Deal or No Deal

October 30 2006

Josh's blog on Deal or No Deal was very interesting.  I've watched the show a few times and find it quite amusing.  Regardless of who is right or wrong in the "challenge of rationale" between Josh and the ego-centric, self-centered, but immature other person (blocking Josh's email just because he didn't want to engage in an intellectual challenge) ... what is quite clear is that people are IRRATIONAL ... each one thinks that they can beat the odds.


In a world of random events and probabilistic outcomes while it is possible to temporarily "beat the odds" in the long run such is absolutely impossible .... a fair coin if flipped enough times will revert to the mean of 50% heads, 50% tails ... and each outcome is independent.


But a person on D or N-D sits there and sees $5, $100, $100,000 ... with an offer to sell out at whatever price ... yet they fixate on the "possibility of hitting $100k".  The real key to the game (aside from the discussion on mean vs expected value) ... is how many briefcases must be opened on the next turn before you get a new offer ... and that is what Josh was alluding to.  If you have $5, $100k, $400k, and a $1M on the table ... and the next step is to open just one brief case ... that looks very different than if you have to open 2 .... In a 1-briefcase scenario the expected value of the next turn would be $168k to $500k (and if the CURRENT offer is below $168k then he should not take the offer).  But, in a 2-briefcase scenario then the range of outcomes is $53k to $700k ... and that my friend is the real challenge .. with such a wide range in the NEXT round ... you would EXPECT to walk away with as much as $700k or as little as $53k ... but it really has nothing to do with the current values on the table ... only the OFFER vs the NEXT round ....


... Josh ... you agree or disagree?

Josh Morgan

October 31 2006
I think I agree (didn't check the exact math), but should add something that I just noticed yesterday. If you take the expected value of the possible following offers, they'll be the current mean! 1 briefcase is easier to consider, so if you have your $5, $100k, $400k, and a $1M, take the average for each possible case being removed, and then average those averages. It will turn out to be the average of the numbers you already have. Eliminate $5 - new average is $500k; Eliminate $100k - new average is $467k; Eliminate $400k - new average is $367k; Eliminate $1M - new average is $167k. Like you said, you can't expect to beat the mean in the long run and it turns out the average mean will always be the same! I should have seen that coming! I suppose this means you have to count on the next offer being an increased percentage of the mean (because it usually is) if you're going to make any money. If you assume the offer will become a greater percentage of the mean and you accept that the mean will not change, then the only deal you can make is the last deal. I'm still chewing on this. Does that sound right?

Josh Morgan

October 31 2006
Oh... to clarify, the average of $5, $100k, $400k, and $1M (original cases) equals the average of $500k, $467k, $367k, and $167k (possible new averages) before rounding error. If you took away two cases or any number of cases, the average new average will still be that first average.