For the proper understanding of this post, you will need to know about the 37% rule and the secretary problem. You can find information on both in my last edition of "Algorithms To Live By". If you do know what they are, you can continue reading without any worries. In the last edition, I mentioned how the 37% rule can be applied to scenarios that are very similar to the secretary problem. The most similar example I could think of was the housing market in places like New York City or San Francisco. Nevertheless the secretary problem can also be seen as a simplified version of more complicated problems, such as partner selection. If you want to choose the best from a pool of candidates, the 37% rule can be applied (with some modifications).
If you recall properly, the 37% rule states that in order to have 37% chance of choosing the best option in a pool of candidates, you need to study or go over 37% of the options without commitment and afterwards leap and choose whoever surpasses the established benchmark. If applied to partner selection, the 37% rule says that if you want to marry the best candidate in the world, you would need to date without commitment 37% of the candidates. Afterwards, you would have to leap and choose whoever seems good enough after collecting all that data and experience. As you might be thinking, this is completely unrealistic. Who has the time, energy, or money to date 37% of the world's men or women to then be able to make a choice? Thankfully, this assumption for the 37% rule is flexible and can be interpreted in different ways. For example, instead of thinking that you want to choose the best woman in the world, you could think that you want to choose the best woman among all the women that you will meet during your reproductive life. Michael Trick did this.
Mathematicians have been having trouble with love since at least the seventeenth century. (Algorithms To Live By, p. 16)
Michael Trick, now professor of operations research at Carnegie Mellon was once a graduate student looking for love. As the scientist that he is, he saw the similarities in looking for a partner and the secretary problem. Therefore, he decided to run the numbers. Assuming that his search for a partner could run from the ages of 18 to 40, the 37% rule gave him the age of 26.1 as the time to stop looking and start leaping. As it turns out, at that point in time he was 26. So, that's exactly what he did. When he met a woman who was a better match than all the ones he had dated so far, he proposed. Neither the algorithm nor all his knowledge of the secretary problem prepared him for what came next. She turned him down. This is playing against the rules according to the secretary problem! Trick had experienced first-hand one of the oversimplifications that the secretary problem has, the possibility of rejection. As it turns out, if the possibility of rejection exists, the maths suggest that you should start leaping after only 25% of looking! In turn, that approach would give you a 25% chance of finding the best match. Still pretty good, I'd say.
I wanted to study this specific case to point out the weaknesses of algorithms to live by. Most of them will be general heuristics that enable you to make better decision but do not ensure a perfect outcome. In the real world, assumptions and rules are broken all the time and many more variables come into play in any decision process. The purpose of this series is to share knowledge and enable any reader to make better decisions. Nevertheless, beware of oversimplifying the situations that you find yourself in. Make sure to study the problems you are facing before choosing a strategy to confront them. Best of luck applying the 37% rule!
If you want to check out other thoughts that this awesome book has evoked, click on these past posts:
- Algorithms To Live By #1: Expectations, Authors, And Algorithms
- Algorithms To Live By #2: Optimal Stopping - How Long To Look For And When To Stop To Find The Best (The 37% Rule)