< Earlier Kibitzing · PAGE 4 OF 4 ·
Later Kibitzing> |
Aug-10-11 | | parmetd: there is no way Esserman had a 2600 performance in this event. He didn't even play enough high rated players for this to be possible. |
|
Aug-10-11 | | Kinghunt: <parmetd> To the contrary, it was strong enough. He scored 2/4 against four GMs: Nakamura (2770), Ramirez (2592), Lenderman (2525), and van Wely (2683). He then scored 5/5 against non-GMs: Mike Webb (1871 USCF), Michael Fitch (2028 USCF), Lorena Zapeda (2175 USCF), Robert Perez (2384 USCF / 2210 FIDE), and Jim Dean (2336/ 2272 FIDE). That's a 2643 performance against the GMs, and a perfect score against everyone else. As for exact GM norm calculation... Keep in mind that for a GM norm, the lowest rated opponent's rating is considered to be 2200 (though a maximum of one rating is changed). If you make that change on one player's rating, use the FIDE ratings on all players with them, and the UCSF ratings for the rest, his average opponent rating was 2384. +5 in 9 rounds against 2384 opposition is a 2606 performance. So, yes, the performance was good enough, it's just that a lot of other requirements weren't met. |
|
Aug-10-11 | | parmetd: Incorrect, the lowest rated opponent is considered *2100* |
|
Aug-10-11 | | Kinghunt: <parmetd: Incorrect, the lowest rated opponent is considered *2100*> I suggest you look at the FIDE title handbook (which you can find here: http://www.fide.com/fide/handbook.h...), section 1.46c, where it clearly states the following: <For the purposes of norms, the minimum rating (adjusted rating floor) for the opponents shall be as follows:
Grandmaster 2200> |
|
Aug-10-11 | | frogbert: kinghunt, the only problem might be that not all of naka's opponents are fide-rated. they have to be <fide>-rated <prior> to the event in order to count towards the number of players/rounds needed and the tpr. otherwise you're right wrt the 2100/2200 question, of course. |
|
Aug-10-11 | | frogbert: <it was quickly fixed after my notice - but i didn't even get a reply ...> maybe someone's reading this! today i got a belated, earnest "thank you" via email. i was also informed that the correction was twittered, mentioning me as the source. so i must apologise for thinking that no credit was given - i must admit that i don't follow 2700chess.com's tweets. :o) |
|
Aug-10-11 | | parmetd: Re: 2100, ah they changed it! My apologies. I must admit I can't be bothered to follow all the rule changes fide institutes. And yes Michael Fitch is *not* fide rated. Also he only had 2 foreign players (Zapeda and Van Wely). Re: frogbert, we were talking about eseerman not naka ;) |
|
Aug-10-11 | | frogbert: i looked up fide-ratings:
mike k webb (1954): 1966
michael fitz (1981): not rated
lorena zapeda: not rated
<If you make that change on one player's rating, use the FIDE ratings on all players with them, and the USCF ratings for the rest> nope, fide-unrated players are calculated differently - but they do count, with certain limitations (so i was wrong saying they don't count at all). first, in a 9 round event <a maximum of 2 players can be fide-unrated>. (see paragraph/tables in 1.49). because webb has a fide-rating, this is ok. however consider this: <Unrated opponents not covered by 1.46c shall be considered to be rated at the rating floor level. See FIDE rating regulations for the current level of the floor.> hence, assuming that we can choose to let an unrated player count as 2200, the 3 players mentioned above will be counted as 1200 (rating floor)
1966
2200
in addition to
2770
2683
2592
2525
2272
2210
ra = 2269, score = 7/9 (78%), tpr = 2489 (for norm purposes) this also follows (indirectly) from paragraph 1.48a:
<1.48a The minimum average ratings Ra of the opponents are as follows:
GM 2380, WGM 2180, IM 2230, WIM 2030.>
in other words, it was (another) im norm, technically speaking. |
|
Aug-10-11 | | frogbert: <we were talking about eseerman not naka ;)> i know, it was just a silly fingerfehler. :o) |
|
Aug-10-11 | | frogbert: <michael fitz (1981): not rated > sorry, michael fitch. i did look up fitch, not fitz btw! |
|
Aug-10-11 | | parmetd: very interesting frogbert, thank you for posting that! |
|
Aug-10-11 | | bartonlaos: Is this common - the creation of an arbitrary rating for a non-FIDE player - and has anyone investigated whether it affects rating inflation? I'm thinking of Loek who competes in these island opens, smashing the locals while bumping his rating up to 2700chess, etc. |
|
Aug-10-11 | | frogbert: also note 1.44a btw:
<1.44a A maximum of 3/5 of the opponents may come from the applicant’s federation and a maximum of 2/3 of the opponents from one federation.> 3/5 of 9, rounding downwards (as specified in 1.44) is 5, while 7 of eseerman's opponents represent the us federation. it also violates the 2/3 of opponents from one federation (7 is more than 6). the only exceptions to this requirement are mentioned in 1.43 a-e, where the only that might apply is <Swiss System tournaments in which the competitors include at least 20 FIDE Rated players, not from the host federation, from at least 3 federations, at least 10 of whom hold GM, IM, WGM or WIM titles. Otherwise 1.44 applies.> i guess there were enough fide-rated players from enough foreign federations, but i didn't check if there was enough title holders. let's see... i found only 9 fide-titled foreign players, but i noticed why i didn't find zapeda - her name is zepeda! so she's fide-rated, with a fide-rating of 2187. hence the corrected list of ratings are: 2770
2683
2592
2525
2272
2210
2200 (unrated, "gm-norm rule")
2187
1966
ra = 2378, score 78%, tpr = 2598
hence, just 3 points shy of the required 2601 tpr. however, if my count of foreign title-holders were right, it still doesn't matter. |
|
Aug-10-11 | | frogbert: < I'm thinking of Loek who competes in these island opens, smashing the locals while bumping his rating up to 2700chess> bartonlaos, unrated players are completely irrelevant for loek's (or any gm's) events - people don't <get> these ratings they are <technically assigned> for the norm calculations, they are only used for norm tpr calculations. if you play a fide-unrated player, it doesn't count towards your fide-rating at all. under special circumstances it can count towards a fide-norm. |
|
Aug-10-11 | | parmetd: there was definitely not 20 foreign players in the US Open. |
|
Aug-10-11 | | frogbert: these are the foreign, relevant title holders i found: GM Kacheishvili, Gior GEO
GM Zapata, Alonso COL
GM Sundararajan, Kidambi IND
GM Van Wely, Loek NED
GM Hansen, Lars Bo DEN
GM Sadorra, Julio C PHI
IM Battaglini, Gabriel FRA
WIM Zepeda, Lorena Marisela ESA
i.e. only 8 foreign title holders (i erroneously counted a wfm first time around). the count might be wrong, as i noticed a couple of norwegian players (without titles) listed under us states. whether this applies to some of the title holders too, is beyond me. but unless such players exist in the line-up, esserman also faced too many us players and too few foreign in order to make a norm, <no matter his tpr> or fulfilling other norm requirements. |
|
Aug-10-11 | | frogbert: <there was definitely not 20 foreign players in the US Open> don't be too sure of that, parmetd. aryan tari (ny) and johan salomon (al) are both norwegian and members of the norwegian federation. :o) |
|
Aug-10-11 | | frogbert: but the short summary of the esserman norm question is rather clear: according to the rules in the handbook, it was not a gm norm. |
|
Aug-10-11 | | parmetd: you missed Tamaz Gelashivili who is also GEO |
|
Aug-10-11 | | frogbert: and timur gareev (with correct spelling) is from uzbekistan, so there were indeed (at least) 10 titled non-us players. so, if one by further examination manages to find a total of 20 non-us players in the line-up, we can conclude that esserman missed out on the gm-norm by facing an average two points lower than the requirement. tough luck. |
|
Aug-10-11 | | bartonlaos: <frogbert> Okay, it was confusing to me because the fide ratings are listed for Nakamura's opponents in international tournaments, yet appear to affect his ratings - for reasons of hierarchy I suppose. example, Bazna Kings on USCF
http://main.uschess.org/msa/XtblMai... |
|
Aug-10-11 | | frogbert: so the list is:
10 previous titleholders
2 norwegians
WFM Vazquez Maccarini PUR
GM Garcia, Gildardo COL
GM Mitkov, Nikola MKD
NM Kustanovich, Michael ISR
FM Berrios, Gabriel Angell PUR
NM Berisha, Enver SRB
Aird, Ian SCO
Angel, Antonio COL
well, that's 20 (and probably more that i don't bother to search for). more than half of them were listed by us states instead of fide federations. |
|
Aug-10-11 | | frogbert: <it was confusing to me because the fide ratings are listed for Nakamura's opponents in international tournaments, yet appear to affect his ratings> now you're talking about <national> rating, like the uscf right? what national federations choose to do with their own rating systems is up to them, but norway does more or less the same thing as in usa: international tournaments played abroad are also rated in the norwegian system, using fide ratings for players without norwegan ratings. in principle it's completely meaningless and utter nonsense to mix ratings and players from different rating pools like this, but the justification in norway is that doing this helps against the systemic <deflation> in the norwegian rating system (yep, we've got the other "problem"). norwegian ratings are generally lower than fide ratings (for basically everyone below 2200-2250 in fide), so the net effect is to inject rating points into the norwegian system. consider the assymetry when a norwegian player (underrated, according to the above) wins or draws against a higher rated opponent: his own (norwegian) rating goes up, while there is no similar deduction from anybody else's rating (as foreign players are typically not added to the norwegian rating system unless they play more than one event in norway). regardless, when they <are> added, only a fraction of their actual games affect their norwegian rating, and hence their norwegian rating quickly becomes very unreliable - exactly like the uscf rating of non-us players who play very few games in the us (and for which i guess the uscf doesn't bother to rate games played elsewhere, for very natural reasons). when i said fide non-rated players don't count, that obviously only applied to someone's fide-rating. |
|
Aug-10-11 | | frogbert: i see that the link says "fide event adjustment". i didn't bother to look it up, but that might imply that there's some kind of adjustment of the fide ratings prior to the calculations for the uscf rating, taking into account that fide ratings genereally are lower than uscf ratings. otherwise i would've expected a much bigger loss for naka in terms of uscf rating if the fide ratings would've been used unadjusted. anyway, this whole business of rating floors (as they are used by the uscf), "fide event adjustments" and other kinds of bogus adjustments all attack some of the premises of the entire relative rating system. the net effect is making it more of a ranking system and less of a measure of <relative> strength within a closed pool. |
|
Aug-17-11
 | | perfidious: <hellopolgar: the thing is, Naka is not a machine, so maybe what he achieved is statistically 57%. but we also have to take into account that players tend to lose their steam very very fast when they go on a winning streak, so every win becomes "harder" as the streak grows. and we also have to take into account that Naka hasn't played rapid chess in a while. last but not least, before he played those 14 rapid games, he had to finish his last game in the open and he was most likely tired. so i am willing to bet that, if we find 100 players of equal strength to naka, and let them play the exact opponents for 14 games, not 57, but less than 50 will manage to score 14-0, but i guess it's kind of pointless.> Unless, of course, your point is to win the argument at any cost, or you're a professional handicapper. If you were indeed the latter, you'd understand that you need a larger statistical size before proclaiming how right you are. |
|
 |
 |
< Earlier Kibitzing · PAGE 4 OF 4 ·
Later Kibitzing> |
|
|
|