at the mercy of the arbiter

[Keith Arkell]

I think Leonard Barden is completely missing my point.

I am of course well aware of ''Chessmetrics",and indeed Leonard's own
work-especially a few decades back- on providing us with current
playing strengths of young players. I am also aware that there have
been many attempts to ELO rate the players of old.

What I am suggsting ,whether I am right or whether I am wrong,is
certainly not done out of ignorance.

I am putting forward a view which,were it to be the case,would
override any mathematical attempts to compare players of different
generations. It is the concept that the standard of play is creeping
up steadily all the time,at all levels,and always has done.

I gave a demonstration by means of a thought experiment to clarify
my point.In this imaginary model universe of 10 players,it is
impossible to show numerically the overall rise in standards of play.
I am suggesting that a watered down version of this scenario is at
work in the real world. Clearly,as I have demonstrated,this rise in
standards would be undetectable by looking at ratings. It would be
hard but not impossible to detect by comparing the quality of games
over a short period,but it would become more readily obvious when
comparing the quality of games played by similar graded players(or
even the same player!) over a bigger gulf of time.

[Leonard Barden]

No, the point was that you downgraded 1930s players to 2050, the level of a weekend tourist.

[Mick Norris]

Keith's basic theory makes sense

A level Maths in 1982 contained what had been degree level Maths from years before - at A level, I was taught by a recently qualified Maths teacher, and he wasn't fantastically better than me for that reason

In 1979/80, I did O level Maths past papers all the way back to when I was born (1964) I think, and was surprised how easy they seemed

This wasn't because I was brilliant, as 3 years at the (then) second best (maths) university in the country proved

I don't see why chess shouldn't have substantially higher standards, doesn't make the 1930s players bad, just means that players like Keith have been able to build on their efforts

[Mark Howitt]

Yep, Keith seems to be basically correct. A computer irons out errors very quickly and easily from even average players... if you look at GM games from the 90s, there are quite a few fairly basic errors in them. Nowadays, any small tactical mistake can easily be punished by a 2000 player who has prepared with his computer against an older 2200 player who hasn't prepared with a recent computer or properly. It doesn't even take THAT much effort by UK players to reach a level which would definitely be in the top 10% by 1930s standards.

[John Moore]

I just wonder how strong Capablanca, Alekhine and Lasker really were! At what would you estimate their ELO (or ECF) since Keith has now agreed they might have been a bit better than 2200 but obviously doesn't agree with the historical ratings.

Oh and I wonder what Mr Howitt does after his preparation with his computer has run out and he has to play his own moves against the old git who doesn't have one (or can't be bothered to prepare) - but is 200 Elo stronger.

[Leonard Barden]

Chessmetrics says the April 1930 No 100 player was c 2450, Keith says he was really 2050.

The top two in April 1930 were Alekhine 2826, Capa 2799, so you have to reduce these numbers by 400 points. Which makes Alekhine a fraction better than Andrew Ledger, Capa slightly worse than Jovanka Houska......

[Stewart Reuben]

I have been presented with data which shows that there has been very little inflation in the ratings of players this century. (There was some inflation in the 1980- early 90s.) We would expect that the ratings will go up, due to greater understanding and opportunities, but that does not mean the higher levels means inflation.

Keith asked about how many players there were in Capa's time. This is an incorrect question. The correct one would have been, 'How many players were there who played internationally? That was a very small number indeed. The explosion since the introduction of FIDE Ratings in 1970 and then 1977 Lloyds Bank Masters has been astonishing. At Leonard's suggestion that event introduced to the world the possibility of players gaining title norms in open Swisses. Of course the reduction in cost of travel has been another major contributor.
To give you a concrete example. I achieved my first FIDE Rating of 2270 in 1976. I was then rated in the top 1500 in the world. I think there were 2500 rated players. But practically none of them was from the USSR. The diaspora of the Russian-speaking players after the USSR break up led to a huge expansion in the number of strong players internationally. There are now about 100,000. But most of these are under 2200.

Inflation is certainly possible. There are a large number of different factors, all of which have an effect, some of which are deflationary. Take for instance the k factor being 10 (becoming 20 1 July) for players 2400 or higher and k=15 (becoming 30) for players under 2400. One player gains more points than the other. This does not seem to cause inflation and I don't know why.

There definitely was inflation in the USCF System and I think this was deliberate marketing. If a player's rating goes up he is more likely to continue playing. I had a USCF rating of about 2100 in 1965. I did not again play until 1991. My rating rose with every event in which I played until it again stablised at 2240. Yet I was a weaker player in 1991 than in 1965.

The deflation in the ECF system has been caused by the systemic undergrading of juniors. This could have been at least partially solved by having perhaps 4 grading lists a year instead of the absurd one.

Of course any enclosed system results in incorrect ratings. Thus the problem with Myanmar and with the fact that Kasparov could never have a 'true' rating. You need to meet a mix of opponents around your own strength to get a true fix.
Stewart Reuben

[Mark Howitt]

Lol someone's a bit sensitive.

Note that I didn't mention Capa or similar players- but the GENERAL level of play outside the elite was certainly lower than today.

To be honest, that kind of response is one of the reasons I play little OTB chess these days- despite having written a 'chess novel' and beaten players who've represented England for their age. The 'old git' attitude just isn't appealing to younger players- there are so many other things we can do to get fulfilment!

Anyway... I wouldn't claim to be too strong a player- but at the club level if one player has prepared properly (not just with the computer but mentally as well) he can easily beat people who don't know how to do that.

[Keith Arkell]

Given how much progress humanity has made in the last century or so(eg with safe water supply,telephones,cars,electric lighting,home refrigeration,nuclear energy,medicine. etc etc etc...),why does it cause offense in some quarters when I suggest that chess isn't somehow immune from this progress? And furthermore Leonard,you have played a part in this progress.I would have thought,given that chess is mathematical,information based,and in principle solvable,it would be a prime candidate for being an area of human advancement.

Would it insult Alexander Graham Bell to sugest that today an average BT technician knows more about telecomunications than he did?
I think it is very easy to get seduced by big names,and it would interest me very much to know how Fritz or Rybka would rate the opening,middlegame and endgame play of ''Andrew Ledger' and ''Ivanka Houska'' compared to that of Alekhine and Capablanca.

They were the greatest players on the planet in their day.Maybe in 80 years time there will be players playing chess at 3200 by todays standard?

[Keith Arkell]

Stewart,In the context of what I was trying to establish,my asking how many chess players there were on the planet in Capa's time was not an incorrect question,though it would have been an incorrect question in the context of the issues you addressed in your reply.

I merely wanted to establish from what sized pool Capa emerged.If the starting pool today is 10 times as big,then probability would dictate that,on raw talent alone,we would emerge with 10 Capa's;and one of the new 9 would most likely be the most talented.

[Stewart Reuben]

It does not work liked that in terms of the absolutely leading people. I eschew the term 'genius' because we would end up in a discussion about that term and I set very high standards.
Shakespeare is reckoned by many to be the greatest playwright and yet very few people were able to read and write in those days and still fewer in English. By analogy there should by now be many contenders for his crown. Perhaps there has been deflation in ability to write in English? Was his ability about the equivalent of Elo 3200 in writing?
Stewart Reuben

[Leonard Barden]

Keith, you are the one claiming without specific evidence that Chessmetrics overrates Alekhine and Capablanca by 400 Fide points, so you have the responsibility to do the Rybka/Fritz comparison to demonstrate or otherwise that Andrew Ledger is the equal of Alekhine....

[Steve Collyer]

It seems Capa fares rather better in comparison with modern GM's than some give him credit for, even when you use modern methods to analyse games:

http://chessbase.com/newsdetail.asp?newsid=3455

Check the average error & % blunders graphs 1/2 way down the page. Also the % best moves further down.

Reduction in chessmetrics elo by 400 points???
Not on the basis of the evidence provided here.

[Keith Arkell]

Good point stewart. Maybe you are highlighting the difference between science and art.From the artistic point of view there have been few players games which have given me greater pleasure to play through than Capa's best.When it comes to enjoying art it isn't so important about such concepts as the level of resistance(within reason).It's a pleasure to enjoy the patterns and manouevres of the game.
Shakespeare and eg Mozart were artists.No new information has emerged in the ensuing centuries which,had they been aware of it,they could have used to improve their work;therefore they are not disadvantaged chronologically.

[Mark Howitt]

I'm sure Stewart himself knows that there is no 'objective' standard of grading writing... in chess by contrast, moves and positions can often be assigned mathematical values which are accurate within at least '0.1 pawn'. Sadly, I believe some older people assign romantic notions to the strengths of players they like, or who were considered strong in their youth.