home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!elroy.jpl.nasa.gov!nntp-server.caltech.edu!cxxh
- From: cxxh@cco.caltech.edu (Chris H. Hall)
- Subject: Game theory/Optimization question
- Message-ID: <1992Jul29.154337.20895@cco.caltech.edu>
- Keywords: game
- Sender: news@cco.caltech.edu
- Nntp-Posting-Host: chairie
- Organization: California Institute of Technology, Pasadena
- Date: Wed, 29 Jul 1992 15:43:37 GMT
- Lines: 115
-
-
-
- The game tree...
-
- E4
- V4
- /W4
- P24/
- /
- E2 /
- V2<
- /W2 \
- / \
- P12/ P25\
- / \E5
- / V5
- E1 / W5
- V1<
- W1 \ E6
- \ V6
- \ /W6
- P13\ P36/
- \ /
- \E3 /
- V3<
- W3 \
- \
- P37\
- \E7
- V7
- W7
-
- En = Expectation (in % of wager) at branch point n. n = 1, ... ,7
- Vn = Variance at branch point n.
- Wn = Wager placed at branch point n.
- Pnm = Probability of going from n to m.
- P12 + P13 = 1
- .
- .
- P36 + P37 = 1
-
- Note: Expectations can be negative.
-
-
- The game is the sequential game tree given above and is
- repeated over and over. The problem is to determine
- W1...W7 with the following constraints.
-
- 1. Minimize risk of ruin.
-
- 2. All wagers must be > 0.
-
- 3. Wagers cannot differ by more than a factor
- of 2 between branches. Thus,
-
- .5*W2 <= W1 <= 2*W2
- .5*W3 <= W1 <= 2*W3
-
- .5*W1 <= W2 <= 2*W1
- .5*W4 <= W2 <= 2*W4
- .5*W5 <= W2 <= 2*W5
- .
- .
- .5*W3 <= W7 <= 2*W3
-
-
- Risk of Ruin:
-
- B + Et*x
- RoR = ------------
- (Vt*x)^1/2
-
- RoR = Risk of ruin measured in standard of deviations
- from the mean.
-
- B = Bankroll (constant)
-
- Et = Expectation (in % of bankroll) of the entire game tree.
-
- Vt = Variance of the entire game tree.
-
- x = The number of wagers when RoR is minimized.
-
-
- With a little work, it can be shown that...
-
- B
- x = ----
- Et
-
- Thus,
- B + Et*(B/Et)
- RoR = ---------------
- (Vt*(B/Et))^1/2
-
- (B*Et)^1/2
- = 2*(------)
- ( Vt )
-
- So, my real question is how does one maximize the ratio (Et/Vt)
- given the constraints above?
-
-
- The method is more important than the answer to the above example; it
- was only used to hopefully give a clearer picture.
-
- If the answer is too long or complicated to post, or is so trivial
- that it is taught is Math 101, a reference would be appreciated.
-
- Also, if I have left something vague or unclear, please email me
- and I will try and clear it up.
-
- Thanks,
- Chris Hall
-
-