vpFREE2 Forums

why CE is better than EV

#1

Just what is the Certainty Equivalent (CE)?

I have searched for definitions, and here's what I found:

1) Certainty Equivalent: A guaranteed return that someone would accept, rather than taking a chance on a higher, but uncertain, return.

2) The amount of payoff that a player would have to receive to be indifferent between that payoff and a given gamble is called that gamble's 'certainty equivalent.'

Note the references to "someone" and "a player." Thus, the CE is a subjective measure, and the CE for a given play is likely to be different for you than it is for me or anyone else.

To actually come up with a meaningful value for CE for a play requires each player to somehow quantize his/her emotional tolerance of Risk. I don't know how to do that. Before using CE, I suggest someone should come up with some way to assign a number to one's emotional tolerance of risk.

The first example used for this thread was dealt 22266 in FPDW. Even though the EV of holding just the deuces is slightly greater than the value of the quints, I have always recommended always holding the quints. This falls in nicely in my definition of Optimum Play.

Only a dedicated pro with a large bankroll should discard the 66.

Dan

ยทยทยท

--
Dan Paymar, developer of "Optimum Video Poker" software
Visit my web site at http://www.OptimumPlay.com
"Chance favors the prepared mind" ~ Louis Pasteur

#2

Dan wrote: "To actually come up with a meaningful value for CE for a play requires
each player to somehow quantize his/her emotional tolerance of Risk. I
don't know how to do that. Before using CE, I suggest someone should
come up with some way to assign a number to one's emotional tolerance of
risk."

True. A degenerate gambler would take a gamble even if the EV was less than the guaranteed amount, and a totally risk averse person would never gamble and would always take the guaranteed amount. You see this at drawings where the prize is say $1000 freeplay or $100 cash and many people take the cash.

I should have specified the Kelly CE, which is approximately EV - Variance/Bankroll/2. Kelly made certain assumptions to derive his CE, so if you disagree with his assumptions you would probably disagree with his approximation for CE.

[Non-text portions of this message have been removed]