krey@gmdzi.UUCP (Juergen Krey) (03/16/90)
Hi! I enjoy playing the solitairy card game Seahaven Towers, and i want to know whether all initial configurations can be solved. I wrote a little simulator with some heuristics to try some really difficult layouts, but til now all config's could be solved. Another problem concerns the Indexor file in the system folder, which contains the latest state of the recent game - rather strangely coded. I suppose the resource TOWR to have the board-info's, but don't manage to decrypt. Any thoughts ???
wilson@csli.Stanford.EDU (Nathan Wilson) (03/17/90)
krey@gmdzi.UUCP (Juergen Krey) writes: >Hi! >I enjoy playing the solitairy card game Seahaven Towers, and >i want to know whether all initial configurations can be solved. It's easy to come up with an intial configuration that is unsolvable. For example, imagine the first column from top to bottom is 7, A, 2, 3, 4 of spades and the two cards in the spaces up above are the 5 and 6 of spades. The 5 and 6 can never move until the 7 is visible. The 4 and 3 can only be moved to one of the open slots. Then all the spots are full and you're hosed. Note that this problem could hold for any sequence of seven cards (of which there are seven), in any suit(4), in any column(10) and that the cards in slots above could be in either order(2). Hence there are at least 7*4*10*2*(45!) shufflings of the deck that would work. That is 66988443684668908954699370437568320436054890905600000000000 possible shufflings (isn't arbitrary precision math fun!). But given that there are 80658175170943878571660636856403766975289505440883277824000000000000 total possible shufflings, you only have a 1 in 674274182400 (that's 600 billion for those who don't want to count those decimal places) chance of ever encountering this particular problem. Now there are probably other more common no win situations that you can get into, but this was the simplest I could come up with off the top of my head. Nathan Wilson Teleos Research nathan%teleos.com@ai.sri.com PS. I don't think I'm really quite as nerdly as this note sounds. I just had Lisp going and it's a Friday so what the hey.
urlichs@smurf.sub.org (Matthias Urlichs) (03/19/90)
In comp.sys.mac.programmer, article <12713@csli.Stanford.EDU>, wilson@csli.Stanford.EDU (Nathan Wilson) writes: < krey@gmdzi.UUCP (Juergen Krey) writes: < < >Hi! < >I enjoy playing the solitairy card game Seahaven Towers, and < >i want to know whether all initial configurations can be solved. < < It's easy to come up with an intial configuration that is unsolvable. Unsolvables are pretty common. I just (for want of anything better to do) wrote a short MPW Pascal program which did an exhaustive search of a given game to see if it could be solved. Out of 10 games, two were unsolvable. Since Seahaven has no way to get the card values out, or the moves in for that matter, I didn't follow it up with a self-play mode and/or a larger data sample to see what the actual percentage is. Also, this was depth-first, and although I did put in some ideas on what moves moves make more sense to try first, the resulting games were _awful_. However, it sure is nice to know if a given game can be solved. Keeps you from slaving away on one for days if the machine can tell you right away "no go". ;-) I probably still have it somewhere. On tape. Unfortuantely the tape drive was eaten by its power supply two weeks ago. Anyone have a 150-MB Archive or TEAC they want to part with? -- Matthias Urlichs