h a l f b a k e r yThe Out-of-Focus Group.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Who hasn't played that wonderful little children's game of public execution? Guess all the letters in the word with fewer than six errors, and the stick man lives, because he doesn't have all his limbs yet. The guesser can be sure that the chooser isn't cheating, because the game is tricky enough when
played honestly. A computer, however, has the advantage over a human of split-second calculations. As a gag, a programmer could design a computer hangman game in which the computer cheats. Each time the player picked a letter, he would be told that his guess was wrong. When he had exhausted his guesses, the computer would search through its library of words an find one that could be formed with the remaining letters. The program might allow a few "correct" guesses to add to the illusion, or to ensure that a word could be formed with the remaining letters. The player might try to win for a long time before he discovered the trick.
exerpt from "The End of Science", by John Horgan
http://suif.stanfor...fop/WWW/wheeler.txt
"negative 20 questions" as an analogy for how a wave function collapses [xaviergisz, Apr 11 2005]
Impossible Hangman
https://repl.it/@RD...ble-Hangman#main.py
I made an Impossible Hangman [RDude654, Aug 29 2020]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Destination URL.
E.g., https://www.coffee.com/
Description (displayed with the short name and URL.)
|
| |
Might be kind of tricky. If a user guesses each vowel, the computer would probably be stuck with a single word eventually.
How about one that stabs you in the throat with a rusty hook before you can guess it? |
|
| |
Related - Years ago, I used a resource editor to swap some of the windows minesweeper game's digit bitmaps around, and "distributed" it around the office. |
|
| |
You are only really matching your computer's word search capabilities to the human player's, whether you choose the word before or after the letters are guessed. Just start with really obscure words. |
|
| |
Otherwise, the player will quickly expose you by picking the vowels and Y. (Always my first choices.) |
|
| |
But I sincerely doubt you will get people to play for long - they play for satisfaction, and if they're not given any, they will walk, whether or not the game is rigged. |
|
| |
*Australopithecus* has yet to result in a stay of execution. |
|
| |
If the player picked all the vowels, the computer would mark some of those guesses as correct and place the vowels on the blanks in such a way that the word field would match as many of the words in the computer's database as possible. |
|
| |
2 fries: "au***a*o*i**e*u*" - I dunno, I think I'd stand a chance. Especially after guessing S and T, the next two obvious choices.
gas: Still think it ain't gonna work. Try it and see. The point that you're missing is that from the point of view of the player, it makes no difference if the word is not chosen until the last moment. |
|
| |
If this game was written to always leave the most possibilities open, based on the player's letter selections and its lexicon, it would be deterministic and hence very easy to beat. Intuitively though, it should be possible to develop a strategy that increases the average number of guesses required to pin down a word. For 3 letter words, if it's not possible to guarantee a loss for the player, it must be close. |
|
| |
Bread from me for the principle of the thing. |
|
| |
[goldilox]: What'd you do that for? More importantly, hwo did you do that? |
|
| |
r-h-y-t-h-m-i-c / s-p-h-y-n-x - get's 'em everytime. |
|
| |
You'd do it to try to make the player lose, in the spirit of the original idea. At each step, after the player had made a guess, the computer would use or not use the letter, in such a way that the longest minimum path (in terms of all possible future guesses) to any word in its lexicon is chosen. This is quite expensive computationally! The computer has more degrees of freedom than in a "normal" game, where the word is pre-selected, so the average number of steps to nail a word should be longer. Assuming that the player knew this strategy and the lexicon, the whole game could be pre-computed, and the outcome known in advance, a bit like tic-tac-toe. Once the word length is specified, there will be three classes of game - a definite winning strategy for the player, a definite winning strategy for the computer, and cases where it comes down to a guess in the last letter / limb. For those games where the player has a definite winning strategy, he simply selects those letters which "force" the computer along its optimal strategy path. It is not immediately obvious to me what class most word lengths would fall into. |
|
| |
Actually, I was talking about the slightly off topic part about minesweeper. |
|
| |
[finrod] At that time, Borland had a resource editor that could open an executable file and edit its resources. I swapped some of of the bitmaps around. I can't remember which, but say it was the '4' and the '5'. When playing, the computer would display the '4' when there were in fact 5 surrounding mines, and vice-versa, making the game even more unplayable than it already was. |
|
| |
Oh. That must have made some people mad. |
|
| |
Make more impossible by using dictionaries from exotic countries. |
|
| |
This is definitely possible and gets my vote. The computer doesn't have to pick a word right away, as DrCurry points out, but it does make a difference:
Like with FarmerJohn's in-between letters, if the player is left with one guess on B E E _, the computer can ensure that the player loses every time. From the program(mer)'s perspective, however, it's probably best not to use words with two of the same letter very often. |
|
| |
Hurray for evil unbeatable games! |
|
| |
Depending upon the number of guesses allowed, and the required 'length' and sensibility of the answers, this may be trivial. |
|
| |
I'll be nice and say I'm not going to use acronyms or proper nouns--just words allowed in the SCRABBLE(R) brand crossword game. You've guessed "A" and "T" and gotten "-AT". You have fifteen more guesses (rather more than usually allowed in Hangman). Good luck. |
|
| |
[zen_tom]: //r-h-y-t-h-m-i-c / s-p-h-y-n-x - get's 'em everytime// |
|
| |
You weren't a fan of Tomb Raider, were you? |
|
| |
Nope, but I did used to play hangman with a deviant. |
|
| |
I once killed the guy with "Frog". |
|
| |
You can do a similar thing with '20
questions' played by a group (one
'guesser' and a 'panel'). Unknown to the
guesser, the panel has not, in fact,
agreed on the answer in advance. Each
panelist in rotation answers yes or no to
each question from the player, and the
final answer evolves as the game
progresses. It's more of a challenge for
the panel than for the guesser, since
they each have to keep their answers in
accord with those of all the previous
panel members. (This isn't my idea, but
I can't remember where I read/heard it.) |
|
| |
s-e-i was a tricky one, but then she fooled me with a-r-t-s. Tried to snow her with t-h-a-g-o-m-i-z-e-r (after the late Thag Simmons, thanks to Gary Larson), but she got that one. |
|
| |
I like the idea of it but, i dont think its not gonna work. I gave it a neg 4 that . Soz |
|
| |
I made the impossible hangman, I didnt see this before i
just searched the web. The game is in links |
|
| |