Skip to contents

Choose a move using a packaged difficulty knob

Usage

mxo_ai_move(game, difficulty = c("easy", "medium", "hard"), seed = NULL, ...)

Arguments

game

An mxo_game object.

difficulty

One of "easy", "medium", or "hard". The mapping is:

  • "easy" – one-ply heuristic; takes immediate wins or blocks immediate opponent wins, otherwise picks the best heuristic move.

  • "medium" – negamax depth 2 with branch_policy = "none".

  • "hard" – MCTS with at most 400 iterations, random rollouts and a two-second cooperative time budget.

seed

Optional integer seed for deterministic stochastic difficulty levels.

...

Named arguments override the chosen engine defaults. For hard difficulty, set time_budget = NULL for a fixed iteration count and reproducible seeded runs independent of machine speed. A timed search can examine different numbers of moves on different machines.

Value

A one-row legal-moves tibble (or a zero-row tibble at terminal states).

Examples

g <- mxo_new_game(n = 3L, k = 3L)
mxo_ai_move(g, difficulty = "easy", seed = 1L)
#> # A tibble: 1 × 5
#>   kind    L_src t_src   idx player
#>   <chr>   <int> <int> <int>  <int>
#> 1 present     0     0    13      1