https://github.com/king-witcher/lsre-4
A Rust tool for estimating player Elo ratings based on a match graph through Bayes Inference.
https://github.com/king-witcher/lsre-4
elo rating-system rust
Last synced: 7 months ago
JSON representation
A Rust tool for estimating player Elo ratings based on a match graph through Bayes Inference.
- Host: GitHub
- URL: https://github.com/king-witcher/lsre-4
- Owner: King-witcher
- Created: 2025-01-22T14:38:12.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-01-22T15:23:48.000Z (9 months ago)
- Last Synced: 2025-01-22T16:31:54.348Z (9 months ago)
- Topics: elo, rating-system, rust
- Language: Rust
- Homepage:
- Size: 16.6 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# How to use
## Extracting graphs
First, you will need a *LSRE JSON Graph*, which represents both the player names and the set of matches and the matches that you want to analyze.
A LSRE JSON Graph comes in the format of:
```ts
type GraphNode = {
name: string
id: number
}
type GraphEdge = {
winner_id: number
loser_id: number
}
/// The LSRE JSON Graph type
type Graph = {
players: GraphNode[],
edges: GraphEdge[]
}
```
Great examples of graphs can be found in the `graphs` directory, based on all Quake Pro League seasons.
The easiest way to get graphs is by scraping data from websites using predefined scripts, such as scripts found in the `scripts` directory, made for Liquipedia.
## Using a graph
Once you got a graph ready, just LSRE it with:
`cargo run --release graphs/my-graph.json`
And LSRE will spit out a JSON with the info you need.