You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
XQV is not fun because there are too few solutions. DES is not fun because there are too many solutions. This is particularly noticeable in Hyperghost when it can take a whole game to exhaust a common set of letters.
The simplest option is to add two random integers: [0,12] + [0, 13] => [0, 25] and use that as the index into the list of letters sorted by frequency.
Another option is to use something similar to the current system, but add two random floats together where we currently use one to walk up the odds table. The odds table would need to be sorted by frequency.
Another option is to work this into a catch-up mechanism. Generate n+2k completely random puzzles, where n is the number of players, and k is an arbitrary integer - maybe 2 or 3. Then scan the word list for matches to each puzzle, and calculate a difficulty rating for each puzzle where more matches and more common words all lower the difficulty rating. Sort the players by their score, but add some noise so they aren't guaranteed to be in exactly the correct order. Throw away the top k and the bottom k puzzles, then assign the puzzles in order of difficulty to the sorted players. The most difficult puzzle will usually go to the leading player.
choose a technique for generating letters
implement it
track the puzzles in a match, and don't repeat them - maybe track the last 100 puzzles
The text was updated successfully, but these errors were encountered:
XQV is not fun because there are too few solutions. DES is not fun because there are too many solutions. This is particularly noticeable in Hyperghost when it can take a whole game to exhaust a common set of letters.
The simplest option is to add two random integers: [0,12] + [0, 13] => [0, 25] and use that as the index into the list of letters sorted by frequency.
Another option is to use something similar to the current system, but add two random floats together where we currently use one to walk up the odds table. The odds table would need to be sorted by frequency.
Another option is to work this into a catch-up mechanism. Generate n+2k completely random puzzles, where n is the number of players, and k is an arbitrary integer - maybe 2 or 3. Then scan the word list for matches to each puzzle, and calculate a difficulty rating for each puzzle where more matches and more common words all lower the difficulty rating. Sort the players by their score, but add some noise so they aren't guaranteed to be in exactly the correct order. Throw away the top k and the bottom k puzzles, then assign the puzzles in order of difficulty to the sorted players. The most difficult puzzle will usually go to the leading player.
The text was updated successfully, but these errors were encountered: