BRITISH CHESS CHAMPIONSHIPS - THE GAMES

[Alex Holowczak]

Right you are. Dunno what I'm watching on the live feed then but it's up to move 114!

Drawn after 118 moves. I suspect the board failed some time ago and they updated the live feed with moves from the scoresheet.

That's exactly what happened; I was the one who was in the wrong place at the wrong time and had to decipher the scrawls.

[andrew martin]

Game of the Day from Rd Four : Gawain Jones v Susan Lalic.


I hope you enjoy it!


http://www.youtube.com/watch?v=K-Ql0fjCuWY

[LozCooper]

Right you are. Dunno what I'm watching on the live feed then but it's up to move 114!

Drawn after 118 moves. I suspect the board failed some time ago and they updated the live feed with moves from the scoresheet.

That's exactly what happened; I was the one who was in the wrong place at the wrong time and had to decipher the scrawls.

I hope the black player has neater writing than Simon

[Matthew Turner]

Hypothetically, I wonder what odds I would get on six White wins on the top six boards? I reckon it should be about 150 or 200 to one and worth a punt.

[andrew martin]

I'm sure Charlie Storey would be happy to give you even money on that bet. Shall I put you in touch?

[Adam Raoof]

I'm sure Charlie Storey would be happy to give you even money on that bet. Shall I put you in touch?

Charlie Storey has withdrawn from the MO.

[John Sargent]

I'm sure Charlie Storey would be happy to give you even money on that bet. Shall I put you in touch?

Charlie Storey has withdrawn from the MO.

I didn't believe this until I woke up and looked at the car park!

[Simon Ansell]

I'm glad you didn't include board 7 , but I reckon it's about 300-1, probably higher to allow for possibilities of short draws.

[Richard Bates]

I'm glad you didn't include board 7 , ...

Quite right. A run of 10 consecutive white wins in FIDE rated tournaments can't go on much longer

[Maxim Devereaux]

Hypothetically, I wonder what odds I would get on six White wins on the top six boards? I reckon it should be about 150 or 200 to one and worth a punt.

Well, if I was a betting man, but didn't have access to any stats on these particular players, styles, past performance vs. opponent, current form etc. and had to do it based purely on general chess stats, I guess I'd have a go at calculating the odds like this:

Based on figures from Chesstempo (since I'm not at home, and don't have access to stats from Chessbase, which I might choose to limit by players over a certain rating), the breakdown of chess games is roughly 35% White wins, 40% draws, and 25% Black wins. So White scores 55%. A score of 55% - 45% is expected by a rating difference of 33-39 points, so when looking at the expected score between these players, we should add 36 to the difference between ratings to account for the in-built White advantage.

Also, given that overall, 40% of games are draw, but 25% are Black wins, Black is scoring 0.2 of his 0.45 expected score from draws, or 4/9. As the percentage of lower-rated wins probably decreases more sharply than that of draws, let us be a bit cautious and say that 50% of a lower rated player's expected score must come from draws, which will make the calculation simpler, in that the percentage chance of a draw will then be the same as the percentage score expected by the lower rated player.

Now we can look at the rating differences for today, add the White factor, calculate the expected score for each game, and use our guess as to the way in which the points are distributed to calculate the odds of White win, draw and Black win in each case.

Code: Select all

White      Black       diff      adj.diff      exp score for Black      White win      Draw      Black win
2687       2625         62          98                0.37               0.445         0.37        0.185
2715       2606        109         145                0.31               0.535         0.31        0.155
2533       2436         97         133                0.32               0.52          0.32        0.16
2489       2418         71         107                0.35               0.475         0.35        0.175
2301       2459       -158        -122                0.67               0.165         0.33        0.505
2457       2278        179         215                0.23               0.655         0.23        0.115

And finally, the chance of a win for all six Whites is just the product of the six figures in the White win column, i.e. 0.006355299 or 1 in 157.349, so Matt wasn't so very far off the mark originally, I'd say.

[Christopher Kreuzer]

Well, if I was a betting man, but didn't have access to any stats on these particular players, styles, past performance vs. opponent, current form etc.

Could you modify the odds given these past performances? Not had time to pick out colours.

Short-Howell - 2 draws
Adams-Jones - 1 win (for Adams)
Gordon-Hunt - 1 draw
Wells-Houska - 3 draws
Griffiths -R.Pert - no data
Hawkins-Rudd - 1 win (for Hawkins)

[Christopher Kreuzer]

I'm glad you didn't include board 7 , ...

Quite right. A run of 10 consecutive white wins in FIDE rated tournaments can't go on much longer

I can't find any games between the two of you. Must be in the larger databases. You must have played quite a few games. Maybe after the game is finished (starting in 15 minutes), one of you will be able to say what the trend is?

[Maxim Devereaux]

Well, it's not really easy to say how the odds should be modified. How much to weight the effect of 1-3 games played between the players? And do you weight results from games with the correct colours more than results with the wrong colour (probably, but how much again?). And what were the relative ratings of the players at the time these previous games were played (e.g. Adams probably outrated Jones by 200 points when he beat him before, as opposed to 100 now).

My approach was just quick and dirty; anything more needs some sort of predictive model, and entire papers and many computer programs have been written trying to do result prediction of this type, based on past results.

However, I would expect lower White-win chances on 1 and 3 than those given by the stats. Two draws in Short-Howell may be a slight indicator to that. Adam Hunt has also played well recently, having finished joint first, ahead of Stephen, in Leros which ended a week or so ago.

Against that, on 5, Griffiths is clearly under-rated and well-prepped, and maybe Richard P's Dragon can be prepped against (or some other part of his repertoire), so his percentage chance of victory is significantly higher than 16.5%, I'd hazard, but whether that offsets the chances of draws in some of the other games, I'm not sure.

[Maxim Devereaux]

4. Qe2 from Short. Avoiding any Berlins, and probably going to transpose into the Worrall attack.

Yet another 9. ...d5 Dragon in Adams-Jones, but Black has varied from the two previous games with 12. ... Bxd4 and then 13. ...Qb6 instead of 12. ...Nxc3.

[Daniel Young]

Déja-vu alert on board 2...