Thursday 29 May 2014

Graphical representation of chess data

I was tipped off* the other day about a series of posts that presented chess related data in graphical ways. The author of these posts, Randal Olson, looked at a number of different chess data sets, covering varied topics, and found a nice way to present the information.
A lot of the work is historical analysis, for example charting the popularity of various openings over the years. In the early years (his starting point is 1850), 1.e4 reigned supreme, but by the 1930's was surpassed by 1.d4. But from 1950 1.d4 became less popular and 1.e4 once again become the most popular opening move.
As a lot of the charts are percentage based, there are some surprising results. In the 1850's the Pirc Defence was played in over 10% of recorded games, although this can be attributed to the Cochrane - Bannerjee matchs, which contributed a sizeable number of games to the data set for those years.
Other topics covered are ratings, length of games, percentage of draws, and even how many games end in checkmate. The work seems to be ongoing, so I'll provide a link to his first chess related post, and you can find your way from there. http://www.randalolson.com/2014/05/24/chess-tournament-matches-and-elo-ratings/

(* It was my brother who alerted me to this work although starting his email with "Hey nerdlinger" is not the way to ensure a Christmas card this year)

No comments: