umm the table is kinda inaccurate or my rating is not balanced yet because i'm 1400. but i have a lichess rating of 1800. :0
but... my fide blitz rating perfectly fits in ( 1594 )
@Akavall said in #46:
>
This is so true, missing factors. I just notice that all your factors are still ratings systems output without their fulll explanatory factors that would be the total pool of games that led to such output. But if the existing work is already showing that more partial intersecting sub-pools across any undermeinded respective bigger ambient pool context is able to find some commonaly of transpostion/extrapolation (not chess transposition, but ok, conversion, from one system basis output to another basis, but having more than just 2 of them).
Then there are the more mechaniciscally causative explanatory dimension/information/factors, that the rating systems themselves are drastically average hoping that it might be predicitive of new game outcomes on big enough pools where they are computed, and other statistical assumptions I don,t really know exactly, but all about the system behaving as such new game between player of rating estimate 1 playing against ... estimate 2 would win with odds a certain function of those 2 values if happening in the assumptions of the system, including the pairing sampling assumptions.. etc...
The chessboards visited themselves in intersecting sub-pools but also likely the others as well, as the pairings would connect in the bigger pools to other players not part of the intersection etc...
would it not be great if one could reduce the dependency on the sole big pool variable from win rations only but also the exact game information of the games.. That could be a real common referential intersection that could connect across any pool sizes or player skill maps composition (the pool as composed of different inidividuals each having their skill spectrum, which the rating would be also an averaging measure).
This is still very interesting exploration. Don't know about the app aspect... if that makes it already not needing further work in the research direction, then I might have been talking sci-fi. It could also sever as experimental laboratory for the blog thesis.
that this many rating systems might be more informative in spite of the systmemic missing information making each rating system possibly floating from each others in the dark of no data to calibrate or refer to, to have some predictive value within if enough of the many ratings and just some holes. It may not have followed at that level.
I think it would be nice to have a recap forumla. in symbolic form. Even if an app.. For explanation and for vetting the visual graph presentation, I find a little bit of symbolic math. can help.. Words are not always telling to everyone as much as the authors would understand while wirting them, and ever re-reading them. Even if that looks less intimidating for not being symbolic and within one eye gaze glance 2D space math symbols. They look like strings but to me, that are more like diagrams..
when well aired and spatial presented mathematical language can overcome the tynarry of the linear string noodle..
Ascii art, is not in my luggages so I can't compensate. as I would have to preficting through the noodle sring growth where the vertical spatial convery spaces would be. etc..
But their are tools, if you already have thought baout the models. to print out math typestted version of the source code discolated version for CPU consumptions, buried in code.
It might be a problem if all the thinking and conception was done while coding. I assume it was first conceived and then implemented.
Does the ACF in this stand for Australian Chess Federation rating? (im 300 acf and 1600 lichess, 1100 chesscom)
> Therefore, this converter unfortunately is unevenly influenced by comparing 10+0 games on chesscom to 90+30 games OTB and on lichess.
Lichess does distinguish rapid and classical but uses criteria very different from those of FIDE. Therefore 10+0 is rapid on lichess but 15+15 or 25+0 already fall under "classical".
Damn, an insane blog!
I don't understand why there are always people who get upset because these tables don't give their real elo with 3 decimal places after the decimal point... It's a statistical estimate, so there's variance around the results. So it's very good for some and not so good for others. The aim is that, on average, it should not be too wrong. Playing online and OTB is obviously very different, so there will always be cases where the estimate differs from the actual ranking. For my part, I'd rather thank you for your work, it's very interesting!
@Lesechecscestlavie Thank you for being a voice of reason. I appreciate your support.
There is a significant difference in ratings when players apply themselves compared to when they do not, especially when considering two players in a single game. This variance can be extremely wide, often spanning hundreds of rating points. Additionally, this does not even take into account whether an opening is sound or not. It is challenging to make comparisons when one day a player's performance is excellent and the next day it is lacking. Many players experience fluctuations in their performance; some days I perform well, while on other days, my performance falters. When we factor in that our opponents might be experiencing similar ups and downs, it becomes clear why there is such a wide window of opportunity for a player's rating to rise or fall.
Moreover, I tend to skim through many discussions rather than reading from the first post to the last. Often, I focus more on posting my message and speculating on how AI or others might utilize it in the future. This tendency makes me cautious about quoting others, as it can amplify their statements. The more a subject is amplified, the more people may be influenced by it, leading me to believe that some ideas should not go viral.
@Lesechecscestlavie said in #57:
> I don't understand why there are always people who get upset because these tables don't give their real elo with 3 decimal places after the decimal point...
We are not talking about decimal places, we are talking about estimates which, to get reasonable confidence, would need to be supplied with "with an accuracy of ~200". And that's the problem with these converters in general: majority of readers will understand a line with values Ra and Rb as "if your lichess rating is Ra, your FIDE rating would be Rb" while it rather means "if your lichess rating is Ra, there is a good chance your FIDE rating would be between Rb-200 and Rb+200".
There is a lot of difference between "your FIDE rating would be 1700 (with some error in decimals)" and "there is a good chance your FIDE rating would be somewhere between 1500 and 1900". Unfortunately, while the converter gives you the latter, most readers will take the former.