Chess Grading Destruction

[Alan Walton]

There are very few players who are truly more concerned about the team result than about their own result - and I see nothing strange about that!

I'm glad I haven't been a captain or player of any time you are involved in, in that case. In my experience, players are very aware of what the team needs in any given context and will strive towards that regardless of the consequences for their ECF grade or FIDE rating.

This kind of player would be quickly kicked into touch if that was his philosophy and was playing in any team I run/captain

[Maurice Lawson]

Your reaction is a very British old school one, and what I expected of course! But it doesn't change the fact. You simply don't like to acknowledge the reality. Chess players try to win their games for their own benefit - it just so happens that this naturally almost always coincides with the team's needs. But it is not the thing which really drives most people! You might be amazed to hear that I am very much regarded as a team player in all the teams I play for. However, I also like to acknowledge the truth of what motivates me!

[LawrenceCooper]

I would echo Alan's and previous comments. The teams I play for are full of players wanting to help the team be successful. I'm normally much higher rated than my opponents in local leagues (apart from when Keith Arkell is in the opposing team!) but I'm happy to halve out if it helps the team or to avoid risks if I feel that board 1 may be the opposing team's best chance in the match. Most of my teammates have similar thoughts. Sadly the last match I played involved the captain losing R+P v R+P after he had been trying to win to save the match.

[Maurice Lawson]

Stewart has made some excellent points! The reminder about the fact that k=20 is around 38 games again shows why going to an annual simple average (where the number of games for an active player might be 60-100) is such an excessive move. Of course the vast majority of players are not 2400+ or even 2200+, so the 30-40 range is far more appropriate for general grading.
It is very clear that all these comments about what the grading system is to be used for, can be coped with in almost any reasonable system. We see tournaments being run with FIDE rating limits as well as with ECF grading limits, and this moves different players into different sections at different times. In practice it makes no noticeable difference to running the tournaments.
Some old things are there because they work, but others are there because people refuse to believe there is anything wrong with them!

[Christopher Kreuzer]

Maurice, is it your grade that motivates you (trying to achieve new heights, maybe a grade over 190?), your results (remaining unbeaten during a season, or remaining unbeaten when playing those of a certain grade or lower), or the quality of your chess (playing well during all phases of the game, learning from your mistakes and not making silly errors or repeating them)?

Would you let a team result come ahead of these considerations? (e.g. drawing to protect a grade when the team needs a win, agreeing a draw instead of pressing for a win if the team needs it, or trying to play perfect chess and losing when more practical play might secure the point or half-point for the team?) Are you prepared to play for a team when you know you might not be in good form?

[Maurice Lawson]

Chris, that's a very good question to reflect on. I believe the answer is, in a sense that they are all relevant. But in any one game I set out trying to win and that is my primary motivation for that game. As the games develops I will either continue looking for ways to improve the position so as to develop a win, or alternatively if the game starts to go against me I will be looking for ways to bring it back to a drawing chance. Nothing particularly surprising in any of that, but the point I am making is that all the other motivations become peripheral during any one game - all that matters is maximising the immediate result. I suspect the reality is the same for most people in most situations.
And yes, I have responded occasionally to the team situations you mention, but in practice they very rarely become that critical as compared with the way the game is heading anyway.

[MartinCarpenter]

Good idea

Teams chess? Sometimes as 'extreme' as stuff like throwing a game away in a despairing effort to salvage a match.

The latter depends on how much the result matters of course On Saturday where we're playing Bradford with the Yorkshire league at stake for them at least. They'll avoid accepting a drawn match unless they're half a chess set down.

In say a random match in a local league where your team routinely finishes very firmly mid table? Would still try of course, but rather less prone to throw yourself off a cliff.

[Roger de Coverly]

Would you let a team result come ahead of these considerations?

The considerations as to whether to offer a draw in a level ending can be team based. If an eight board match is 4-3, it can be rational to offer or force a draw, rather than trying an Arkell style grind, even if in an individual tournament, you might continue. That's particularly so, if more often than not, you overestimate your chances and overreach.

[Alan Walton]

Your reaction is a very British old school one, and what I expected of course! But it doesn't change the fact. You simply don't like to acknowledge the reality. Chess players try to win their games for their own benefit - it just so happens that this naturally almost always coincides with the team's needs. But it is not the thing which really drives most people! You might be amazed to hear that I am very much regarded as a team player in all the teams I play for. However, I also like to acknowledge the truth of what motivates me!

No, it is because I play for a club (and other clubs) that always believes the result that matters for the team and not the individual

This is especially prevalent in tight matches against strong teams (for example in the 4NCL), having people with your point of view within the team can put undue pressure on the rest of the team members

Remember there is no "i" in Team

[John McKenna]

This thread reminds me of the bad old days of gore that we have not seen lately!

What's the point of having two grades per year that are both 'smoothed' and hence convergent?
Might as well just have one such grade (end of July) per year.
A second grade (end of January) that highlights divergence must be of some significance and use.

A two-state system that meets grading radicals halfway is preferable to a single-state one that satisfies only grading conservatives, and a reasonable compromise. (This grading glass is only either half empty or half full no matter how you look at it.)

After all, we have a two-state system of competitions - club & county matches that run continuously throughout the year, and tournaments that last a couple of weeks max. (Club & county time controls usually differ from tournament ones.)

Simply regard the (stable) July grades as primarily a match grade for the whole season and the (volatile) January grades as primarily a tournament grade for the second half of the season, and calculate them appropriately, differently.

[Carl Hibbard]

I would prefer at this time we all played nice.

[MartinCarpenter]

Oh definitely

If anyone wants to consider things, consider the following seasons (my games in Yorkshire):
2008_9 17 games at 174,
2009_10 12 games at 185,
2010_11 14 games at 175,
2011_12 11 games at 199(!),
2012_13 14 games at 168,
2013_14 12 games at 177
2014_15 24 games at 171,

ECF Standard grades: 177,179, 179, 185, 181, 176, 172
30 game window: n/a, 179, 176, 184, 181, 175, 174
Yearly Yorkshire grades :177, 179, 177, 190, 171, 179, 170
Grade bands of my live grades in those years: 177-185; 177-184; 173-179; 180-190; 171-193; 171-179; 170-184.

Which set is 'right'? Goodness knows You can see how ECF does smooth things out a bit though.

I rather prefer that myself, finding the Yorkshire live grades rather melodramatic. Others might prefer something different.
(Yorkshire grades get very stable if you play 40+ games a season, not really on this number of games.).

My 'real' ECF has more/different sets of games in it, so behaved rather differently to this. Contrived to arrange things to manage a 22pt drop in 6 months in 2013

[Carl Hibbard]

I think the changes were sensibly motivated, but won't make a huge amount of difference - certainly not enough to justify such a vitriolic post.

Some posts moderated mainly for a quieter life whilst the sun is out.

[MartinCarpenter]

Actually just one more post about this. Basically judging the old (new ) ECF style - the 30 game window one - as tracking list.

Two major flaws. Firstly your form in your last game wasn't judged vs your existing game, it was judged vs your form 29 games ago.

So you can have a bad result and if it replaces a horrible one your grade still goes up Ditto with a good one and your grade going down. Results tend to cluster a bit (a given tournament etc) so this sort of effect could definitely be sustained for periods of 5 games (1 congress!).

Really not desirable behaviour as it'll confuse people, and it did introduce a sort of 'delayed reaction' into the calculations.

The second thing is the principle of trying to track fluctuations in form using a 6 month grading list. It just doesn't work By the time you publish the list the form is often 3-4 months old, and by the time it stops being valid it'll be more like 10. People very rarely sustain runs of form for that long and again people won't be seeing their grade going up/down in response to their recent play.

In all honesty, with 6 month lists, I think you likely do better to just aim for a stable estimate. At least its good for something that way Get it more frequent and you can try and show runs of form. The Yorkshire live list does that OK, although mostly it just demonstrates how stable most peoples grades actually are.

[Maurice Lawson]

Mike, you make the point that the 30 game window was effectively replacing the current game, whilst dropping off one game from 30 games previous.

But the 12 month system is massively worse - As soon as you move past the most recent grading date a whole six months of games drops out of your records, and is slowly replaced by however many games you play during the next six months. It is much more perverse than the one by one replacement created by taking a fixed number of games.

As for your argument about six months information going out of date, well it seems to me that the point you make is far more applicable to twelve months. And even longer periods would increase stability yet further.

So, for all the discussion and points made it still seems to me that the argument for a twelve month system is primarily based upon the result being smooth, and that is regarded as a virtue by those who support it. However I simply don't believe that should be the primary purpose of the grading system, and I see no reason at all why that assumption should be made. On the contrary I regard smoothness as a failing in that it never tells me anything interesting. What is the point of publishing a number which has been so smoothed as to hardly ever vary? If you are looking for stability then why not base the grading on two years or even three years of game data. Then things will get very stable indeed, which surely is what the supporters of stability are looking for. I just find that whole approach very boring indeed, and it simply isn't what I'm looking for in a rating system. Neither it seems are many others - otherwise why is there so much interest in the live ratings which shift every time a game is played. Ideally that is exactly what I would like to see (although I'm not saying it's doable for the time being, but it certainly will be one day!) - it would immensely liven up the game and you would find people would be extremely interested in it. Only the practical reality of the data input at present prevents this from being possible, but it will certainly come in the future.

I don't believe that our playing performance (or our actual playing strength) actually does remain stable. It moves up and down over time. It seems that many like to describe this as noise, whilst I describe it as change in performance, and I would like to see it measured so as to recognise the real fight which is going on in the chess world. At present there is too much old style thinking pervading the ECF which is holding back the development of chess.

The revised system will hold up the grades of older people (like me) who are likely to decline in playing strength, whilst holding down the grades of people who are improving. This is absolutely perverse. It will be challenged when people really start appreciating what is happening, and as the ability emerges to process game data in near to real time. Whilst the old guys continue to hold the power, and want to keep it that way, it will be bad for the development of chess. But eventually the young chess players will simply sweep away all these old style arguments about stability and we shall start to make English chess a much more exciting sporting activity.