Eleusis

ec2-3-15-15-56.us-east-2.compute.amazonaws.com | ToothyWiki | RecentChanges | Login | Webcomic

A game played not by logical deduction (figure things out from the rules and the data) but by induction (figure out the rules themselves, from the data).

There are [several] [webpages] [about] the game. The first now [also] includes rules on Eleusis Express - essentially the same game, but after a correct play you may guess the rule, ending the round if correct.

Briefly: One player takes the role of "Nature" or "God", and writes a secret rule on a piece of paper, which they then hide. The other players are then dealt hands of cards from a normal playing card deck, and take turns attempting to play a card to the end of a sequence. To each play, "God" says "Yes" or "No", and it's added to the sequence if "yes" or placed in some visible sideline if "no", so that all previous successes and failures are visible.

The aim of the game for the players is to discover the secret rule, and for "God" is to make the rule reasonably guessable - not too easy or too hard.

Sample rules are:

If the last card was red, play a black card, otherwise play a red card.
Play a card whose number is one higher than the number of the last card played.
If the last card is even, play a black card; if it's odd, play a red card.

AlexChurchill used to play it several years ago, and found it fascinating. He would be interested in playing it again.

: ChrisHowlett too - either at GE or on the Wiki. But perhaps not until after finals on the Wiki.

The thought occurs that it would be quite possible to play a game on the wiki, using ToothyWikiInternals/SignatureServer to encode the secret rule...

: Indeed. Although it would need to be modified somehow - so as to give people a selection of cards to choose from. suggest either: Publically available lists of what everyone has to play, or: Anyone may play anything which has not yet been played.

: (PeterTaylor) You mean you need a card-shuffler which deals a deck of cards between n people for reasonable values of n? If you want card-shuffling code (in Java, not Perl, but shouldn't be too hard to port - Perl does have arrays, doesn't it?) I can provide.

: Hmm. I reckon the hand of cards isn't a very important aspect of the game anyway. Keeping a public "score" = handsize for each player would do, I think, and allow each player to play any card from the deck. --AlexChurchill

: From the rules as given, the handsize of each player doesn't seem overly relevant either? - Kazuhiko

: Well, there is a concept of scoring, based on the different players' handsizes. It's spelled out on the more detailed rules pages linked to above. (They also have a concept of a "Prophet", which I've not come across in the ones I've played before: basically a person who thinks they know the rule starts "predicting" God's answers, and can potentially score quite well from this, as long as they don't get one wrong!) It's kinda necessary, altho the play shouldn't focus too much on the scoring IMHO. --AlexChurchill

: I've seen Hofstadter's version - which does indeed have that mechanic. -- TheInquisitor

This is a NoteToSelf? to re-suggest starting a game, but probably only once I'm likely to have long-term reliable net access. I like the prophet mechanic, and I also think that rather than allowing any play, we should have 8 or so cards anyone can play. Or claim that none of the 8 are playable. --CH

CH rediscovers this page. Anyone interested?

: Isn't this just the chairman's game, but less amusingly presented? --Vitenka
: Focused on the interesting induction bit, rather than the political metaphor and randomly slapping people bit, yes. --AC

DR asks: do you think it improves or detracts from the game, if you limit the scope of possible rules to being specific properties of the cards: order played, number and colour. eg orientation they land on, whether facing east, number of second passing since last play, if the player spoke or not, etc are not valid data upon which a rule may act.

Personally, I think it'd be best to follow the same rule as Zendo (which is basically the same game played with IceHouse? pieces): the table should be treated as an infinite plane in a bubble universe, and the validity of a rule must be deducible from only the cards without reference to any other information. But IceHouse? pieces have a few more variables than playing cards (physical configuration relative to each other), so I'm not sure if that leaves a good enough Eleusis rule space; I suspect it does, and think it'd certainly be best to start under such restrictions. --AC

Yes - the "You can play a card if you are wearing green" rule is bad and wrong. "Card must be played in a non-wonky way" is slightly less bad. I'd stick with just "This is the card played, and this is the previous sequence of cards played" as being good enough. --Vitenka (Though, as long as people know in advance that positioning etc. can count, it may be ok, though significantly less wikiable)

: In which case, how would you feel about a BNF rule generator ? --DR
: Would seem to defeat most of the fun of being the judge, but I don't see any conceptual problem with it. --AC

CategoryGames