Lichess rating vs normal rating

@giulatestacoglione

It’s nonsensical to me to ignore than online creates more variables that impact player performances and ratings than OTB. I would never use an example of beating a 1900+ OTB player by winning a classical online game against them while they were taking care of their 1 year old as demonstration I can compete against such strong players, nor beating a 2000+ OTB player while they were at their sisters birthday party similarly. But competing against three 2000+ OTB rated players in an OTB event and scoring 1/3 gave me a reason to think I could rise to their level if I worked on chess more. Online ratings are fickle, who knows if a rating is accurate or not, or if you’re playing someone who’s there to do their best or just pass some time on a bus ride home. OTB, when people pay money, carve out a weekend of their time, they are there to play chess and do their best.

Show me a model of correlation between online and OTB and I’ll demonstrate how inaccurate it is. It simply doesn’t work. You get a ball park guess that anyone could make by asking some OTB players to assess you could without any formula needed just by their playing experience in OTB play.

-Jordan

@jg777 I dont think the avarage player on this site is a father of a one year old child....and in any case as you reason there should be no difference playing a 1500 or a 2000 on this site because ratings are non representative....i dont have watched your profile but i dont think you have won 2500 classical player....then i could be wrong...and im the only one that when play chess on this site has no baby to look after.

@giulatestacoglione

Right, the point is there’s ample reason to believe online environments create factors that negatively or falsely represent ratings. OTB there are less. A great indication of that is how FIDE has online and OTB ratings- they are considered different if they felt there weren’t significant differences why wouldn’t they rate their online matches to their OTB ratings?

-Jordan

Lets think about it in a statistical fashion.

The total population of players online vs Fide Tournaments allows us to draw larger conclusions on the standard deviation in terms of "Reliability and Consistency" (of ELO). These conclusions, due to the shear size of Opportunities (Games played) are statically significant inputs onto the output value of online ELO measurements/scores. (Even with these data sets conforming to a non normal distribution, the ELO formula's & match making would compensate for this lopsidedness of the player population.

The "Ohh but there are cheaters online" Ok if we took the sample size and said it equaled the population fraction that we end up playing over the course of set time and said this has a z value of 90% (10% cheaters) and divided that by our error of estimation, (n=p(1-p)(z/b) you would see that with a consistent value of (n, games played) that 10% of the player base can be cheaters and this would not cause ELO values to leave a range of acceptable quality levels.

So my question would be, why does Fide not figure out a way to increase its (N, games played) to have a more statically relevant ELO score for its player base.

edit
@jg777 you are Right, there would not be a correlation, because the Fide OTB scores are less statically significant and i am sure your p values would be much less then .4