Wednesday, March 12, 2014

Overrated

We've been complaining about too-low grades on S&B recently.

Here, by contrast, is a player with a grade that's far too high.


When I were a lad, players of 200* and above were supposed to be good. It was as late as 1984 before I ever beat one. Alan Trangmar was graded 202 and I was nearly four weeks past my nineteenth birthday.


I don't think Alan played very well - well of course he didn't, he lost to me - but even so, it felt like a real accomplishment and I'd waited a long time to achieve it.

Now I am invited to believe that it's the same accomplishment if somebody beats me. I doubt it. I doubt it's really something worth waiting years for.

Now I know that:
  • a lot of ECF ratings got boosted a few years ago for perfectly good reasons ; and
  • my rating is not actually based on very many games, seeing as I live abroad now. (Matter of fact it'd be higher still if I hadn't lost the last game I played.)
Nevertheless it's still bizarre to see 200+ next to my name. I don't believe I'm playing my best chess at the age of forty-eight. Still less, my best chess by a distance.

Point is, though, 200+ used to be - or seemed to be - some kind of gold standard, something that told you "the bearer of this grade is a really strong player". Which I am not and never will be. And just as my problem with the FIDE Master title is that I've beaten seven or eight of them, and a "Master" ought to be somebody I can't beat seven or eight times -so it seems to me that the qualification for reaching 200+ ought to be that the likes of me don't do it.

[* Of course all this will be incomprehensible to readers who don't know the English rating system. Just take it that 200+ used to be really good.]

15 comments:

Anonymous said...

All of the ECF grades got boosted for arguably spurious reasons. Average players and the average grade used to be in the 105 to 120 range. This was boosted to 130 to 145 with greater increases lower down and lesser ones higher up. As one might expect, this is having an effect of feeding upwards. A shortage of ambitious eighteen year olds taking points off established players contributes. Those of IM standard have observed that whilst their International ratings remain in the same range, their English grades are increasing. They've also commented that some of the 200 graded players they face regularly are not of that standard.


RdC

Ashish said...

Why don't the English teach their children how to ... rate games by Elo? Is this just another example of English exceptionalism?

Jonathan B said...

Ashish, English club games are played on weekday evenings and so typically can't be rated because the time control is not long enough

Matt Fletcher said...

From a sample of 1 person over 10 games, it's difficult to draw any conclusion - if we assume Justin's 'true' grade is 189 (which based on his Elo and past performance looks to be the right ballpark) then he'd have to score 6.5/10 against an average 189 to get his grade of 204. I reckon (based on some fag-packet calculations) that there's about a 1% of this happening by chance - so unlikely but not totally impossible (if 10 bloggers played 10 games a year for 10 years, you'd expect one to end up over-rated by 15 points or more at some point).

If it was based on a sample of 30 games, it would be much less plausible to have happened by chance, and much more likely to be indicative of a problem with the system.

Jonathan B said...

(if 10 bloggers played 10 games a year for 10 years, you'd expect one to end up over-rated by 15 points or more at some point).

Yeah, but it bloody well wouldn’t be me - a bitter blogger writes.

Matt Fletcher said...

Yeah, but it bloody well wouldn’t be me :) Yeah - but you play far too many graded games to get a "wrong" grade by chance - if you want to get to 200+ you'll just have to improve!

I reckon my first grade on returning to chess in 2010-11 (180) was at least 10 points high - it was based on only 16 games and my previous highest grade had been 165. My chess has improved a lot since then whilst my grade has remained static (but now I'd like to think it's a little bit too low...)

Jack Rudd said...

That sample also includes a game against me, which in itself is always going to boost the range of potential scores.

Jack Rudd said...

Here's another thought, which I'll put out there because I've had similar thoughts about my own grade: the 200 mark seems much less special to you now than it did in 1984 because you're now a better player than you were then. The closer you are to whatever figure you were aiming at, the less it seems like a truly special figure that only rare people can achieve.

ejh said...

because you're now a better player than you were then

This is plausible, I think, but I don't think it's plausible enough. I'm probably better than I was thirty years ago, but not that much.

Anonymous said...

If you read the comments to Thursday's blog entry, it's possible this is a case of mistaken identity and Justin's six games at Hampstead won't give him a grade at all unless he comes back to England and plays a few more.

RdC

Anonymous said...

This was the ECF list that finally jumped the shark.

I've lost one of my last twelve longplay games v IMs and GMs yet I'm in danger of going below 200 without getting an iota worse.

Meanwhile someone I know has a very nice grade, in fact seven points higher than the strongest player he's beaten in his life.

PG

Anonymous said...

I think the time control is more an issue with FIDE not rating evening games, rather than the inability to Elo rate them

Jonathan B said...

Are you sure?

Adam Raoof’s tournaments at Hampstead are Game in 60 plus 30 second increment from move 1. They’re played at the weekend and limited to u2200s because - I thought - FIDE wouldn’t rate games that short for stronger players.

Anonymous said...

Sorry - maybe I'm not making myself clear. I was attempting to comment on the reply to the second comment by Ashish.

FIDE will rate Blitz and Rapidplay games using an Elo system. They may not rate certain games (e.g. games played of an English evening) for their long-play (or whatever it is called) list if the time control is not satisfactory.

That is a separate question from whether English games, played on a weekday evening under a fast time control, can be Elo rated, as it would appear to be entirely possible so to do.

Why the ECF uses its own system is something I can't answer.

AWIC

Anonymous said...

In response to PG, under the ECF grading system, you are only ever as good as the results of your most recent 30 games or last 6 months of play. This contrasts with Elo type systems that can have a longer built in memory.

In response to AWIC, the ECF grading system is one of the oldest in the world, dating back to the 1950s. If we retain it, it is because the increasingly elderly English chess public, many of whom have been playing since the 1960s or 1970s are familiar with it and its foibles. The International system is plainly struggling with all the improving players that it let in by reducing its floor to 1000, so it isn't just a question of flicking a switch to use Elo methods for national ratings without working out how to handle Juniors and others showing rapid improvement.

RdC