hanabi.rs/README.md

78 lines
2.9 KiB
Markdown
Raw Normal View History

2016-03-19 22:14:29 +01:00
# Simulations of Hanabi strategies
2016-03-30 07:24:29 +02:00
Hanabi is a cooperative card game of incomplete information.
Despite relatively [simple rules](https://boardgamegeek.com/article/10670613#10670613),
the space of Hanabi strategies is quite interesting.
2016-04-02 23:29:46 +02:00
This project provides a framework for implementing Hanabi strategies in Rust.
2016-03-30 07:24:29 +02:00
It also explores some implementations, based on ideas from
[this paper](https://d0474d97-a-62cb3a1a-s-sites.googlegroups.com/site/rmgpgrwc/research-papers/Hanabi_final.pdf).
2016-04-02 23:29:46 +02:00
In particular, it contains an improved version of their "information strategy",
which achieves the best results I'm aware of for games with more than 2 players ([see below](#results)).
2016-03-19 22:14:29 +01:00
2016-04-02 23:29:46 +02:00
Please feel free to contact me about Hanabi strategies, or this framework.
2016-03-19 22:14:29 +01:00
2016-04-02 23:29:46 +02:00
Most similar projects I am aware of:
2016-03-19 22:14:29 +01:00
- https://github.com/rjtobin/HanSim (written for the paper mentioned above)
- https://github.com/Quuxplusone/Hanabi
## Setup
2016-04-02 23:29:46 +02:00
Install rust (rustc and cargo), and clone this git repo.
2016-03-19 22:14:29 +01:00
2016-04-02 23:29:46 +02:00
Then, in the repo root, run `cargo run -- -h` to see usage details.
2016-03-19 22:14:29 +01:00
2016-04-02 23:29:46 +02:00
For example, to simulate a 5 player game using the cheating strategy, for seeds 0-99:
2016-03-19 22:14:29 +01:00
```
2016-04-02 23:29:46 +02:00
cargo run -- -n 100 -s 0 -p 5 -g cheat
2016-03-19 22:14:29 +01:00
```
2016-04-02 23:29:46 +02:00
Or, if the simulation is slow, build with `--release` and use more threads:
2016-03-31 19:17:22 +02:00
```
time cargo run --release -- -n 10000 -o 1000 -s 0 -t 4 -p 5 -g info
```
2016-04-02 23:29:46 +02:00
Or, to see a transcript of the game with seed 222:
```
2016-04-02 23:29:46 +02:00
cargo run -- -s 222 -p 5 -g info -l debug | less
2016-03-31 19:17:22 +02:00
```
2016-03-19 22:14:29 +01:00
2016-04-02 23:29:46 +02:00
## Strategies
2016-04-04 09:26:42 +02:00
To write a strategy, you simply [implement a few traits](src/strategy.rs).
2016-04-02 23:29:46 +02:00
The framework is designed to take advantage of Rust's ownership system
so that you *can't cheat*, without using stuff like `Cell` or `Arc` or `Mutex`.
Generally, your strategy will be passed something of type `&BorrowedGameView`.
This game view contains many useful helper functions ([see here](src/game.rs)).
If you want to mutate a view, you'll want to do something like
`let mut self.view = OwnedGameView::clone_from(borrowed_view);`.
An OwnedGameView will have the same API as a borrowed one.
Some examples:
- [Basic dummy examples](src/strategies/examples.rs)
- [A cheating strategy](src/strategies/cheating.rs), using `Rc<RefCell<_>>`
- [The information strategy](src/strategies/information.rs)!
## Results (auto-generated)
2016-03-19 22:14:29 +01:00
2016-04-01 09:14:13 +02:00
To reproduce:
```
time cargo run --release -- --results-table
```
To update this file:
```
time cargo run --release -- --write-results-table
2016-04-01 09:14:13 +02:00
```
On the first 20000 seeds, we have these average scores and win rates:
2019-03-09 14:41:18 +01:00
| | 2p (stderr) | 3p (stderr) | 4p (stderr) | 5p (stderr) |
|---------|------------------|------------------|------------------|------------------|
| cheat | 24.8594 (0.0036) | 24.9785 (0.0012) | 24.9720 (0.0014) | 24.9557 (0.0018) |
| | 90.59 % (0.21 %) | 98.17 % (0.09 %) | 97.76 % (0.10 %) | 96.42 % (0.13 %) |
| info | 22.3249 (0.0128) | 24.7278 (0.0046) | 24.8919 (0.0029) | 24.8961 (0.0027) |
| | 09.81 % (0.21 %) | 80.54 % (0.28 %) | 91.67 % (0.20 %) | 91.90 % (0.19 %) |