Gibraltar : Draws in less than 30 moves

[Carl Hibbard]

The 3-1-0 probably doesn't reduce the actual number of draws but it puts a halt on coasting to draws in the last couple of rounds making a tournament (especially on-line) far more interesting?

I am all for it...

[Alex Holowczak]

When I get some free time, I'm going to get some statistics that debunk the 3-1-0 myth.

Good luck with that Alex. My recommendation would be to "do it" rather than "talk about it".

However, going from this preliminary set of data, I think Alex, you're going to have a pretty hard time showing that the 3-1-0 system is a "myth". As far as I can see it's a reality in action.

I have other priorities for the next few days. I'll do it over the weekend, writing it up properly in LaTeX and stuff.

Let's not count chickens before they're hatched.

[Peter Rhodes]

Let's not count chickens before they're hatched.

My advice stands as before - less talk - more action - "just do it" !!


I understand your point Sean, it's similar to arguing about the effectiveness of the DP as a deterrent - people compare America and Europe when the only real comparison is to compare America with the DP and America without the DP - which is of course impossible.

In these kind of scenarios it's never going to be possible to conduct a double-blind trial, so we just have to run with intuition and the limited value from the data before us.

Personally, I consider the difference between the two sets of data : 26% and 19% pretty significant, but I'm open to different interpretations.

[Alex Holowczak]

Let's not count chickens before they're hatched.

My advice stands as before - less talk - more action - "just do it" !!

I was about do it here (at University), but this Ubuntu doesn't have Texmaker, and the admins appear to have disabled sudo (i.e. I can't download it either).

[Peter Rhodes]

Can't you just grind some numbers old-style ?

[Alex Holowczak]

Can't you just grind some numbers old-style ?

Pffft. Where's the fun in that?

[IM Jack Rudd]

And today's effort...

[Event "8th Gibtelecom Chess Festival Masters"]
[Site "Gibraltar, England"]
[Date "2010.02.03"]
[Round "9.1"]
[White "Gustafsson, Jan"]
[Black "Bacrot, Etienne"]
[Result "1/2-1/2"]
[WhiteElo "2627"]
[BlackElo "2713"]
[PlyCount "29"]
[EventDate "2010.??.??"]

1. d4 d5 2. Nf3 Nf6 3. c4 c6 4. cxd5 cxd5 5. Nc3 Nc6 6. Bf4 a6 7. Rc1 Bf5 8. e3
Rc8 9. Be2 e6 10. O-O Nd7 11. Qb3 Na5 12. Qa4 Nc6 13. Qb3 Na5 14. Qa4 Nc6 15.
Qb3 1/2-1/2

[Ian Thompson]

At Coulsdon so far this year, there have been 544 games of which there have been 122 draws (22.43%).

This can be split into games that use the 3-1-0 system and games that do not use the 3-1-0 system.

The competitions that use the 3-1-0 system are :...

Under this category, there have been 303 games of which 59 resulted in a draw (19.47%).

The competitions that DO NOT use the 3-1-0 system are :...

Under this category, there have been 241 games of which 63 resulted in a draw (26.14%).

I thought the proponents of the 3-1-0 scoring system claimed that it results in more fighting chess than the normal scoring system. If this is the case then quoting figures on the percentage of games that are drawn is not the right thing to do (even if it was meaningful to compare the draw ratios in different types of event). You want to know how many games were drawn without either player making much effort to win, which can probably be estimated by looking at the number of short draws. Without this information I think your figures show nothing.

Taking the Wijk aan Zee A group as an example, 63% of the games were drawn, but only 9% of the games were drawn in under 25 moves, which suggests to me that there were plenty of hard-fought games that eventually ended in draws. In this event, which of these was the better game:

1. Smeets-Van Wely, where Van Wely misplayed the opening and lost in 23 moves
2. Short-Kramnik, where Short came very close to winning, but failed to find the endgame win and only drew (in 63 moves).

I vote for Short-Kramnik.

[Ola Winfridsson]

Peter, even a drawing percentage of 26 is well below the average!

[Sean Hewitt]

And today's effort...

[Event "8th Gibtelecom Chess Festival Masters"]
[Site "Gibraltar, England"]
[Date "2010.02.03"]
[Round "9.1"]
[White "Gustafsson, Jan"]
[Black "Bacrot, Etienne"]
[Result "1/2-1/2"]
[WhiteElo "2627"]
[BlackElo "2713"]
[PlyCount "29"]
[EventDate "2010.??.??"]

1. d4 d5 2. Nf3 Nf6 3. c4 c6 4. cxd5 cxd5 5. Nc3 Nc6 6. Bf4 a6 7. Rc1 Bf5 8. e3
Rc8 9. Be2 e6 10. O-O Nd7 11. Qb3 Na5 12. Qa4 Nc6 13. Qb3 Na5 14. Qa4 Nc6 15.
Qb3 1/2-1/2

Pitiful. Let's hope the arbiters have the stomach to do something about it.

[Jonathan Bryant]

Personally, I consider the difference between the two sets of data : 26% and 19% pretty significant, but I'm open to different interpretations.

The problem is that whether these figures are 'significantly' different - in a statistical sense or otherwise - isn't really the question. The issue for this thread is does 3-1-0 cause a difference. Your figures could be significantly different for reasons nothing to do with the scoring system.

Even if we assume your figures are not problematic for any reason to do with sample size or are not skewed in any way (frankly a rather geneorous assumption) I'd caution against generalising across the board. In particular I'm not sure we can necessarily assume that what goes for amateur chess players will hold true for professionals. Actually it seems to me there are very good reasons to think that professionals and amateurs will respond differently to various scoring systems.


That said, thanks for the stats you put up - they're interesting in their own right whether or not they prove anything for the subject at hand.

[Peter Rhodes]

In particular I'm not sure we can necessarily assume that what goes for amateur chess players will hold true for professionals.

Indeed Jonathan, this was a caveat I made in my message in which I outlined the data.


People react to incentives. I think it is as simple as that.

[Jonathan Bryant]

In particular I'm not sure we can necessarily assume that what goes for amateur chess players will hold true for professionals.

Indeed Jonathan, this was a caveat I made in my message in which I outlined the data.


People react to incentives. I think it is as simple as that.

Oh yes so you did - sorry I missed that in the middle there.

I think you're right about the people reacting to incentives thing but the trouble is incentives themselves aren't simple.

[Ola Winfridsson]

In particular I'm not sure we can necessarily assume that what goes for amateur chess players will hold true for professionals.

Indeed Jonathan, this was a caveat I made in my message in which I outlined the data.


People react to incentives. I think it is as simple as that.

That's indeed true, but I doubt that changing the scoring system to 3-1-0 is the right incentive. Too many draws, short or otherwise, is certainly not much of a problem at amateur level (in your limited set of data three quarters of all games are decisive with the normal 1-½-0 scoring system, I don't think that's much cause for concern). At the professional level I'd even say that the problem of short draws has been exaggerated. In so far as it is a problem it can be solved by various other means, such as not allowing sharing of prize money (through conventional tie-break systems or a playoff system like the one at Gibraltar, for instance).

[Roger de Coverly]

In so far as it is a problem it can be solved by various other means, such as not allowing sharing of prize money

In some cases it may be the non-sharing of prize money which is the problem itself. A professional player might well be prepared to play for a win in the last round on sporting grounds, but not if it caused them to lose a needed perceived minimum payment. So perhaps if the potential prize money is £ 2500 for a win , £ 2000 for a draw and £ 1500 for a loss, then you get a game , whereas if it's £ 4000 for a win, £ 2000 for a draw and £ nil for a loss, then you get a non-game.