Myth of Chess Elo rating inflation

Myth of Elo rating inflation in over the board Chess

This article was posted here in 2021 but the text written about 10 years ago only the Correspondence Chess section is new since I did not play Correspondence Chess when I wrote down my thougts.

Many older players keeps saying that there is inflation in Elo rating. They simply can not accept the fact that the general playing strength has increased much and that training methods has improved. This myth is widly spread and at Wikipedia is even written as that this claim is true, while this is pure bullshit and the older players are simply jealous and angry that more players has thier Elo rating or higher so that they are not high rated anymore.

A fact that they ignore are for example that Kenneth Regan, a statisticer and IM, has investigated rating inflation in a academic paper and he compared every Elo from 1600 and up and he concluded that players for example players with Elo 2000 when Elo rating was created played same percent of computer moves as players with Elo rating 2000 today and its the same thing for Elo 2100 , Elo 2200, Elo 2300, Elo 2400, Elo 2500, Elo 2600, Elo 2700 and Elo 2800 as well as for every lower rating. Thus inflation is a myth.

Its easier to get higher Elo today becouse of many factors. Internet has made it more easy to play training games, one can play 24 hours a day. Its also possible to watch elite level play live in tournaments with commentators and engine evaluations displayed. Computers has also effected human games, today one gets it in black and white exactly where one went wrong and where one played right and learn better moves and other ideas among the whole game. Many more of todays tournaments are Elo rated and there are more players to met with about the same strength. Games vs the about same rating are more stimulering and one learns more.

The quelety of books has increased. There are more and better books about every aspects of the game and better biografis written for average club players with better annotations. Todays book are also more easy to read and has better layouts. Modern players has studied and learned from not only the old masters but also from today elit players as well. Its also possible to watch Chess DVDs and other things to learn more about Chess.

Ambitius players can more easy get 1800 Elo today than for 20 years ago. In the old days it requered much more self study. It has also effected class system in Swedish championship. 1800 used to be the floor for class I. Today 1800 is the floor for class II.

Its often said that todays junior players knows more concrete opening theory than Bobby Fischer at his peak and I would not be surprised is it is true.

For my own part I used to have 1600 in Swedish LASK rating. Then I started to study modern books about the old masters and my rating jumped up almost 300 Swedish LASK rating and my Elo in long games has since that varied between 1820 and 1907. When this text was written I have 1840 in Elo. I have however much higher rating in both blitz Elo and in rapid Elo. Swedish LASK rating does by the way not exist anymore and Sweden only uses Elo rating.

I think of the old masters as great pioners but one can realy not compare todays elite players with older elite players since todays elite knows the game better and players better Chess and also mets better opponents thus learns more from theire own games. If any of todays top 200 players traveled back in time 30 years in a time machine they would become World Champion.

The World Champion that often gets lowest numbers amount of mistakes are Capablanca. The reason for that are that he played very boring postions and very simple Chess positions where there often are only one or two moves to consider. He was 30 years ahead of his time in endgames and often won objective lost endgames. Hes opponents in that times elite was signifikanly weaker than todays or even only 30 years later hes times elite so they gave him less concrete problems to solve.

Deflation Elo Correspondence Chess

In Correspondence Chess computers has realy effected the game. In many of todays elite tournaments one single win and draw all other can lead to tournament victory. The same is also true in some Final Events in thematical games.

In Correspondence Chess Chess engines and tablebases has made that 500 elo rating difference means very litle if both players are rated over at least 2200 Elo. It only means that the player played Correspondence Chess before computer era or that the player has higher Fide Elo thus got higher start Elo since the player could play in events with higher average Elo. I have played draws vs 500 Elo higher rated multipe times, including in rated games. But I have also played draws vs lower rated as well.

In Correspondence Chess the Elo one has is incredible depending on start Elo since the Elo after 30 games are based is based on performence vs opponents average rating. Thus if one has higher Fide Elo one can play in higher average Elo events than 1800 Fide Elo players can join. And high start Elo means that one can continue to play in closed higher average Elo events so rating stays high.

Also K factor drops fast thus a wins suddenly win gives very litle and wins are also rare becouse of computers. Even draws vs 500 Elo higher gives very litle. I managed to play up my start Elo more than 200 Elo and managed to get a CCE title but now my K factor is so low and most my games are draws so I can hardly get that much higher Elo to be able get any higher title since higher norms requeres an unrealistic score unless opponents in the event has high average Elo. I play in lots of open events in order to maybee met som high rated players but one also mets lots of low rated as well thus no rating gain.

Many high rated titles players like SIM:s that I have met has delayed clear draws for years by refusing draw offers even when I have the more fun side of the draw and take leave mode in protest and does no seem to understand or else they pretend that computers has not changed the game. All 2350+ Elo or players has titled ranked above CCM that I have met has played has all play Correspondence Chess since long before computer era. They think that Elo 2200 is much inferior player and thus are living in theire alternatrive reality where they are not overrated and no player is underrated, all rated bellow them are much weaker players. They has never heard the terms start Elo, the relation between start Elo and FIDE Elo or heard how good modern Chess engines are or heard the word tablebases.

Rating only matters if one of the players are not using computer engines or uses a weaker engine (often an older engine) or does not have access to tablebases.

Becouse of computers and Elo deflation the CCE title has almost become the title IM and the title CCM has almost become the title GM. This is a bit funny since female players can chooose between name of title between CCE and LIM and CCM and LGM. Also I have also met IM:s etc with lower rating than 2200 Elo.

In the modern Computer age is very difficult to win regardless of rating difference. High rating does not mean strength it meen either high FIDE elo = higher start ELO or that the player played Correspondence Chess before Computers reached todays standard. Titles also meens little these days, I have even met IM:s with lower rating than me. At higher level a single win and no loses can lead to tournament victorys, this is also true in some Finals in Thematical events.

To win a game something from the list bellow must happend:

  • Either player messes up mover order in opening.
  • Either player blindly follows an opening book that are not good.
  • Either player plays an opening that is either dubious or better for opponent.
  • Either player makes some other kind of opening mistake.
  • Either players does not use computer check or use to low depth or uses and outdated engine or has not got tablebases.
  • Either players analys one move but manage to commit an other move or analysed the wrong position or was confused by similar ongoing games positions.
  • Either players loses on time.