Main objective: Guess the elo rating of a player having as input a list of their games (probably more than one)

Key Questions:
- Is there a better method rather than measuring the move accuracy?
- A partial answer to the first question is the opening. Classify which openings are most likely to be played in each rating level, and in what degree of accuracy.

Dataset: For a free dataset, lichess's dataset is massive ( https://lichess.org/api ). The issue is that it mainly focuses on blitz games, which would make any rating estimator much harder
For classical games, the biggest and most reliable one is chessbase, which is not free. If anyone has access to it, that'd be great. If not, there are other free options like https://theweekinchess.com/twic but they 're smaller in size and less organized.

For now, I 'd just want to see if there are people interested in it before deciding whether to move forward with this, so do contact me on Slack if you are!