dumbmatter
commented
4 months ago I appreciate your interest, but I recommend you don't do this, it's too difficult! Some complications:
- Same feature should exist for all 4 sports
- It's not just clock and score that matter, it's all the "state" of the game, such as who has the ball, which player is shooting FTs, how many timeouts left, etc.
- Chart needs to handle multiple events happening at the same time on the x-axis (if the x-axis is game clock) such as FTs with a stopped clock
- Live sim code is pretty messy, not the easiest thing to work at. Might need some serious refactoring to give you access to the info you need to calculate a win probability
- My wife just had a baby so I am less available than normal to help
So this is probably something better for me to do myself eventually.
Now in theory if you come up with something that solves all those edge cases and any others that might exist, that'd be fantastic. But it's a big complicated task.
Also, for this and any other contribution, make sure you're okay with the license (not open source) and CLA https://github.com/zengm-games/zengm/blob/master/README.md#license-and-contributor-license-agreement
This is in reference to https://github.com/zengm-games/zengm/blob/d4ad5f2ac58c069e86cb5606163c4162fb565633/TODO#L99C1-L99C47 I believe that I may be able to complete this.
I anticipate having to make a component such as a BoxScore.probability.basketball.tsx which can be placed within BoxScore.basketball.tsx that will take perhaps the playbyplaylogger of the game and the two TeamGameSim objects and compare stats between the two teams to determine a baseline win probability and change depending on current score and current stats of the game.
Could maybe use parts of how the over under is calculated though I currently cant find how it is implemented.
Additionally, should I was thinking I would something similar to PlayerGraphs/ScatterPlot.tsx to graph the points.
Guidance would be apprecaited! Thank you.