On 6/24/24 10:39 AM, juan via agora-discussion wrote:

HIGHEST NUMBER GAME

Some thoughts...

1. To avoid turning this into a contest of who can type the most nines into an
email or abuse Knuth notation the most, maybe make the winner the player with
the median number (higher of the two middlemost if an even number of players).
Or the *second* highest number, and if only one valid player remains, e
automatically wins the round.

2. Putting commit before bet or combining the two will speed up the game a
little. A well chosen fingerprint algorithm shouldn't leak any meaningful
information.

3. What if each time someone bets or raises e also picks and answers a Boolean question about eir number. To call or raise, one must first answer all
the previous questions. Agorans being Agorans, we'd probably need to put some
reasonableness bounds on it -- e.g., provided someone's secret number is X or less, and given a math background of Y, one must be able to determine the answer by hand without unreasonable effort, for suitable X and Y. (Someone can pick a number above X at eir own risk.) Questions that can be answered by checking against a table of known results online would also qualify.

4. Explicitly disallow any conditionality, randomness, or external references in the secret number, even if clearly described.


Overall, I like this: a simple, elegant concept. Hopefully I'm not ruining
that too much!

--
Mischief
(steampunk hat)
(Invulnerable, 1 Bang)

Reply via email to