|
There is an Elise console application project ("eliseconsole") in the source code distribution. This version of the engine is driven by text commands. Here are the most useful commands in the console application and what they do; these explanations are fairly brief, please consult the source code for details:
Used to change the current game type. Select the desired game type from the menu.
Prints current game state information (such as the current board, the scores, and the number of tiles left in the bag.)
Generates possible moves using the hi-leave algorithm. Will perform end game search if the bag is empty.
Exits the program.
Prints the current game board.
Makes a move (will generate the possible moves, then you select the desired move from the menu.)
Makes a move -- this can include illegal or phony moves. You must type in the word, the square and the direction, with tiles represented by blanks in lowercase.
Allows you to set the rack for the current player.
Passes the turn.
Has Elise play many games against itself, using move generation and evaluation algorithms that you specify.
Empty both players' racks.
Starts a new game of the same type as the current game.
Check whether a specified word is in the current lexicon.
Plays 100 quick games and reports the amount of time taken.
Toggles whether to generate down moves as well as across moves on the first turn.
Prints the current contents of the well (the tiles in the bag plus tiles on opponents' racks, if they are hidden in the current game type).
Change the current lexicon.
Lookup a specified word in the DEFINITIONS file for the current language.
Allows you to manually set the entire board.
Set the players' scores.
Set the current player's turn.
Toggle whether fastsim sorts moves by win probability or point spread.
Merge two definitions files together.
Generate rack statistics for the current game type and lexicon.
Works as findmoves, except it uses experimental strategy.
Saves the current position as an .elise file.
Loads a position from an .elise file.
Used to generate word-play frequency data for the current lexicon.
Output words in the current lexicon with no definitions in the current definitions file.
Toggle on median rack draws.
Fill the current rack from the bag.
Show info about the current rack and its possible rack leaves.
Toggle opponent rack inferences in simulation.
Run the opponent rack inference code and print the results.
Set tiles that are known to be on opponent's rack (this overrules any tile inferences.)
Set the clock for the current player.
Make a "clock move" -- a move is chosen with simulation settings according to the amount of time left on the current player's clock.
Performs a simulation of a specified ply level.
Generates a win probability table from a table of game outcomes (generated by simwinprob).
Play many games and generate a table of game outcomes; this table can be converted into a win probability table using outputprobtable.
From the full-rack simulation data for the current game, generate a table of partial rack values.
Toggle the use of end game heuristics (in simulation) on or off.
Calculate a win probability value based on the specified game state.
Rate the current rack draw against all possible rack draws.
Toggles on or off the greedier end game search code.
Outputs a blank trie (tribble) for the current lexicon.
Runs timetrial 20 times.
Gives Elise's static end game evalution of the current position (if the bag is empty). Static end game evaluation generates all possible moves for both players and attempts to find a winning end game using only those moves -- that is, it does not re-generate moves during the search. This gives the same result (win/loss/draw) as an exhaustive search between 90% and 95% of the time, but is much faster than an exhaustive search.
Auto-play a game until the bag is emptied (we reach the endgame). Useful for testing endgame changes.
Load a Scrabble® game saved as a .GGG file.
Set Elise's current "mood", where -3 is extremely pessimistic, 0 is serene and levelheaded, and 3 is extremely optimistic. Moods bias the set of opponent racks used in simulation.
Plays many games and outputs tile leaves after "small" moves. This data can be used to create a tile leave statistics (.tl) file for use with the tile inference code.
Compiles a tile leave statistics (.tl) file from the tile leave simulation results.
Plays many games and outputs "fun" games in .GCG format. "Fun" games include come-from-behind games and games with long (10+ letter) moves.
Outputs the "rare racks" in the currently loaded rack statistics -- all racks that make below a specified threshold number of appearances. Rare racks, if present, will be used in generating rack statistics.
Toggles rack sorting on or off during the pre-end game.
Another method of making a play: type the word you want to play, and Elise prints the possible locations for the word; select the desired move from the menu. |