@jg777
Agreed. FIDE ratings tend to be slightly lower than online blitz ratings for higher rated players. This is generally the recognized pattern. (and supported by the formula)
It is also true for lower rated players.
The entire issue here is that the formula predicts a HIGHER rating for lower rated players, contrary to what is expected.
Get it?
There is no data to support that players below 1500 online have a FIDE rating which is higher. All evidence points to a lower rating, the same as players rated above 2000.
This is the issue. The formula projects a higher FIDE rating for lower rated players. There is an obvious flaw in the formula. (as I've explained, the constant "187" is added to the equation for both higher and lower ratings, jacking up the result for lower rated players to show a FIDE rating above their online blitz rating.)
So this is the entire issue? Right here? Good. Mow we can pinpoint the issues at hand for both sides. According to the graphs, which represent the actual data completely devoid of doctoring, bad players tend to have their OTB ratings higher than their blitz ratings. Go check. That is the actual data. Where do most of the bad players fall? Regardless of what we expect, that is the data. @mdinnerspace
This is not as important as my last point (which is my main one) but I will say that my temperature example does show that the use of constants is acceptable in mathematics, and that STRICTLY in terms of accepted mathematical practice, having an x larger than its corresponding y in one part of a function and vice versa in another part of the function is not inherently unsound.
Whos data? LOL
You mean the data collected by the OP, which is based entirely on voluntary information provided in an online profile ?
Oh, that data. Yep. I get it now.
OK, so now we are off of the math part, and it's the data collection method that you are concerned about. While there are absolutely people who gave goofy values, the OP used methods that diminish outlier effect. But, @mdinnerspace , did you see that correlation in the majoroty of thr graph? That does not just happen by chance! Either the OP is flat out lying to us or that data is legitimate.
New players that begin playing chess online in a fast time control setting (relative to OTB 90 minute games) develop special skills. Often they are quite successful in achieving ratings >1500.
It is fact, that those same players when playing their 1st OTB tournaments, do not out perform their online rating and achieve a higher FIDE rating than their online rating.
But where is the data? What backs that claim up? The OP collected data, like we just discussed. It supports the opposite of what you claim.
@mdinnerspace.
@Jacob531
Whether or not the OP is lying to us or not about the data and how it was collected is not the issue.
The data is based on voluntary information provided in a members profile. Specifically the member states his/her FIDE rating.
How often can that be trusted.
Verifiable evidence. The scientific method. Not a formula based on voluntary information.
I can be trusted.
@mdinnerspace the data would not show such a strong correlation if it contained mostly random or insane values.
This topic has been archived and can no longer be replied to.