Chess 960 tournaments: variety

I agree with what you say.. it seems that certain SP tend to favour players with black. (i believe I mentioned that too in my previous posts)

Kokino also you mentioned the difference on one specific config  between the earlier computer DB,  the current one, and your 250000 game DB. But I would be interested to see the same query I ran  -basically just sort the 960 positions from your DB by winning % (not overall %) for white/black respectively and see if it lines up with what I got for the current computer DB - not in terms of specfic row correlation, but the overall distribution.  Of course I guess you would have to make the DB available as well and maybe you're not ready to do that.

If its computer played games, computers are looking several moves ahead.  Sure, in some 960 configs white can advance a pawn on his first move and have his bishop bearing down on an unprotected pawn of black, on the same diagonal possibly as a black rook behind it that cannot move.  But the first move by black he can protect that pawn, so white will not benefit from it. Even a half-way decent player could miss something like that though, especially if he's never played 960.  That is why the results in group tournaments between 960 groups and other groups are really skewed.

I realize that you mentioned that, but can you give a specific initial position and explain how it favors Black ?

That only suggests that the White, being unfamiliar with the position, may make a certain opening move that looks immediately attractive but turns out to be bad. It doesn't mean that it was the only move in the position, or that the initial position objectively favors Black. In fact, this supports my suspicion that even the top players play the opening only semi-competently in 960.

I didn't mean that was a position favoring black.  Just that that particular feature doesn't really favor white.  For another reason, black can do the same thing to white, if white doesn't protect its own pawn in the corresponding position.  What is the only difference between black and white whether its 960 or standard - white gets to move first.  I cannot think of a 960 position where white's advantage is somehow multiplied over the advantage in the traditional setup simply because he gets to move first.  Whatever he can do to black, black can do to him with the disadavantage of course of having to move second.

It may be that, with more games and more developed 960 opening understanding, we wouldn't be seeing significant differences in White / Black percentages from standard chess. Or it may be that certain positions do favor White more, or less, than the standard position. Only time and practice (and maybe theory) can show this.

grivei, I was speaking about chess magazines printed on paper. I am not sure if such information can be also found in databases or free online resources.

Great thread! I already picked up the organizational aspects in the first post and used them in an offsite post on my C960 blog...

Practical Issues for Hosting Chess960
http://chess960frc.blogspot.com/2009/12/practical-issues-for-hosting-chess960.html

...Now I see that the discussion has shifted to other aspects of C960 and that my name has been brought into it (starting with #16 by Eberulf). Yikes! I'd better be careful what I say, so I'll come back later when I've digested it all. - Mark

I'm back and will start with a couple of posts by Eberulf.

---

31st March 2010, 07:17pm; by Eberulf : 'He leaves out the most important stat it seems - the average expected advantage for white for a random 960 position.  He gives the highest scoring starting 960 positions for white and black, so why not the average.  For what purpose would that be left out.'

You're referring to this post...

Theoretical Advantage in Chess960 Start Positions
http://chessforallages.blogspot.com/2008/12/theoretical-advantage-in-chess960-start.html

...which starts, 'One of the nagging questions about chess960 relates to the fairness of the different start positions. Do White and Black have the same chances from any initial position or do some positions naturally favor one color?'

I left out the average of all positions because it wasn't an objective. I was looking for start positions that favor one color.

---

3rd April 2010, 02:32am; by Eberulf : 'Compare that to Weeks results - in his first table its 76% white vs. 41% black. Even in his second table he has 61% white vs. 53% black.  That's WAY off from my results, and my results are correct.  Not to rub it in as anyone can make a mistake I guess.'

You've misinterpreted my results, the column header says '%White', which I explained in the footnote: 'The column labelled '%White', taken from the original data, was already calculated using one point for a win plus one-half for a draw.' A %White = 41% means 59% for Black.

Discussing a botched table is counterproductive. It is more interesting to focus on the corrected table...

Advantage in Chess960 Start Positions Revisited
http://chessforallages.blogspot.com/2009/01/advantage-in-chess960-start-positions.html

...where I am more confident of the data and the calculations. Here you say, 'in his second table he has 61% white vs. 53% black'. I'm not sure to what you are referring. The important column is 'overall score for White (%S)', from which the overall score for Black can be derived.

---

Back to my original blog post on the subject ('Theoretical Advantage in Chess960 Start Positions') you wrote, 'it seems possible to have been compiled originally for propoganda purposes or something'. How could you possibly know what my motives were? - Mark

This is to Kokino.

---

3rd April 2010, 03:35am; by kokino: 'clearly, this is wrong as he is giving 50% wins for black and obviously, it should be for white with an overal score of 61%.'

The numbers I gave are the numbers calculated by CCRL. Even the current version (April 13, 2010) shows SP518:RNBQKBNR at less than 50% overall for White. As you pointed out, the data sample is too small to be reliable.

---

Incidentally, I just noticed that if you click on the CCRL column headers...

CCRL 404FRC : Downloads
http://www.computerchess.org.uk/ccrl/404FRC/opening_report_by_eco.html

...the data is sorted by the values in that column. There is no need to download and calculate. The current best SP for White is RKNBNQBR 68.2%; for Black it is RNKRBBNQ 37.9%.

BTW, the CCRL link was indeed broken for a month or so. - Mark

The first paragraph is a bold, almost outrageous, statement. Have you heard this directly from GM participants at Mainz or are you basing it on some other, perhaps anecdotal, experience?

In one of his radio interviews, Fischer sniggered at Kasparov's suggestion to limit the number of start positions in use. The idea protects the interests of GMs more than anyone else. - Mark

Fischer didn\t invent 960, it was an adaption from Shuffle chess, and hasn\t got any kind of monopoly on it. Certainly he memorized a load of opening variations when he was in standard play.

Fischer didn't invent shuffle chess, but he did invent the castling rules, including the restrictions on the initial placement of the King and Rooks. This is what sets chess960 apart from other types of shuffle chess and brings a random start position quickly into patterns similar to traditional chess (SP518: RNBQKBNR). So, yes, he did invent chess960. - Mark

Are those stats available online?  I'm curious how big the discrepancy is.  How much of an advantage does white have in the worst case?

white will always have the advantage as long as u stay with the plan no mis.

how do you play chess 960?

a