No ratings inflation; "GM" Paul Morphy was a master or expert, but no grandmaster

[raylopez99]

A friend sends me this information, from a computer scientist's presentation. Check it out.

What is shows is:

Magnus Carlsen is really good.

Female champion Hou Yifan can play really well at times, much higher than her normal rating. Is she cheating? She almost won the Tradewise Gibraltar last year for example.

Paul Morphy played at the 2124 to 2344 Elo level, just below the 2400 level associated with a modern IM. So he was a master, or expert, but hardly a modern GM.

Alekhine played really well against Capa in the 1927 WC.

There is probably no ratings inflation, or at least not a strong ratings inflation. So the fact there are over 40 players above 2700 Elo today vs two back in the days of Fischer means we are witnessing more strong players today. Stronger players like me, nearly an IM.

Check it out yourself at the link below.

RL


http://www.cse.buffalo.edu/~regan/Talks/SkillInferenceCornell.pdf

Magnus Carlsen: 2983 at London 2011 (Kramnik 2857, Aronian 2838, Nakamura only 2452). 2855 at Biel 2012.
Bobby Fischer: 2921 over all 3 Candidates' Matches in 1971. 2650 vs. Spassky in 1972 (Spassky 2643).
2724 vs. Spassky in 1992 (Spassky 2659). Hou Yifan: 2971 vs. Humpy Honeru (2683) in Nov. 2011.
Paul Morphy: 2344 in 59 most impt. games, 2124 vs. Anderssen. Capablanca: 2936 at New York 1927.
Alekhine: 2812 in 1927 WC match over Capa (2730).

[raylopez99]

On Thursday, January 17, 2013 4:30:18 PM UTC+2, raylopez99 wrote:

> Bobby Fischer: 2921 over all 3 Candidates' Matches in 1971. 2650 vs.
> Spassky in 1972 (Spassky 2643).
>
> 2724 vs. Spassky in 1992 (Spassky 2659). Hou Yifan: 2971 vs.
> Humpy Honeru (2683) in Nov. 2011.

What the above shows is that in the second Fischer-Spassky match of 1992, both players played better--statistically--that the first time they met in 1972. Of course the first time was marred by forfeits and blunders, but still, it's the exception to the rule that players get weaker with age (for example Sam's play over time has deteriorated with age).

RL

[The Master]

On Thursday, January 17, 2013 9:30:18 AM UTC-5, raylopez99 wrote:

> A friend sends me this information, from a computer scientist's presentation. Check it out.
>
> What is shows is:
>
> Magnus Carlsen is really good.

**Looks** really good. This is because we have not yet seen him play Houdini.

> Female champion Hou Yifan can play really well at times, much higher than her normal rating. Is she cheating? She almost won the Tradewise Gibraltar last year for example.
>
> Paul Morphy played at the 2124 to 2344 Elo level, just below the 2400 level associated with a modern IM. So he was a master, or expert, but hardly a modern GM.
>
> Alekhine played really well against Capa in the 1927 WC.
>
> There is probably no ratings inflation, or at least not a strong ratings inflation. So the fact there are over 40 players above 2700 Elo today vs two back in the days of Fischer means we are witnessing more strong players today. Stronger players like me, nearly an IM.
>
> Check it out yourself at the link below.
>

> http://www.cse.buffalo.edu/~regan/Talks/SkillInferenceCornell.pdf
>
> Magnus Carlsen: 2983 at London 2011 (Kramnik 2857, Aronian 2838, Nakamura only 2452). 2855 at Biel 2012.
>
> Bobby Fischer: 2921 over all 3 Candidates' Matches in 1971. 2650 vs. Spassky in 1972 (Spassky 2643).
>
> 2724 vs. Spassky in 1992 (Spassky 2659). Hou Yifan: 2971 vs. Humpy Honeru (2683) in Nov. 2011.
>
> Paul Morphy: 2344 in 59 most impt. games, 2124 vs. Anderssen. Capablanca: 2936 at New York 1927.
>
> Alekhine: 2812 in 1927 WC match over Capa (2730).

It appears as if you may have made a considerable number of typos above (or else the article itself may well be junk).

One thing we must never forget is the fact that these so-called computer 'scientists' as a group have a horrific track record!

These are the type of guys who used to get published everywhere, regarded as 'experts' on matters relating to computer chess, always predicting that computers could not do this or that, would never be able to beat a human expert, later changed to master, later still to grandmaster, etc. They were nearly always wrong, these supposed 'experts.'

One thing you got right though: Alekhine had to play very well in order to defeat Capablanca (just as Capablanca had to play very well in order to defeat Lasker). Of course, everyone already knew that.

[MikeMurray]

On Friday, January 18, 2013 1:11:52 AM UTC-8, The Master wrote:

> One thing we must never forget is the fact that these so-called computer 'scientists' as a group have a horrific track record!

Regan's a 2400 player.

> These are the type of guys who used to get published everywhere, regarded as 'experts' on matters relating to computer chess, always predicting that computers could not do this or that, would never be able to beat a human expert, later changed to master, later still to grandmaster, etc. They were nearly always wrong, these supposed 'experts.'

It wasn't mostly the computer science types who were claiming that, rather philosophers, etc.

[The Master]

On Friday, January 18, 2013 11:51:56 AM UTC-5, MikeMurray wrote:

> On Friday, January 18, 2013 1:11:52 AM UTC-8, The Master wrote:

> > One thing we must never forget is the fact that these so-called computer 'scientists' as a group have a horrific track record!
>
> Regan's a 2400 player.

I have not yet read the article linked to by Phillip 'Ray' Innes, but I expect you probably mean its author.

The authority term selected and used by Dr. Innes was 'scientist,' not nearly-an-IM (like himself). Thus, his credentials relate to his supposed use of *scientific methodology*, not chess expertise.
My point was that in spite of alleged embrace of this powerful methodology, those historically described as 'scientists' have a horrific track record indeed. My guess is that their numerous failures have resulted from simple failure to actually utilize the scientific method --along with careful, sound reasoning.

Now that the weekend has arrived I may actually attempt to read through the article itself, not just Phil-Ray's rgc discussion thereof. But Phillip made it sound like a complete waste of time... .

> > These are the type of guys who used to get published everywhere, regarded as 'experts' on matters relating to computer chess, always predicting that computers could not do this or that, would never be able to beat a human expert, later changed to master, later still to grandmaster, etc. They were nearly always wrong, these supposed 'experts.'
>
> It wasn't mostly the computer science types who were claiming that, rather philosophers, etc.

Can you cite an example of a 'philosopher' who, like all those computer scientist types I described earlier, was published in chess magazines or got his predictions all wrong (just as they did)?

I am trying to remember whether or not former rgc participant Bob Hyatt made any such predictions, but I can't recall. (Anyway, he was a bus driver, not a philosopher.)

[raylopez99]

On Friday, January 18, 2013 6:51:56 PM UTC+2, MikeMurray wrote:
> On Friday, January 18, 2013 1:11:52 AM UTC-8, The Master wrote:

[Junk by the Minor deleted]

>
> It wasn't mostly the computer science types who were claiming that,

Right. Minor is trolling as usual, but while I troll, I also provide useful information.

RL

[MikeMurray]

On Saturday, January 19, 2013 12:07:15 AM UTC-8, The Master wrote:

> Can you cite an example of a 'philosopher' who, like all those computer scientist types I described earlier, was published in chess magazines or got his predictions all wrong (just as they did)?

Hubert Dreyfus was probably the best known of the philosophical skeptics who have had to eat their words (or, more usually, act as if they never wrote them).

[The Master]

On Saturday, January 19, 2013 7:01:44 AM UTC-5, raylopez99 wrote:

> > It wasn't mostly the computer science types who were claiming that,
>
> Right.

Got any evidence that he's right? Or just lots of opinions, as usual.

> Minor is trolling as usual, but while I troll, I also provide useful information.

If I were going to start trolling, I would begin by creating a significant number of **original** threads (sorry for using all this technical jargon, Phillip).

Instead of, say, commenting here and there on threads introduced by geniuses like yourself and Sam Sloan, I would get creative and invent my own threads, comming up with new (and whacky) ideas and of course disguising them so they looked like serious opinions or questions. In other words, my posting style would begin to look a lot like your own, or that of Mr. Sloan!

Am I correct in remembering that your old pal Larry Parr used to insist that Dr. Blair was 'trolling' when he stubbornly and persistently corrected the record with regard to Larry's own considerable nonsense? In any case, Larry tried every trick in the book, but to no avail. Dr. Blair did not easily suffer fools, as Mr. Parr soon discovered. But back to the main idea.

I would be only too happy if either of you two fellows could provide any actual examples of what you seem to believe are 'philosophers' doing what the academics or 'scientists' obviously did: made claims which in retrospect (and at least to some of us, even back then) were obvious blunders. Let me give you an example of what I mean...

Right here in Indiana (a handy quote pops to mind: 'If there's a bright center to the universe, you're on the planet that it's farthest from.' --Luke Skywalker), there was a very strong chess master who got his stuff published in multiple places, and whose predictions, like so many others of this stripe, have proved wrong. No biggie-- everyone makes mistakes. But the whole reason they published his opinions ...and that's all they were, opinions... was on account of him supposedly being an authority, someone who knew his stuff and therefore could be trusted to make carefully reasoned conclusions, restrain from over-reaching, etc. Boy were they wrong (as usual).

You see, along with royal titles and political power and academic titles and money and good looks and rating points and all that sort of thing, comes an unwelcome 'bonus,' in the form of hubris. (Another handy quotation leaps to mind: 'A man's got to know his limitations.' --was it the movie Dirty Harry?)
Thank goodness I don't have to suffer the consequences of having any of those sorts of things. And you're not exactly leading man material yourself, Phillip.

When I read the term 'philosopher,' I think of Socrates. Then I realize he was a Sophist, tried and convicted of corrupting the youth, etc. So my mind wanders a bit, recognizing that most who have at some time or other been labeled 'philosophers' fell well short of the mark. But maybe Mr. Murray has different (read: lower) standards. Anyway, I have deliberately refrained from reading the next post in this thread until *after* I have responded to this one, though Mike Murray's name becons at the bottom of my screen.

Have a nice day, Phillip, and thanks for the laugh (rgc's permanently resident troll accusing someone else of trolling). Well, I _would have_ laughed if I had been in the right mood.

[The Master]

Allow me to clarify one thing: making a living as a professor, or teacher, of Philosophy (the subject) is not at all the same thing as actually being a philosopher, a 'lover of wisdom.'

College professors, armed with titles and degrees and considerable power at the local (University) level, tend to stray from the course of wise humility into that of hubris. [Rgc's Dr. Blair being a rare exception, kept humble by the automatic transfer of half his rating points to Dr. Phil IMnes, which of course had the predictable result on the latter's ego.]
These guys write and publish books, for crying out loud-- no self-respecting philosopher would ever do THAT.

Seriously, when you wrote 'philosopher,' I thought you were going to name someone like Socrates (a Sophist), Diogenes (a layabout), or perhaps Aristotle (a Natural Philosopher).

[The Master]

On Thursday, January 17, 2013 9:30:18 AM UTC-5, raylopez99 wrote:

> A friend sends me this information, from a computer scientist's presentation. Check it out.
>
>

> http://www.cse.buffalo.edu/~regan/Talks/SkillInferenceCornell.pdf


Well, I read it and it was just a 'brainstorm list' of stuff, including some ideas and ratings using arbitrary data samples. Where's the beef?

[Phil Innes]

On Thursday, January 17, 2013 9:30:18 AM UTC-5, raylopez99 wrote:

> A friend sends me this information, from a computer scientist's presentation. Check it out.

> http://www.cse.buffalo.edu/~regan/Talks/SkillInferenceCornell.pdf
>
>
>
> Magnus Carlsen: 2983 at London 2011 (Kramnik 2857, Aronian 2838, Nakamura only 2452). 2855 at Biel 2012.
>
> Bobby Fischer: 2921 over all 3 Candidates' Matches in 1971. 2650 vs. Spassky in 1972 (Spassky 2643).
>
> 2724 vs. Spassky in 1992 (Spassky 2659). Hou Yifan: 2971 vs. Humpy Honeru (2683) in Nov. 2011.
>
> Paul Morphy: 2344 in 59 most impt. games, 2124 vs. Anderssen. Capablanca: 2936 at New York 1927.
>
> Alekhine: 2812 in 1927 WC match over Capa (2730).

(In passing I note that Sophia Polgar's perfrmance in Greece is not noted here, she scored 2950+.)There has always been much contention with this kind of analysis, and Ken Regan will be aware of most of the arguments which are generally these: Such rating are often just of the 'performance type as with my Polgar reference, but apart from that factor there is a combination of achieved ratings and retrograde estimates which is to measure actual performance with speculations. Before any consistent application of ELO there were no ratings and thereby no measurements. When this subject was exercised here before there was a general consensus that if a chess engine could measure players games and assigned them a value that would provide at least a common index to what was measured and enable some uniformity of assessment. Anecdotally Morphy's games don't look like Kasparov's since they are relatively simple tactical forays on the romantic play theme — Morphy was merely the best player of the romantic period, and if you peg Morphy at something like 2300, how strong were his opponents?

Some evidence of this lies in historic players knowledge of openings which was slight, and that compared with software which has book=on [which is illegal and can't be included in ELO calculations]. Most historic players would find themselves at a severe disadvantage by moves 10-15 if playing a computer with book=on, and analysis software even quantifies this disadvantage. As we know when the advantage is above 1 it is definitely evident, and above 3 likely fatal in middle games. Positionally the computer is not o good at evaluating opening play since it does not quantify material to initiative very well. But at move 15 most historical players would be trailing even a reasonable amount of modern knowledge of openings, and also how to restrict opponents play limiting wild romantic attacks.

Whether historic players could adapt to modern conditions is very uncertain. If the best of them were really in some range 1800-2300 how much better do we think they would get with more knowledge? Or are they already playing at the limit of their ability to incorporate knowledge? After all, there are very many 1800-2300 playing adults these days who, despite modern knowledge do not significantly improve and therefore this factor alone might make a mockery of any retro-active predictive calculations or models by Ken Regan. Phil Innes

[raylopez99]

On Monday, January 21, 2013 2:18:15 PM UTC+2, Phil Innes wrote:
[some excellent points deleted]

I made some excellent points.

RL

[Roper]

These are conclusions from guys that never studied chess seriously in
their life and I am surprised that guys that did study chess seriously
believe them.
There is rating inflation.
I was 2200 35 years ago when I stopped chess.
And I see 2200 players as they are today.
They are much worst and the reason is simple.It's easier to get to 2200
today since the number of players that are above 2200 is ten times
higher.
Back then to get a 2200 you had to win IMs and GMs because these were
the only ones that had above 2200.Today there are (at least in my
country) more than 200 CMs,NMs and FMs that are above 2200 and half of
them are seriously overrated.
The same is the case to higher levels.Aronian has higher rating than
players like Botvinnik ,Petrosian , and Tal.What does that mean?That he
is better than them?Of course not.But it's easy for Aronian to be in
2800 when there are GMs like Arkadi Naiditsch.Check the game Van
Kampen-Naiditch Group B 8th round.A 2700 GM blunders a piece.He takes a
pawn that is protected.
Anyone that watches Tata Steel has already realised that it is a
blunderfest.
Simple combinations are coninuously missed by 2600 and 2700
grandmasters.
Sokolov was unable to win a totally won endgame against Nakamura.
In Group B , 9th round , GM Turov(2630) fails to see a simple mating
pattern and fails to defend against a simple attack.GM Ernst(2556)
blunders a piece because he missed a simple check and many many others.
In Group A the level is high but Group B reminds more players that fight
for an IM norm than players that are 2500+ grandmasters.
You have to study the games to realise how good players were Petrosian ,
Tal , Botvinnik , Fischer (and others) were.
Even players like Gligoric , Miles , Timman , Larsen ,at their prime
would wipe the floor with many contemporary 2700+ grandmasters.




--
Roper

[mdavis]

I'm not arguing here, just making some observations.

First, with today's vastly superior computer assistance for openings,
analysis and competition, I would expect players to be better overall
than in pre-computer days, just as today's athletes are better
conditioned, better trained and old records continue to fall (track and
field or swimming for example).

Second, when we play over games of past GMs we are usually only afforded
examples of the best games, the won games, the successful tournaments
played during their prime. The outright blunders are seldom seen, lost
to posterity.

Third, today's GMs have studied old games and found better lines and
flaws that are missed OTB or without the aid of computer analysis. Bad
lines are known and avoided.

I'm not saying there isn't a difference between today's Elo equivalents
and those of yesteryear. But I don't think there is so much difference
that older players in their prime would monopolize tournaments under
today's playing conditions.

[raylopez99]

On Wednesday, January 23, 2013 4:00:06 AM UTC+2, mdavis wrote:
> I'm not arguing here, just making some observations.
>
>
>
> First, with today's vastly superior computer assistance for openings,
>
> analysis and competition, I would expect players to be better overall
>
> than in pre-computer days, just as today's athletes are better
>
> conditioned, better trained and old records continue to fall (track and
>
> field or swimming for example).
>

Yes, true--a NY Times graphic showed that today's high school champion in track could beat the great Jessie Owens from the 1930s. Same for mind games: IQ steadily increases generation by generation. This is called the Flynn effect (http://curiosity.discovery.com/question/what-is-flynn-effect)

>
>
> Second, when we play over games of past GMs we are usually only afforded
>
> examples of the best games, the won games, the successful tournaments
>
> played during their prime. The outright blunders are seldom seen, lost
>
> to posterity.

Yes, true. Taylor Kingston who visits this board has written an article about the games lost by famous grandmasters.

>
>
>
> Third, today's GMs have studied old games and found better lines and
>
> flaws that are missed OTB or without the aid of computer analysis. Bad
>
> lines are known and avoided.
>

True. And to be a master you need to have a very good, if not photographic, memory.

>
>
> I'm not saying there isn't a difference between today's Elo equivalents
>
> and those of yesteryear. But I don't think there is so much difference
>
> that older players in their prime would monopolize tournaments under
>
> today's playing conditions.

I agree. I also think the prior poster Roper made good observations too.

RL

[raylopez99]

> Roper

I think you make excellent points. I also think the chess statistician Jeff Sonas, a world expert on statistics, agrees with you, since I notice he makes adjustments to Elo scores of yesteryear, so that they are "normalized" (made higher) to be comparable to today's 'inflated' Elo scored. For example, if you read though Sonas' explanation here: http://chessmetrics.com/cm/CM2/Formulas.asp?Params= you will see he is trying to compare groups to each other in the best way possible, and since Capablanca never played Fischer, it will be tough to compare past to future players, but Sonas does a credible job of it. His "lifetime best" Elo list looks credible, see here: http://chessmetrics.com/cm/CM2/PeakList.asp?Params=199510SSSSS5S000000000000111000000000000010100 (five year peak champions)

Another technique I've seen is to see how many blunders or deviations from the best play as scored by a strong computer (say Rybka or Houdini, playing on the best hardware so at a very high level). This way (and Chessbase.com has several articles on this) shows the list of all-time champions corresponds to Sonas' list. But that said, I bet a lot of today's 2700+ Elo supergrandmasters also score high.

Also, I once read that Artur Yusupov, when he was already or nearly already a grandmaster, did not know you could castle long when the square not traversed by the king (i.e. a1 or b1) was under attack by a bishop. So even back 20 or more years ago you had these holes in knowledge. And for example, as discussed on this board, the standard exchange sac of RxNc3 as in certain lines of the Sicilian was unknown until the early 20th century, so indeed there were gaps in knowledge back then.

But your points are well taken.

RL

[The Master]

You two seem to get along together very well.