Do openings affect your rating directly?
„I didn’t like that opening“, he said.Everybody, whether online or OTB, already experienced that: Your opponent is playing an opening you don’t like or don’t know, and suddenly a piece is missing, mostly one of yours. So, do specific openings harm your rating? I wanted to find out.
But let’s start some days ago...
The beginning
I had to cure my boredom, and while playing on Lichess, suddenly a link to the database, with all games ever played on this website, popped into my eyes. Even though I already was aware of the database, I never thought about how I could use such a huge amount of data.
Even only the games of last month had a zipped file size of 24 gigabytes
That’s quite a lot and I wanted to use that. So, I started coding, and not long after that, I was able to read Lichess PGN files. For those who don’t know: PGN is the file format in which all the games on Lichess (and literally anywhere else) are saved.
Now I had to think about the analysis itself...
Analysis
It was pretty clear what exactly has to me calculated: The average rating change after a game with a specific opening. Luckily, Lichess already provides this information in the PGN files. So it would all be pretty simple.
So I broke it further down:
Opening with less than 10 games played will be eliminated. Basically because new players to chess have no idea of opening theory, so only a few, very exotic games were played with very exotic openings, resulting in drastically change because of the rating deviation of new players.
Further, to filter high deviation games (games used to initially rate the player), I only used games where both ratings changed by a similar amount of points.
Then I executed three turns of analysis: I filtered the games by average rating.
The first turn only contained games below a rating of 1300.
The second one contained games between 1300 and 2000.
The third turn contained games with more than 2000 rating points on average.
The results
All ratings, 123981 games analyzed:
I think, some explanations would be very helpful now. At the lower edge of this picture you can see some strange combinations of letters and numbers. Those are the ECO Codes. ECO stands for "Encyclopedia of Chess Openings", where all the important openings are indexed.
On top of each bar is a number. Those represent the average rating gain in perspective of white. So negative numbers mean that those openings are advantageous for black.
For this graph, I'll copy the "extreme" opening keys here. You can look 'em up yourself: List of chess openings - Wikipedia
- B73 Sicilian, Dragon, Classical, 8.0-0
- C71 Ruy Lopez, Modern Steinitz Defence including Noah's Ark Trap
ALL THE OTHER GRAPHS
It actually looks interesting. Quite a lot openings from the graph above do not occur in the lower bandwidth.
WHAT ARE YOUR THOUGHTS?
First of all, I want to excuse for my eventually bad language use. I am not a native english speaker. Next, thank you for reading this. If you have any ideas for further coding projects (what would you personally be interested in?).
And discuss everything in the comments.
Thank you :)
