package macondo

Get desktop application:
View/edit binary Protocol Buffers messages

• optional GameHistory game_history = 1

• optional EvaluationRequest evaluation_request = 2

• BotRequest.BotCode bot_type = 3

• int32 millis_remaining = 4

Used in: ipc.GameRequest, BotRequest

• HASTY_BOT = 0

• LEVEL1_COMMON_WORD_BOT = 1

• LEVEL2_COMMON_WORD_BOT = 2

• LEVEL3_COMMON_WORD_BOT = 3

• LEVEL4_COMMON_WORD_BOT = 4

• LEVEL1_PROBABILISTIC = 5

• LEVEL2_PROBABILISTIC = 6

• LEVEL3_PROBABILISTIC = 7

• LEVEL4_PROBABILISTIC = 8

• LEVEL5_PROBABILISTIC = 9

• NO_LEAVE_BOT = 10

• SIMMING_BOT = 11

• HASTY_PLUS_ENDGAME_BOT = 12

• SIMMING_INFER_BOT = 13

• FAST_ML_BOT = 14

  harder bots here? inference? ml, etc.

• RANDOM_BOT_WITH_TEMPERATURE = 15

• SIMMING_WITH_ML_EVAL_BOT = 16

• CUSTOM_BOT = 20

  custom bots can be used to run custom algorithms. They should be used with the BotSpec below.

• UNKNOWN = 100

  This should have probably been number 0.

• oneof response

  • GameEvent move = 1

  • string error = 2

• optional Evaluation eval = 3

• string game_id = 4

Used in: ipc.GameRequest, ipc.ServerChallengeResultEvent, GameHistory

• VOID = 0

• SINGLE = 1

• DOUBLE = 2

• FIVE_POINT = 3

• TEN_POINT = 4

• TRIPLE = 5

EndgameMove is a single move within an endgame variation

Used in: EndgameVariation

• string move_description = 1

• int32 score = 2

• int32 move_number = 3

EndgameVariation holds a move sequence from endgame analysis

Used in: TurnAnalysis

• repeated EndgameMove moves = 1

• int32 final_spread = 2

Used in: BotResponse

• repeated SingleEvaluation play_eval = 1

Used in: BotRequest

• string user = 1

  Evaluate for this user

  Later: evaluation type (static/sim/etc)

GameAnalysisResult contains analysis results for a completed game

Used in: analysis_service.GetAnalysisResultResponse, analysis_service.SubmitResultRequest

• repeated TurnAnalysis turns = 1

• repeated PlayerSummary player_summaries = 2

• int32 analysis_version = 3

• string analyzer_version = 4

This should be merged into Move.

Used in: ipc.ServerGameplayEvent, BotResponse, GameHistory, PuzzleCreationResponse, puzzle_service.AnswerResponse

• string nickname = 1

• string note = 2

• string rack = 3

• GameEvent.Type type = 4

• int32 cumulative = 5

• int32 row = 6

• int32 column = 7

• GameEvent.Direction direction = 8

• string position = 9

• string played_tiles = 10

• string exchanged = 11

  An event will not have all of these; it depends on the type of the event.

• int32 score = 12

• int32 bonus = 13

• int32 end_rack_points = 14

• int32 lost_score = 15

• bool is_bingo = 16

• repeated string words_formed = 17

  words_formed is a list of all words made by this play, in user-visible pretty form. The first word is the "main" word, anything after it are cross-words.

• int32 millis_remaining = 18

• uint32 player_index = 19

  The player who played this move is encoded in player_index. This should be the index in GameHistory.players.

• uint32 num_tiles_from_rack = 20

  num_tiles_from_rack lets us know how many tiles from the rack were either played or exchanged. It is only populated for those two fields.

Used in: GameEvent

• HORIZONTAL = 0

• VERTICAL = 1

Used in: GameEvent

• TILE_PLACEMENT_MOVE = 0

• PHONY_TILES_RETURNED = 1

• PASS = 2

• CHALLENGE_BONUS = 3

• EXCHANGE = 4

• END_RACK_PTS = 5

• TIME_PENALTY = 6

• END_RACK_PENALTY = 7

  Only for international rules (or after 6 zeroes)

• UNSUCCESSFUL_CHALLENGE_TURN_LOSS = 8

  Lose a turn for challenging a word incorrectly (only for double challenge)

• CHALLENGE = 9

  Issue a challenge

GameHistory encodes a whole history of a game, and it should also encode the initial board and tile configuration, etc. It can be considered to be an instantiation of a GCG file.

Used in: game_service.GameHistoryResponse, ipc.GameEndedEvent, ipc.GameHistoryRefresher, BotRequest, puzzle_service.PuzzleResponse

• repeated GameEvent events = 1

• repeated PlayerInfo players = 2

  players are in order of who went first.

• int32 version = 3

• string original_gcg = 4

• string lexicon = 5

• string id_auth = 6

• string uid = 7

• string title = 8

• string description = 9

• repeated string last_known_racks = 10

  last_known_racks should only be set in an incomplete / in-progress game. if set, player racks should be set to these values.

• bool second_went_first = 11

  Add letter distribution (basically game name), board config, and more in the future.

• ChallengeRule challenge_rule = 12

• PlayState play_state = 13

• repeated int32 final_scores = 14

  Final scores of the game; the order is in the order of the listed players!

• string variant = 15

  The variant is the game variant used. If blank, should default to "classic".

• int32 winner = 16

  The index of the player who won. It's not always the person with the highest score, because there can be timeouts, etc. If it's a tie, it will be a -1.

• string board_layout = 17

  The board layout is just the name for the layout of the board. It should have a sensible default, if blank.

• string letter_distribution = 18

  The letter distribution is the name of the distribution of tiles used for this game. If blank, should default to "english".

• string starting_cgp = 19

  If provided, the starting CGP is a crossword-game position string.

Used in: TurnAnalysis

• PHASE_EARLY_MID = 0

• PHASE_EARLY_PREENDGAME = 1

• PHASE_PREENDGAME = 2

• PHASE_ENDGAME = 3

Used in: TurnAnalysis

• NO_MISTAKE = 0

• SMALL = 1

• MEDIUM = 2

• LARGE = 3

PEGOutcomeInfo holds the outcome for a specific tile draw in PEG analysis

Used in: PEGPlayInfo

• string tiles = 1

• PEGOutcomeType outcome = 2

• int32 count = 3

Used in: PEGOutcomeInfo

• PEG_OUTCOME_UNKNOWN = 0

• PEG_OUTCOME_WIN = 1

• PEG_OUTCOME_DRAW = 2

• PEG_OUTCOME_LOSS = 3

PEGPlayInfo holds per-play statistics from the pre-endgame solver

Used in: TurnAnalysis

• string move_description = 1

• int32 score = 2

• string leave = 3

• bool is_bingo = 4

• double win_prob = 5

• repeated PEGOutcomeInfo outcomes = 6

• bool has_spread = 7

• double avg_spread = 8

• bool is_played_move = 9

• bool is_ignored = 10

Used in: ipc.ServerGameplayEvent, GameHistory

• PLAYING = 0

• WAITING_FOR_FINAL_PASS = 1

• GAME_OVER = 2

Used in: GameHistory

• string nickname = 1

• string real_name = 2

• string user_id = 3

  user_id is an internal, unchangeable user ID, whereas the other two user identifiers might possibly be mutable.

PlayerSummary contains aggregate statistics for one player

Used in: GameAnalysisResult

• string player_name = 1

• int32 turns_played = 2

• int32 optimal_moves = 3

• double avg_win_prob_loss = 4

  Average losses

• int32 small_mistakes = 6

  Mistake breakdown

• int32 medium_mistakes = 7

• int32 large_mistakes = 8

• double mistake_index = 9

  Overall rating

• double estimated_elo = 10

• int32 available_bingos = 11

  Bingo tracking

• int32 missed_bingos = 12

PlyStats holds per-ply score and bingo statistics for a simmed play

Used in: SimmedPlayInfo

• int32 ply = 1

• double score_mean = 2

• double score_stdev = 3

• double bingo_pct = 4

Used in: PuzzleGenerationRequest

• int32 index = 1

• int32 size = 2

• repeated PuzzleTag includes = 3

• repeated PuzzleTag excludes = 4

• string game_id = 1

• int32 turn_number = 2

• optional GameEvent answer = 3

• repeated PuzzleTag tags = 4

• int32 bucket_index = 5

Used in: puzzle_service.PuzzleGenerationJobRequest

• repeated PuzzleBucket buckets = 1

Used in: PuzzleBucket, PuzzleCreationResponse

• EQUITY = 0

• BINGO = 1

• ONLY_BINGO = 2

• BLANK_BINGO = 3

• NON_BINGO = 4

• POWER_TILE = 5

• BINGO_NINE_OR_ABOVE = 6

• CEL_ONLY = 7

SimmedPlayInfo holds per-play statistics from Monte Carlo simulation

Used in: TurnAnalysis

• string move_description = 1

• int32 score = 2

• string leave = 3

• bool is_bingo = 4

• double win_prob = 5

• double win_prob_stderr = 6

• double equity = 7

• double equity_stderr = 8

• repeated PlyStats ply_stats = 9

• bool is_played_move = 10

• int32 iterations = 11

• bool is_ignored = 12

Used in: Evaluation

• double equity_loss = 1

• double win_pct_loss = 2

• bool missed_bingo = 3

• bool possible_star_play = 4

• bool missed_star_play = 5

• bool top_is_bingo = 6

TurnAnalysis contains analysis for a single turn

Used in: GameAnalysisResult

• int32 turn_number = 1

  Turn context

• int32 player_index = 2

• string player_name = 3

• string rack = 4

• GamePhase phase = 5

• int32 tiles_in_bag = 6

• string played_move = 7

  Moves

  e.g., "8D QUIXOTIC" or "(exch AEI)"

• int32 played_score = 8

• string optimal_move = 9

• int32 optimal_score = 10

• double win_prob_loss = 11

  Loss metrics

  For sim/PEG phases (0-1)

• int32 spread_loss = 12

  For endgame phase

• bool was_optimal = 13

  Mistake categorization

• MistakeSize mistake_size = 14

• bool blown_endgame = 15

• bool is_phony = 16

  Phony tracking

• bool phony_challenged = 17

• bool missed_challenge = 18

• bool optimal_is_bingo = 19

  Bingo tracking

• bool played_is_bingo = 20

• bool missed_bingo = 21

• repeated SimmedPlayInfo top_sim_plays = 22

  Enriched data (v2+)

• repeated PEGPlayInfo top_peg_plays = 23

• optional EndgameVariation principal_variation = 24

• repeated EndgameVariation other_variations = 25

• string known_opp_rack = 27

  Known opponent rack (from challenged phony)

  e.g., "JAM" - tiles revealed by challenged phony