With careful play, I've won most of my recent games. But there is a lot of room for improvement!
Unsurprisingly, accuracy is a good predictor of who wins a game. Whether both players are playing with relatively poor accuracy, high accuracy, or one player is playing with markedly higher accuracy than the other, the winner is usually the player who makes the most accurate moves (according to the "accuracy" percentage given during a chess.com post-game analysis). This is the case in nearly all games so far, actually. My own accuracy ranges from about 40% to 95+% from game to game, as I am decidedly more capable in certain positions and under certain circumstances. When the tools are more robust, I can probably find statistical insights there about when I am playing the most accurately, when I am playing the least accurately, and exactly what to study to play with consistently high accuracy.
Finished an early version of the game data analysis database and graph visualization tools. It has some kinks to work out but it's functional now. In the future I want to hook it up to the chess.com API (I've been told one exists) and use it to automatically convert my games to the JSON database I have here. I also want to make a page for the site which replays chess games with some extra analysis (not the same stuff which the chess.com post-game analysis does, but some other things) and just for giggles. I will also want to hook it up to the Lichess API (if there is one), and scrape my several hundred games from that platform.
The Graph visualization, I am writing from scratch as a math exercise (it's got some unique wrinkles), and the project in general is an opportunity to get more proficient with JavaScript and HTML5 in general. The Database is just a big JSON file which I am importing games to manually via a node script. Eventually, I will find a more robust database solution and make the whole project more modern, as it could be a decent long-term project in itself.
In the very long term it would be neat to hook the whole thing up to a chess engine and start doing some analysis that way, just to get some experience in how to make something like the chess.com or Lichess game analysis tools. While duplicating them would probably not be worthwhile, I may find things to do which they aren't doing, so that it could be a useful supplemental tool. Similarly, I am trying not to duplicate all the graphs in the chess.com "Insights" section, and instead look for graphs I can make that explore slightly different things, as their tools are really good already.
Currently, the database is only tracking the games I've manually input since March 19th, 2023, using hand-picked metrics, all played online through chess.com's Android app. In the future I will want to automate scraping the data on all my games from the platform itself, and retrofit the database to include whatever metrics they offer in addition to my own. I will eventually want to start a separate category for my Lichess games, and also for games not played on an online platform. Eventually, this will result in a data structure that combines all of the metrics from these things into something universal for my purposes, and possibly useful in general.
I suppose this project and site will also serve as a kind of blog, given the format, but its scope is restricted to chess and the tools for analyzing my game as I improve.
back to index