When did Elo inflate?

I looked at the old Elo rating lists, at Olimpbase. Someting that struck me was that the ratings were low - and fairly stably so.

While the tops fluctuated with individual achievements, like career and retirement of Fischer, let´s look at the rating of Nr. 100 ranked player:

1. 1969 - 2480
2. 1970 - 2480
3. January 1971 - 2480
4. July 1971 - 2480
5. July 1972 - 2480
6. 1973 - 2480
7. January 1975 - 2490
8. 1976 - 2490
9. 1977 - 2495
10. January 1978 - 2500
11. 1979 - 2500
12. 1980 - 2500
13. January 1981 - 2495
14. July 1981 - 2500
15. January 1982 - 2495
16. July 1982 - 2495
17. January 1983 - 2495
18. July 1983 - 2495
19. January 1984 - 2490
20. July 1984 - 2495
21. January 1985 - 2500
22. July 1985 - 2500
23. January 1986 - 2505
24. July 1986 - 2505
25. January 1987 - 2515

So when did real inflation come? 1969 to 1987 was 35 points in 18 years.

The number of high-rated players hasn't been stable, though. Has it? The effective value of the 100th spot would depend on the number of high-rated players.

In 1969, there were just 4 players with Elo of over 2652:

1. Fischer - 2720
2. Spassky - 2690
3. Korchnoi - 2680
4. Botvinnik - 2660

In January 2019, there are 91 players rated 2660 or more (Melkumyan and Leko share positions 90-91, at 2660 points). Number 4 is Ding Liren, at 2813 points.

Does Melkumyan, in 2019, play as good chess as Botvinnik did 50 years ago? Or have Elo points inflated?

Personally, I believe that rating inflation is a real phenomena... but it's nearly impossible to prove, or to separate it from other factors such as the effect of computer prep on the standard of play.

After all, the rating system was designed to compare playing strengths within a pool of interacting players... not to compare playing strengths across two different pools of players.

Continuing - just January

1. 1988 - 2515
2. 1989 - 2530
3. 1990 - 2540
4. 1991 - 2545
5. 1992 - 2555
6. 1993 - 2565
7. 1994 - 2565
8. 1995 - 2575
9. 1996 - 2580
10. 1997 - 2580
11. 1998 - 2585
12. 1999 - 2587
13. 2000 - 2595
14. 2001 - 2596
15. 2002 - 2596
16. 2003 - 2598
17. 2004 - 2606
18. 2005 - 2613
19. 2006 - 2618
20. 2007 - 2619
21. 2008 - 2627
22. 2009 - 2634
23. 2010 - 2643
24. 2011 - 2649
25. 2012 - 2652
26. 2013 - 2650
27. 2014 - 2652
28. 2015 - 2653
29. 2016 - 2653
30. 2017 - 2655
31. 2018 - 2653
32. 2019 - 2652

Hm. Since 2011 has been pretty stable

GM Melkumyan technically plays better chess today than Botvinnik did 50 years ago.  And yet he will not be considered as great as Botvinnik.   Not even close.

Have ELO points inflated?  Looks like it, but I have seen citations by a IM or GM Regan who is a professional mathematician who argues that there isn't rating inflation.

I don't actually think ratings have inflated much at all. I think people have just gotten better over time.

All those old school players like Fischer, Korchnoi, Botvinnik, etc., were great for their time, but they simply wouldn't be able to compete with modern day theory, computer analysis, etc.

People nowadays are just more knowledgeable than people 50 years ago, because the information is right in front of us, on the internet.

Because of course they wouldn't use the computer analysis, or the new theories, or the internet. 

The question is, would today's talent be as strong if they were back in the 60's - 70's?

Over 5 year periods:

1. 1969-1973 constant 2480 - 0
2. 1973-1979 - 2480-2500 - +20
3. 1979-1984 - 2500-2490 - -10
4. 1984-1989 - 2490-2530 - +40
5. 1989-1994 - 2530-2565 - +35
6. 1994-1999 - 2565-2587 - +22
7. 1999-2004 - 2587-2606 - +19
8. 2004-2009 - 2606-2634 - +28
9. 2009-2014 - 2634-2652 - +18
10. 2014-2019 - 2652-2652 - 0

The observed changes of general Elo level are basically stable 1969-1985, then increase 1985-2011, then stable since 2011.

Does this match timing of changes in chess theory and training/analysis opportunities?

Well, early - mid 80s is when the personal computer started really coming on. 

I would say it has kind of leveled off in the last decade or so. 

So that may be a big factor. 

the flaw with that thinking is that if most people have access to those things today(which they do) , then there is no reason that would cause rating inflation. ratings are not some empirical formula based upon what you know about chess. but are rather based upon is how you fare statistically against other players of today (i.e. your peer group. 

Ratings don't compare what Botvinnik knew against what Nakamura knows today about chess. 

Now see the rating of tops:

1. 1969 - inflation 0, Fischer nominal 2720, deflated 2720
2. 1970 - inflation 0, Fischer nominal 2720, deflated 2720
3. 1971 - inflation 0, Fischer nominal 2740, deflated 2740
4. 1972 - inflation 0, Fischer nominal 2785, deflated 2785
5. 1973 - inflation 0, Fischer nominal 2780, deflated 2780
6. 1975 - inflation +10, Fischer nominal 2780, deflated 2770
7. 1976 - inflation +10, Karpov nominal 2695, deflated 2685
8. 1977 - inflation +15, Karpov nominal 2690, deflated 2675
9. 1978 - inflation +20, Karpov nominal 2725, deflated 2705
10. 1979 - inflation +20, Karpov nominal 2705, deflated 2685
11. 1980 - inflation +20, Karpov nominal 2725, deflated 2705
12. 1981 - inflation +15, Karpov nominal 2690, deflated 2675
13. 1982 - inflation +15, Karpov nominal 2720, deflated 2705
14. 1983 - inflation +15, Karpov nominal 2710, deflated 2695
15. 1984 - inflation +10, Kasparov nominal 2710, deflated 2700

Note that Karpov never matched even the nominal Elo of Fischer.

Sure, but I know much less than grandmasters so I fare much worse statistically against them. 

There are a few factors that tell whether you will be better than someone else but knowledge is one of those factors. 

If I am much more athletic than Tom Brady but do not have the knowledge of football I can never play football as well or better than him or his peers. 

My point is that ratings do not care or compare knowledge over time. Yes, certainly having more chess knowledge generally means you will play better and at a higher rating. That is compared to your peers of today. The pool that you are paired against. 

My point is that ratings DON'T have anything to do with how your knowledge may compare to that of people who are long dead and played in a different rating pools decades ago. 

1. 1985 - inflation +20, Kasparov nominal 2715, deflated 2695
2. 1986 - inflation +25, Kasparov nominal 2720, deflated 2695
3. 1987 - inflation +35, Kasparov nominal 2735, deflated 2700
4. 1988 - inflation +35, Kasparov nominal 2750, deflated 2715
5. 1989 - inflation +50, Kasparov nominal 2775, deflated 2725
6. 1990 - inflation +60, Kasparov nominal 2800, deflated 2740
7. 1991 - inflation +65, Kasparov nominal 2800, deflated 2735
8. 1992 - inflation +75, Kasparov nominal 2780, deflated 2705
9. 1993 - inflation +85, Kasparov nominal 2805, deflated 2720
10. 1994 - inflation +85, Karpov nominal 2740, deflated 2655
11. 1995 - inflation +95, Kasparov nominal 2805, deflated 2710
12. 1996 - inflation +100. Kasparov nominal 2775, deflated 2675
13. 1997 - inflation +100, Kasparov nominal 2795, nominal 2695
14. 1998 - inflation +105, Kasparov nominal 2825, deflated 2720
15. 1999 - inflation +107, Kasparov nominal 2812, deflated 2705
16. 2000 - inflation +115, Kasparov nominal 2851, deflated 2736
17. 2001 - inflation +116, Kasparov nominal 2849, deflated 2733

So one could make the argument that Morphy is/was a greater player than Magnus, and the Magnus fan can't use ELO rating as a rebuttal, yes?

Not sure what you are saying / asking? do I think if Morphy and Magnus went head-to-head? I think that Magnus would crush Morphy if it was Morphy from the 1800's versus Magnus of the 2000's.

If you could transport Magnus back in time and take away his tools and the knowledge that had been gained since Morphy's time, or conversly transport Morphy forward in time and give him all the tools and knowledge that players of today possess - not so sure. 

I do know that in his time (again compared to HIS peers - not chess players of all time) Morphy was head and shoulders above his peer group. so in that sense, I would say that Morphy was a better player than Magnus. 

You could only make a valid comparison of ratings if Morph and Mag were in the same ERA and played against each other. 

1. 2002 - inflation +116; Kasparov nominal 2838, deflated 2722
2. 2003 - inflation +118; Kasparov nominal 2847, deflated 2729
3. 2004 - inflation +126; Kasparov nominal 2831, deflated 2705
4. 2005 - inflation +133; Kasparov nominal 2804, deflated 2671
5. 2006 - inflation +138; Kasparov nominal 2812, deflated 2674
6. 2007 - inflation +139; Topalov nominal 2783, deflated 2644
7. 2008 - inflation +147; Kramnik nominal 2799, deflated 2652
8. 2009 - inflation +154; Topalov nominal 2796, deflated 2642
9. 2010 - inflation +163; Carlsen nominal 2810, deflated 2647
10. 2011 - inflation +169; Carlsen nominal 2814, deflated 2645
11. 2012 - inflation +172; Carlsen nominal 2835, deflated 2663
12. 2013 - inflation +170; Carlsen nominal 2861, deflated 2691
13. 2014 - inflation +172; Carlsen nominal 2872, deflated 2700
14. 2015 - inflation +173; Carlsen nominal 2862, deflated 2689
15. 2016 - inflation +173; Carlsen nominal 2844, deflated 2671
16. 2017 - inflation +175; Carlsen nominal 2840, deflated 2665
17. 2018 - inflation +173; Carlsen nominal 2834, deflated 2661
18. 2019 - inflation +172; Carlsen nominal 2835, deflated 2663

Note that while Kasparov and Carlsen, unlike Karpov, did surpass the nominal Elo of Fischer (and so did Topalov and Kramnik), none of them has been as dominant over Top 100.

Nice info for reference, kudos for compiling it.

Note, incidentally, the reason there was no World Chess Champion before 1886.

After 1862, Morphy was sulking and refusing to play chess at all. After 1873, Steinitz did play - but at blindfold and simultaneous games, refusing to meet masters.

Meaning that they were not themselves Champion, yet disqualified anyone else from being one.

It was only after Morphy died in 1884 and Steinitz agreed to play Zukertort that there was a Champion.

If Morphy, by sulking and refusing to play, could disqualify Zukertort as World Champion ever and Steinitz till 1884, how about Fischer?

Was Karpov ever a real Champion?