Talk:Mental poker

From Wikipedia, the free encyclopedia

Contents

[edit] Collusion

Hi - is it possible to design a system where, to help prevent cheating, some centralised agency (e.g. the Poker Room) gets sent all the encryption keys after the hand is dealt, so that it can investigate any possible collusion? Would this be possible? Evercat 11:46, 27 August 2005 (UTC)

Doesn't that defeat the point? If there is a trusted centralised agency, couldn't it do the shuffling and dealing of cards. 157.161.173.24

[edit] Where's Shafi?

The whole concept of "mental poker" was invented by Goldwasser and Micali. I'm surprised she isn't credited in the text. I'd also expect to see at least some mention of Oded Goldreich (Weizmann Kavod).

[edit] Commutative encryption?

Which encryption schemes are commutative, besides RSA with identical moduli? Because I heard there was an attack against using RSA to encrypt two messages with two different moduli. --Zemylat 16:04, 31 May 2006 (UTC)

2 possible cards (heck why not 52) packaged as 1 card, when 'A' key is sent it unencrypts the card of choice from the host who initialy encrypted... Same thought occurred to me, anyone have any references on such code?  —Preceding unsigned comment added by 96.2.147.222 (talk) 06:01, 23 December 2007 (UTC)

[edit] Question

Doesn't this algorithm have the weakness that Bob has no way of knowing if Alice replaced the deck with one that contains duplicate cards, say to increase the odds of picture cards etc. There would need to be a step at the end that decrypted all of the cards to ensure that no alterations had been made to the freqencies of the cards in the deck by the first encryptor. Another possible problem would be for the final shuffler to replace the cards that would be dealt to him with plain text cards, since he has the right to see them from the beginning, he can request the keys to decrypt them, but throw them away. I suppose again this would be solved by insisting that every card in the game is decrypted at the end so that everyone can check that the deck wasn't modified. This way, at least you'd be sure of knowing if someone cheated.Kybernetikos 09:37, 25 October 2006 (UTC)

Duplicate plain-text cards would result in duplicate encypted cards. Easilly detectable.
Replacing cards would result in Bob getting big errors when trying to decrypt his dealt cards. Also I think I remember that there are ways of checking cards with zero-knowledge proofs... but of that I'm not certain... crypto experts will have to answer that one. --J-Star 10:30, 25 October 2006 (UTC)