Remove and close a file handler from the logger.

Compute residuals for a model where each tournament is predicted to be 50% 50%

Compute residuals for a model where each tournament is predicted to be the mean of the scores

Compute residuals for a model where each tournament is predicted using the logit of the first round's score. scores: shape (num_tourname

Compute residuals for a model where each tournament is predicted using the logit of the previous round's score. scores: shape (num_tourn

Execute a single training round.

Run the main agent for the current round.

Serve the folder's replays until interrupted, opening the index in a browser.

Runs tournament with the players and updates their measured elo.

Execute a single training round, i.e., run the game, then run the agent.

Higher score => better rank (0 is best). Ties share the same rank.

Update mirror agent's codebase with the main agent's changes.

Enable or disable static mode

Set up the Frozen-Flask freezer with proper configuration

Calculate Shannon entropy for a list of commands. Shannon entropy measures the "surprise" or information content in the distribution. -

Convert the GameContext to a dictionary for rendering prompts in the agent

* Update the existing chart with data for a new file

Process a single agent and update their agent_stats. Returns True if any updates were made.

Write LaTeX table comparing bootstrap metrics between methods. Args: bootstrap_results: Dictionary mapping bootstrap type to metrics dict

Write LaTeX table with ELO results to file. Args: results: Dictionary mapping game name to fit results output_dir: Directory to s

Write LaTeX table with plain ELO scores and uncertainties (no bar charts). Args: results: Dictionary mapping game name to fit results

Write results in JSON format for website consumption. Creates a leaderboard JSON with rankings per arena and overall. Format: { "

Args: files: Dict mapping file paths to their contents command_outputs: Dict mapping command prefixes to their output

The CodeArena class is responsible for running games, i.e., taking a list of code from different agents/players and running them against each

This class builds a win matrix from a log directory, it doesn't fit anything yet. It also adds a "ALL" game to the win matrix, which is the su

Fit Bradley-Terry model to a win matrix Args: win_matrix: Dictionary mapping player pairs to win counts regularizatio

Args: name: The name of the game. draw_probability: The probability of a tie in the game. repetitions: Th