home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!psinntp!cadkey!eric
- From: eric@cadkey.cadkey.com (Eric Holtman)
- Subject: Game theory / Probability Question
- Message-ID: <1993Jan12.161250.11378@cadkey.com>
- Sender: eric@cadkey.com (Eric Holtman)
- Organization: cadkey
- Date: Tue, 12 Jan 1993 16:12:50 GMT
- Lines: 49
-
-
-
- O.K., here's a probability/game theory question...... being mostly
- ignorant of such things, I'll toss this out to the net.
-
- There's a game with N players. Each player antes one unit into the kitty.
- Each player is randomly dealt a card from a deck with the numbers 1 to 1000
- on it. Then, each player decides whether to stay or drop out. If you
- stay and have the highest number of those who stayed, you win the kitty.
- If you stay and are beaten by someone else who stayed, you match the kitty.
- If you drop out, there is no penalty. Everyone decides to stay or drop
- or simultaneously, so you don;t know how many stayed until after you make
- your choice.
-
- You never put more than 20 units into the kitty, even if you're matching
- a higher amount. The kitty grows (obviously) when 3 or more players stay,
- since one rakes in the winnings, and two others match (up to 20 units each).
-
- To complicate matters (ignore this rule if you wish), if *no player* stays,
- the player with the highest card matches the kitty (subject to 20 unit
- limit).
-
- My (uneducated) analysis:
-
- The pot, when below 20 units is always offering you even money, i.e.
- you risk 4 units to win 4 units. So, how high does your number have to
- be to justify staying in?
-
- In the cases where the pot has grown beyond 20 units, your getting
- odds on, i.e. risking only 20 to perhaps win 50 units. Does this mean the
- number you draw might not need to be so high to stay? Again, how high should
- your number be to justify staying in??
-
- Also, obviously, the number of players should effect your choice (i.e.
- a number such as 700 is very good when N=2, not so great when N=100).
-
- Why do I care?
-
- Well, the game is actually played with a real deck of cards, but I'm just
- toying with finding the best strategy for winning. I assume I can rank the
- different hands to get to 1 to 1000 (or some similar spread). Just curious,
- and all.
-
- Any probability gurus out there have some insights??
- --
- Eric J. Holtman |
- Product Development | "You fat, bloated, Eeediot!!!"
- Cadkey, Inc. |
- | - Ren Hoek ("that's Ho-ak, not HOKE!!")
-