385feeb6ba
- When there are less than 5 players, and we're near the discard threshold, prefer hinting over discarding, even if there are known useless cards. - We now ask questions like "what's the first playable card in this list?" This means that if a playable card is in the asking player's list, the player will learn that it's playable, and that every card before it is not playable. Additionally, if a player doesn't know of any dead cards in their hand and there is enough information available, we use this mechanism so that if the player doesn't have a playable card, they will learn about one dead card in their hand. (These were two commits that got joined in a rebase accident, sorry.)
2.8 KiB
Simulations of Hanabi strategies
Hanabi is a cooperative card game of incomplete information. Despite relatively simple rules, the space of Hanabi strategies is quite interesting. This project provides a framework for implementing Hanabi strategies in Rust. It also explores some implementations, based on ideas from this paper. 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).
Please feel free to contact me about Hanabi strategies, or this framework.
Most similar projects I am aware of:
- https://github.com/rjtobin/HanSim (written for the paper mentioned above)
- https://github.com/Quuxplusone/Hanabi
Setup
Install rust (rustc and cargo), and clone this git repo.
Then, in the repo root, run cargo run -- -h to see usage details.
For example, to simulate a 5 player game using the cheating strategy, for seeds 0-99:
cargo run -- -n 100 -s 0 -p 5 -g cheat
Or, if the simulation is slow, build with --release and use more threads:
time cargo run --release -- -n 10000 -o 1000 -s 0 -t 4 -p 5 -g info
Or, to see a transcript of the game with seed 222:
cargo run -- -s 222 -p 5 -g info -l debug | less
Strategies
To write a strategy, you simply implement a few traits.
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).
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
- A cheating strategy, using
Rc<RefCell<_>> - The information strategy!
Results
On seeds 0-9999, we have these average scores and win rates:
| 2p | 3p | 4p | 5p | |
|---|---|---|---|---|
| cheat | 24.8600 | 24.9781 | 24.9715 | 24.9570 |
| 90.52 % | 98.12 % | 97.74 % | 96.57 % | |
| info | 20.9745 | 24.6041 | 24.8543 | 24.8942 |
| 04.40 % | 75.07 % | 89.59 % | 91.53 % |
To reproduce:
n=10000 # number of rounds to simulate
t=4 # number of threads
for strategy in info cheat; do
for p in $(seq 2 5); do
time cargo run --release -- -n $n -s 0 -t $t -p $p -g $strategy;
done
done