength, but now that you've played a few more games and won them all, we've decided you're not as good as we thought you were."
The trouble with determining how well one has performed over a set number of games is that the FIDE rating system itself does not give you the means to do that. The mechanics of the rating system do not define a TPR, nor give you an easy means of determining it for a given set of opponents and scores.
What they do define is an expected score; a 2450 playing the set of opponents you generated would have an expected score of (0.5*6)+(0.92*3)=5.76 over those nine games, and so would make a minor rating gain (almost certainly 2.4 rating points) against it if he scored 6. Against a field of nine opponents rated 2317, his expected score is presumably more than 6 (hence the norm target of 6.5), and so he'd lose rating points for scoring 6 against those opponents.
A few years ago, the rating system assumed all your opponents in a tournament were of the strength of your mean opponent, and the consequence was that it was possible to get tournaments in which, for any given game, you either drew with a player of at least your strength or won, and yet lost rating points in the tournament overall. It has now thankfully moved away from this assumption; perhaps it is time for the TPR calculations to do the same.
For the purposes of a WIM norm the three victories are assumed to produce a performance of "at least" 2200. For the purposes of an IM norm the three victories are assumed to produce a performance of "at least" 2450. For the purposes of a GM norm the three victories are assumed to produce a performance of "at least" 2600.
Whatever, the basic argument is that a tournament organiser can manipulate the composition of the tournament (to produce appropriate number of foreign players etc) by including a load of people whose only purpose is to lose to the norm seekers and reduce the effective norm requirements to be over a reduced number of games. It's just another stage in bringing the whole title system into disrepute.