1

After seeing a video of an analysis on the chess game in A Game of Shadows between Sherlock Holmes and Moriarty, it made me wonder what their fictional ratings would be around. Other games like the Harry Potter chess puzzle and Einstein vs. Oppenheimer have also interested me, so what would I analyze to predict their ratings?

A few ideas I have had was to compare general rating trends, such as ratings when openings are mastered, when prominent use of strategy is used, the speed (I think it was a 5-min blitz in AGOS. Also looking at how far they calculate ahead, the rate they calculate, how they prepare for future threats, etc. I could also look into the history and check the openings used, and the general theories of the time and how much they diverge.

I also note how these games are fictional and may not be perfect or even set up with the era in mind, but it still seems like an interesting thing to do.


Also, if you have any other examples they are appreciated.

closed as off-topic by Brian Towers, Phonon, Glorfindel, SmallChess, Ywapom Apr 20 '18 at 22:35

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "This question does not appear to be about chess within the scope defined in the help center." – Brian Towers, Phonon, SmallChess, Ywapom
If this question can be reworded to fit the rules in the help center, please edit the question.

  • I would appreciate if those who have closed this one can explain why it’s off-topic. It seems like an interesting, original, answerable question about concrete games of chess (historical or fictional). Particularly baffled by smallchess who gave a very good answer but then voted to close. Thanks – Laska Apr 26 '18 at 1:10
3

You will need statistics to estimate playing strength from a single game. It's of course possible, but you will be need to read journal articles.

https://ailab.si/matej/doc/Using_Heuristic-Search_Based_Engines_for_Estimating_Human_Skill_at_Chess.pdf

"Using Heuristic-Search Based Engines for Estimating Human Skill at Chess" is considered as a baseline for estimating human skills from a single game. The idea is simple; assume a computer engine plays much stronger chess than humans; we should be to use the difference in evaluations for statistical inference.

I have personally implemented this paper for a chess company. The system is working.

2

Trying to estimate rating based on a single game is just about impossible or at least will come with a very large error margin. Even more so, if you don't know either players rating.

The ideas you have are all very valid, but will in the end only give you the difference between absolute beginners and reasonably advanced players. Opening knowledge, etc might be acquired at very different rating levels Lack of it can be compensated by say better tactical knowledge. Or said differently, you might have two players at the same rating level where one is a good tactician (and poor at openings) and the other particularly good at openings (and poor at tactics).

In a single game you are hardly going to be able to assess all the factors mentioned by you.

When it comes to more advanced players who are reasonably good at everything, it will become even more difficult. A player of say 2200 strength will play more or less similarly to a GM, just that the GM is on average better at all the points you mention.

Another thing to take into account is that just looking at the moves you are not able to assess all the evaluation/thinking/calculation being done by the players, which will be more/in shorter time by stronger players.

Not the answer you're looking for? Browse other questions tagged or ask your own question.