J4, tell me if this satisfies the conditions.
There are approx. 1 million English words. The experiment is to see if you can guess a randomly selected word from a randomly selected book.
To encode the answer, I type the ISBN number of the book in a text editor or word processor, e.g.
9780140289206 (13 digits)
I complete the code by adding the page number (3 digits), line number (2 digits) and word number (2 digits).
Input mask: BBBBBBBBBBBBBPPPLLWW (20 digits)
Example: 97801402892060961001
(ISBN for
Godel, Escher, Bach: an Eternal Golden Braid, page 096, line 10, word 01)
Next, I take the word ("football"), and convert it to binary with an online converter, e.g. http://www.unit-conversion.info/texttools/convert-text-to-binary/
Result: 01100110 01101111 01101111 01110100 01100010 01100001 01101100 01101100
I paste it into the text editor mentioned, then remove all the spaces:
0110011001101111011011110111010001100010011000010110110001101100
Next, I connect the two numbers by adding "base 10" behind the book code, "base 2" behind the binary word code and "*" between them:
97801402892060961001 base 10 * 0110011001101111011011110111010001100010011000010110110001101100 base 2
I copy and paste the combination to Wolfram Alpha: http://www.wolframalpha.com/
Wolfram Alpha multiplies the numbers and returns:
721895711186429677656686373548203609676
Next, I publish the number in our guessing game. If a person guesses the word correctly, it is easy to run the process in reverse. Simply divide the given number by the binary of the guessed word. If it is the correct word, the following will happen:
1. The result will be a 20 digit number
2. It will follow the pattern BBBBBBBBBBBBBPPPLLWW
3. If you paste the first 13 digits into google, it will show you a book
4. If you can get the book, or search it online, you will see the word on page PPP, in line LL, word number WW.
If the person has guessed the word incorrectly, the probability of finding that wrong word in a wrong book at that wrong place would surely be even smaller than guessing 1 out of a million words.
It may be hard to find that specific book, but at least it would be verifiable. I'm sure it would be pretty easy to write a script to automate the process, but I wouldn't know where to start.
Thoughts?
Edit: Hi X, agree with everything you said. Much info = many syncs