## statistics and applied probability

### national university of singapore

In case you ever wondered:
“If a man/woman can expect to meet exactly N eligible partners in his/her life, what strategy should he/she use to maximize his chances of choosing the very best one?”
http://www.mathpages.com/home/kmath018/kmath018.htm

## 1 Comment

1. You can also use it to optimise your secretary! (It’s sometimes called the secretary problem.)

Good to see one of the simplifying assumptions being removed (namely that your objective function is to get the best or nothing) although instead of calling it a weighting function it could just be called a utility function, for that’s what it is.

But still, the main two problems to me seem to remain:

• you have to know how the spouse scores in absolute terms, even compared to the potential spouses you haven’t met yet;
• you have to know the number of potential spouses you will meet over your lifetime.

I think this is crying out for a proper Bayesian analysis, with a prior distribution for the number of potential spouses and for the “quality” of these spouse (i.e. his/her utility to you!). The latter would be especially interesting as you could update it as each courting happens. My suspicion is that the simplistic decision rule would no longer work.

Theme by Anders NorenUp ↑