finding the statistically best Hangman word

Of the Tabletop, and other, lesser varieties.

Moderators: SecondTalon, Moderators General, Prelates

btilly
Posts: 1877
Joined: Tue Nov 06, 2007 7:08 pm UTC

Re: finding the statistically best Hangman word

Postby btilly » Wed May 27, 2009 10:19 pm UTC

Lagren wrote:I had a flash of inspiration, based on "wax."
How about "Wan"?

Once people get the "an", they're likely to waste guesses on the very large list of "An" words.
Ban, Can, Fan, LAN, Man, Nan, Pan, Ran, Tan, Van, Wan.
As they run down this list, when they come to "Wan" it won't even sound like anything, because all the other words are pronounced with hard A's while "Wan" is a long A.

I think you're mixing up "wan" (meaning to look pale) and "wane" (the opposite of wax). Because I guarantee that "wan" is pronounced with a short A just like ban, can, fan, etc.

There ought to be a special prize for misspelling a word in hangman and accidentally winding up at a real word.
Some of us exist to find out what can and can't be done.

Others exist to hold the beer.

User avatar
bigglesworth
I feel like Biggles should have a title
Posts: 7461
Joined: Sat Apr 07, 2007 9:29 pm UTC
Location: Airstrip One

Re: finding the statistically best Hangman word

Postby bigglesworth » Wed May 27, 2009 10:25 pm UTC

Where I am at least, wan is pronounced "waan", can is pronounced "can" and wane is pronounced "wayne"
Generation Y. I don't remember the First Gulf War, but do remember floppy disks.

User avatar
Adacore
Posts: 2755
Joined: Fri Feb 20, 2009 12:35 pm UTC
Location: 한국 창원

Re: finding the statistically best Hangman word

Postby Adacore » Thu May 28, 2009 2:18 pm UTC

Words with X and Z in them tend to fare very well indeed, since nobody guesses them early. As already mentioned, the best word I've found is Lynx - no vowels, and neither of the two most commonly guessed (in my experience) constonants 's' and 't'. While words such as wax and wan are great, we normally play with a minimum word length of 4 characters.

Of course, if you know the people you're playing with you get an element (to use the Mafia forum terminology) of WIFOM. They know you're using obscure words, so guess obscure letters far more frequently. This, of course, changes the 'optimal' word choice.

User avatar
Lissa
Posts: 34
Joined: Mon Apr 13, 2009 6:51 pm UTC
Location: Oslo, Norway
Contact:

Re: finding the statistically best Hangman word

Postby Lissa » Thu May 28, 2009 3:04 pm UTC

bigglesworth wrote:a, I, O.

That's all I think.


A friend of mine used π (pi) as a single letter hangman word :P

Zalzidrax
Posts: 94
Joined: Sat Sep 13, 2008 3:41 am UTC

Re: finding the statistically best Hangman word

Postby Zalzidrax » Thu May 28, 2009 3:56 pm UTC

Discounting cheap shots (4 letter words or less that could be any number of things, even if you have all but one letter), I've had good success with the word quetzal.

And y's can be bad because many of the "hard to guess" words use them.

phider2
Posts: 71
Joined: Sun May 24, 2009 10:52 pm UTC

Re: finding the statistically best Hangman word

Postby phider2 » Thu May 28, 2009 8:18 pm UTC

Xaddak wrote:I saw a few people using phrases, so how about "the quick brown fox jumped over the lazy dog"? I think it would only work with crowds that aren't too geeky, though, because Word used to use that phrase to show fonts, if you went into the Font window itself. It still might, actually, I'm not sure.

It's funny, because I tried that at school today. It was after finals, so there wasn't really anyone at school (only people that couldn't get a ride home or were on the debt list), and I was playing hangman with a few teachers and another student. Sort of as a joke, I did "the quick brown fox jumped over the lazy dogs," saying it was really easy (having all the letters in the alphabet, it's impossible to lose :P). Surprisingly, nobody got it until it was down to the last few letters.

User avatar
SecondTalon
SexyTalon
Posts: 26528
Joined: Sat May 05, 2007 2:10 pm UTC
Location: Louisville, Kentucky, USA, Mars. HA!
Contact:

Re: finding the statistically best Hangman word

Postby SecondTalon » Thu May 28, 2009 8:31 pm UTC

The quick brown box jumps over the lazy god? That doesn't make any sense!
heuristically_alone wrote:I want to write a DnD campaign and play it by myself and DM it myself.
heuristically_alone wrote:I have been informed that this is called writing a book.

User avatar
Shivahn
Posts: 2200
Joined: Tue Jan 06, 2009 6:17 am UTC

Re: finding the statistically best Hangman word

Postby Shivahn » Fri May 29, 2009 7:34 am UTC

SecondTalon wrote:The quick brown box jumps over the lazy god? That doesn't make any sense!


Maybe it's one of those sleeping gods?

Or a Shinto god. Or something.

Anyway, I've found syzygy to be a most excellent word for hangman. It's so excellently obscure that Firefox is telling me that I misspelled it.

It's probably one of the best hangman words. It's rare, and if you give people five guesses (I think that's the usual amount given), they'll probably spend time on vowels like a,e,i,o and u. And if they guess y, they'll just be looking at -y-y-y, which is confusing. Also, though s is probably decently common, z and g are both somewhat obscure. You only get to use it once though, because every time you play hangman after that, if your word has six letters the guessers will guess "y" first.

I have succeeded in altering my entire family's hangman behavior.

User avatar
clockworkmonk
I'm on a horse!
Posts: 649
Joined: Fri Aug 03, 2007 12:53 am UTC
Location: Austin

Re: finding the statistically best Hangman word

Postby clockworkmonk » Fri May 29, 2009 1:27 pm UTC

one thing that might work are any of the words that have a Q but no U, such as Qat, and Qabala. people tend to guess vowels, and when they get all the A's, they generally won't guess the Q.
418 I'm a teapot

User avatar
Armadillo Al
Posts: 934
Joined: Thu Aug 14, 2008 4:17 pm UTC
Location: Illinois

Re: finding the statistically best Hangman word

Postby Armadillo Al » Fri May 29, 2009 4:34 pm UTC

If only I could ever use anything like that...when I play, it's usually with people who would prefer if you used words that people have actually used in a sentence at some point in the history of the English language.

One of my favorite YouTube clips is a round of Wheel of Fortune where everyone pretty much wanted to hang themselves trying to get the puzzle "OXIDIZED". When you can't buy vowels, that one is *mean*.
Compiling.. wrote:xkcd... where EVERYONE loves EVERYONE...

User avatar
Blatm
Posts: 638
Joined: Mon Jun 04, 2007 1:43 am UTC

Re: finding the statistically best Hangman word

Postby Blatm » Sat May 30, 2009 4:10 am UTC

AvalonXQ wrote:How many single letters qualify as proper hangman words?


Define "proper". Every letter has its own entry in every dictionary I've ever seen.


I support "lynx" and "qat", though "qabala" isn't as good because if they know the word, they'll probably guess it after seeing "_ a _ a _ a", and similarly for "syzygy". More words with q can be found at wikipedia. Most of them are not really English though. At the bottom of the article, they say:

Wikipedia wrote:In Scrabble in North America, the only acceptable words with a Q and not a U are qi, qat, qaid, qadi, qoph, qanat, tranq, faqir, sheqel, qabala, qabalah, qindar, qintar, qindarka, mbaqanga, and qwerty, along with their plurals (such as qats and sheqalim).[ TWL ] Other words in this list, such as suq and qiviut, are also acceptable, but since these contain a U, they are less likely to be useful.


Amusingly enough, Firefox recognizes only "qi" and "qwerty" from that list. Does anyone know of a good Firefox dictionary? (Ideally using British spelling)

User avatar
Internetmeme
Posts: 1405
Joined: Fri Jul 25, 2008 3:16 pm UTC
Location: South Carolina, USA

Re: finding the statistically best Hangman word

Postby Internetmeme » Mon Jun 01, 2009 8:42 pm UTC

Xyzzy

The three least-guessed letters in one package.

Also, why not use words or phrases without a "u" after "q", such as qitar, or Iraq. That is a good way to throw people off.

If all else fails, obscure mesoamerican gods also work.

Itzhama.
Spoiler:

User avatar
Uber_Apple
Posts: 134
Joined: Wed Aug 20, 2008 10:07 am UTC

Re: finding the statistically best Hangman word

Postby Uber_Apple » Wed Jun 03, 2009 4:48 pm UTC

the best word is *drum roll*


hangman


seriously it took me and another person about 10 minutes before we gave up.

User avatar
Internetmeme
Posts: 1405
Joined: Fri Jul 25, 2008 3:16 pm UTC
Location: South Carolina, USA

Re: finding the statistically best Hangman word

Postby Internetmeme » Wed Jun 03, 2009 5:53 pm UTC

Fun hangman game today:
Told my friend that it was sudden death. If he misguessed one letter wrong he lost.

The phrase:
The quick brown fox jumps over the lazy dog.
Spoiler:

User avatar
Tommy2995
Posts: 27
Joined: Fri Nov 07, 2008 8:00 am UTC
Location: Pallet Town.

Re: finding the statistically best Hangman word

Postby Tommy2995 » Fri Aug 07, 2009 8:26 pm UTC

I find good words are Ketchup, Box and August.
I am Dave! Yognaught and I have the balls!

User avatar
lu6cifer
Posts: 230
Joined: Fri Mar 20, 2009 4:03 am UTC
Location: That state with the all-important stone

Re: finding the statistically best Hangman word

Postby lu6cifer » Sat Aug 08, 2009 4:15 am UTC

I've used "Brobdingnagian" before.

Sometimes the best strategy is to use a really obscure word.

Though that can also backfire. I've sometimes guessed the word "syzygy" prior to any letters in hangman games because nobody thinks anyone will guess 'y' or 'z', hence, it's used very often.
lu6cifer wrote:"Derive" in place of "differentiate" is even worse.

doogly wrote:I'm partial to "throw some d's on that bitch."

User avatar
Vieto
Posts: 1558
Joined: Sun Jun 22, 2008 10:44 pm UTC
Location: Canada

Re: finding the statistically best Hangman word

Postby Vieto » Tue Aug 11, 2009 9:05 pm UTC

bigglesworth wrote:But a single letter is a bastardly move in any hangman game.

This idea is very interesting.


Not really. I think the only proper single letter words are A and I, with Z and S being sometimes words (catch some z's, 's = is), unless you are using abbreviations, in which case any random letter or string of characters works, or if you are using the letters in certain contexts (in math, replacing unknowns with X or G, for instance) and slang.

Of course, for words that only use one letter (A, I, X..., Zzzzzz, etc.), you have more options. _ _ _ _ _ _ _ _ could easily be Zzzzzzzzz....

sithwalrus
Posts: 13
Joined: Thu Feb 26, 2009 8:12 pm UTC

Re: finding the statistically best Hangman word

Postby sithwalrus » Thu Oct 08, 2009 4:22 am UTC

Yay!!! Complicated Problems which distract me from homework!!!!
in a fair game of hangman, only the 26 letters are generally used, you sometimes can get away with something like n00b, or l33t, or ?, but after you do it once they will probably tell you to stop. using every character in the English language is for people who can't think of any good ones normally, also since there are pronouns for everything, and new ones are made constantly, they are cheap. you only have truly whooped your opponent if it is a word they actually could have known.
with this junk in mind, the important things to consider in a hangman word are as follows;
1) Word Length
a) a good word is short. if there are more than 10 letters, the probability of any given letter being in it increases. Also, the number of words it could be falls considerably, and there will probably be enough repeated letters that someone can find the complete word after only a couple guesses.
b)a good word is not too short. there are three one letter words, so someone is guaranteed to never make more than two wrong guesses. 3 is similar, and you only start to see a really wide range of possibility's once you get to around 4 letters.
- IMAO the best #of letters is inbetween 5 and 8. on either side of that range you start seeing vocabulary limitation, which can help your opponent.
2) Letter Distribution
oh, the mother of all hangman words. to use an optimal letter distribution, you need to know that there are basically four kinds of players, each with their own weaknesses. a n00b who likes vowels (or even picks randomly, heaven forbid!), a Enlightened n00b who likes e,t,a,o,i,n, and s, a skeptical n00b who will quickly switch over to memorable uncommon letters x, j, q, and z after they miss a few, and the pro, who can reverse engineer your strategy and counter it based on the same tedious psychological manipulation which you are using.
3) Similarity
after someone guesses their first couple letters they may start to see patterns in the words that it could still form. some words are easy to finish after guessing less than half of the letters, since there are only a few combination's of letters that could possibly finish the word, but others leave dozens of options for how you could finish them. for this reason, words like "Quiz" and "Lynx" lose some of their value as words (though they are still awesome for inverse distro), while words like "Hack" or "Soft" actually are more challenging then they look.
4) Obscurity
if nobody is familiar with the word they are trying to guess, then they can't use similarity's to guess the next letter, in fact, banking on something looking like other words and trying to guess letters according to this plays right into a 3 from the hangman. obscure words are pretty much always better than common words, but this is hard to do if you are playing against someone who is more fluent than you, so a bit of research somewhere like http://www.islandnet.com/~egbird/dict/dict.htm can be useful.

with all this said, my favorite hangman words are winx, jerque, qasida, dyvors and fruzz.

User avatar
Cynical Idealist
Posts: 1124
Joined: Mon Sep 15, 2008 10:48 pm UTC

Re: finding the statistically best Hangman word

Postby Cynical Idealist » Thu Oct 08, 2009 8:16 pm UTC

SecondTalon wrote:Twyndyllyng

That's archaic and might not be allowed. I'm fond of syzygy myself.
The internet removes the two biggest aids in detecting sarcasm:
1)The tone of voice
2)the assumption that the other person is sane
Elvish Pillager wrote:See? All the problems in our society are caused by violent video games, like FarmVille.

User avatar
Armadillo Al
Posts: 934
Joined: Thu Aug 14, 2008 4:17 pm UTC
Location: Illinois

Re: finding the statistically best Hangman word

Postby Armadillo Al » Fri Oct 09, 2009 2:03 pm UTC

Twyndyllyng is archaic and syzygy isn't? I must be playing Hangman with the wrong people.
Compiling.. wrote:xkcd... where EVERYONE loves EVERYONE...

User avatar
Cynical Idealist
Posts: 1124
Joined: Mon Sep 15, 2008 10:48 pm UTC

Re: finding the statistically best Hangman word

Postby Cynical Idealist » Fri Oct 09, 2009 10:04 pm UTC

Armadillo Al wrote:Twyndyllyng is archaic and syzygy isn't? I must be playing Hangman with the wrong people.

I'm now going to do some of the worst citing I have ever done in my life:
Regarding twyndyllyng:
Wikipedia: (If archaic words and spellings are considered, there are many more, the longest perhaps being twyndyllyngs, the plural of twyndyllyng.)
Wiktionary: 1. (obsolete) A twinling; twin.
Regarding syzygy:
Wikipedia has an "In popular culture" section on it.
The internet removes the two biggest aids in detecting sarcasm:
1)The tone of voice
2)the assumption that the other person is sane
Elvish Pillager wrote:See? All the problems in our society are caused by violent video games, like FarmVille.

xnull
Posts: 58
Joined: Sun Jan 07, 2007 12:57 am UTC

Re: finding the statistically best Hangman word

Postby xnull » Sat Feb 27, 2010 10:56 pm UTC

I wrote code to do this and then, when Googling to see if someone else had done the same, encountered this thread. I'm resurrecting it.

Code: Select all

#include <iostream>
#include <cstdlib>
#include <string>
#include <cmath>
#include <set>
#include <map>

using namespace std;

int main(int argc, char** argv) {
        if (argc != 2) {
                cerr << "[!] Usage:\t" << argv[0] << " [min_word_length]" << en$
                return 1;
        }

        int minwordlen = atoi(argv[1]);

        string word;

        map<char, double> pdf;
        map<double, string> hangmandict;

        pdf['a'] = 0.08167;
        pdf['b'] = 0.01492;
        pdf['c'] = 0.02782;
        pdf['d'] = 0.04253;
        pdf['e'] = 0.12702;
        pdf['f'] = 0.02228;
        pdf['g'] = 0.02015;
        pdf['h'] = 0.06094;
        pdf['i'] = 0.06966;
        pdf['j'] = 0.00153;
        pdf['k'] = 0.00772;
        pdf['l'] = 0.04025;
        pdf['m'] = 0.02406;
        pdf['n'] = 0.06749;
        pdf['o'] = 0.07507;
        pdf['p'] = 0.01929;
        pdf['q'] = 0.00095;
        pdf['r'] = 0.05987;
        pdf['s'] = 0.06327;
        pdf['t'] = 0.09056;
        pdf['u'] = 0.02758;
        pdf['v'] = 0.00978;
        pdf['w'] = 0.02360;
        pdf['x'] = 0.00150;
        pdf['y'] = 0.01974;
        pdf['z'] = 0.00074;

        while(cin >> word) {
                set<char> charmap;
                set<char>::iterator itr;
                double prob = 1.0;

                for (int i = 0; i < word.length(); i++) {
                        charmap.insert(word[i]);
                }

                for (itr = charmap.begin(); itr != charmap.end(); itr++) {
                        char c = *itr;

                        double p = pdf[c];

                        if (p == 0) {
                                prob = 0.0;
                                break;
                        } else {
                                prob *= (1 - p);
                        }
                }

                prob = pow(prob, 6);

               if (prob > 0 && word.length() >= minwordlen) {
                        hangmandict[prob] = word;
                }
        }

        map<double, string>::reverse_iterator itr = hangmandict.rbegin();

        for (int i = 0; i < 10; i++) {
                cout << itr->first << ": " << itr->second << endl;

                itr++;
        }

        return 0;
}


The code makes the assumption that letter positions are conditionally independent (which is obviously not true). Feel free to modify it and run it for diglyphs and triglyphs.

To run

Code: Select all

# cat /usr/share/dict | ./a.out 4
Where the number corresponds to the minimum length word you will allow.

This code really likes the word muumuu (a kind of boot).
This statement is false.

User avatar
BurningLed
Posts: 561
Joined: Tue Feb 09, 2010 5:42 pm UTC

Re: finding the statistically best Hangman word

Postby BurningLed » Sun Feb 28, 2010 12:49 am UTC

Qi. Q would be the last letter guessed before the 'i,' almost certainly. Or better yet, 'qat' since there is 'hat,' cat,' 'baat,' and a number of other '*at' words that would be guessed as well.
Axman wrote:Some people blow their cash on watches that they show off to people who think said watches make a person cool. Some people spend a weekend buying everyone fake gifts in a game of make-believe.
I think the latter group is awesome.


Return to “Gaming”

Who is online

Users browsing this forum: No registered users and 11 guests