chess-heatmaps

Heatmap representation of chess moves, it shows the squares that each piece has been moved to more times, it can also show what's the differences between the moves from each player.

Usage

I only tested with pgn files from the lichess.org website but most of them follow the same format. I'll leave in a assets/ folder one example of a pgn file in case you want to have a look.

Make sure you give a path for the pgn file and the username of the player and it will only look for white games as default.

The the heatmaps are from the pawns, bishop, knight, rooks, queen, king, respectively

Options

• -u , username - mandatory to provide
• -p , path to pgn file - mandatory to provide
• -b, look for black games of a certain player, it will only look for white as games as default
• -t , to option - means that it will only consider the first x moves
• -f , from option - it will only search from the x move foorward
• -l, activate the logarithmic difference between moves. logarithmic difference is a cheap way to calcular the ratios between 2 numbers

Here is an example in how you to run the simpler way. In this example I'm looking for the white games of the user "jotaalvim" and I also provide a pgn file whith his games

$ python densidade.py -u jotaalvim -p assets/lichess_jotaalvim_2022-03-12.pgn

Here is an example of the heat map representation of the same user as before but now it filters the chess moves from the move 20 to the move 40.

python densidade.py -f 20 -t 40 -u jotaalvim -p assets/lichess_jotaalvim_2022-03-12.pgn

The output is the following:

I'll add the way to calculate differences between players soon

Outputs

At the moment the output goes to the "ouputs" folder to the heatmap.png file

Dependencies

Make sure you have jjcli module instaled, you can install it by:

pip install jjcli

I also use mathplotlib

pip install matplotlib

some generated examples

These are the heatmaps from the knight, bishop, queen, rooks , pawns and the king, respectively, in the player's white games.

---

Heatmaps from a weak player (800 blitz lichess)

---

Heatmaps from a medium/average player (2000 blitz lichess)

Heatmaps from a very strong player (2700 blitz lichess)

It's cool but not that informative

From simply looking at the maps, there's not much information that we can collect, we can perhaps conclude that the weaker players often move the queen to the more obvious d1-h5 diagonal instead of placing in c2 or d2. Such thigs happen a lot when using different openings. Another, is that the dark bishop in played fewer times by the weaker and medium players then by the stronger player. It's impossible to conclude something interesting due to the fact that people are probably playing different openings.

Differences between players

A more useful and interesting data would be the logarithm differences between the moves from each player. Here is an example of the differences between a stronger and a medium player, in this case the squares more yellow/highlighted are squares not played as frequently by the medium player compared to the strong one, and therefore tend to be "better" squares for the pieces.

Opening analysis

The heatmap but restricting the search for only the x amount of moves. Here's an example of the first 10 moves from:

weak player games

---

medium player games

---

strong player games

---

Where does each piece captures others?

Here's an example of the squares where each piece tends to capture others:

todo

• Where does each piece goes to die
• Where does each piece stays the longest