Monday, February 11, 2008

I wax statistical...

I'm watching "Deal or No Deal" -- they have a new gimmick where they're putting up multiple cases with $1 million (12 of 26 for the current contestant), replacing the other high dollar amounts on the board. However, I'm not sure it actually changes the gameplay at all.

Computing the expected value at any point is straightforward. The bank's offers, however, are not solely dependent on the expected value; the number of turns appears to be a factor. A comment on this blog post suggests offer = <value> * turn / 10 (where <value> is the expectation value). The offers the bank is making in the early rounds of the 12 $1 million case version, however, seem low -- almost as if the values on the board were not $1 million but the previous values. The episode ended before the contestant finished, so I'm left wondering at this point.

One thing which strikes me, however, is that the endgame is probably the same. If he/she makes it to the final round, in all likelihood (though I have yet to confirm this statistically) the choice will be between a low dollar amount and a $1 million suitcase. Maybe it's somehow easier to get to this endgame? Hm. You know, I'm going to have to do the math on this one.

No comments: