Contents

### Short description

A new article explains how to apply the mathematics of the “discerning bride problem” to dating, to determine when to choose a partner. Based on the “37% rule” that says applicants should reject the first 37% of potential partners in order to find the most ideal match, the storyline said daters cannot know how many women they will meet, so will have to set reasonable expectations, start by viewing enough options, then choose the first, the best seen before. To determine the best option, users could use a simple rating from one to ten or ask applicants to fill out criteria forms.

## A mathematical approach to choosing a girl. The 37% rule

Nowadays, many people find their soul mate on the Internet: on thematic forums and online communities, in games and on dating sites and applications like Tinder, where dating is generally put on the conveyor belt. If ten years ago, 22% of all marriages in the United States began with dating on the Internet, now the share of online dating has exceeded 39%. In fact, the Internet has become the main way of meeting men and women, both long-term and short-term. This is very convenient for geeks and techies because we get a competitive advantage using familiar tools. For example, you can support dozens of chat sessions in a desktop application or apply numerical analysis methods in Excel/Google Sheets.

*Note. By “girl” here and further is meant any object that is considered in turn from a limited pool of similar objects with different characteristics. It can be not only a girl, but also a man, an apartment for rent, a car on the secondary market, a house in the village, an employer, etc.*

If you have worked with ~~conveyor belt~~ applications for dating, then you understand the main problem: how to choose the best option from the many options that are considered in turn? Simply put, *when to stop*?

If the option is almost suitable according to a number of parameters, then should you stop at it or reject it – and continue the search? What will be the optimal strategy in conditions of unpredictability and lack of information? Fortunately, this problem has a clear mathematical solution.

In 2017, a mathematician from the University of Cambridge, Marianne Freiberger, derived the 37% rule, which she called the basic rule of strategic dating.

## ▍ Strategic dating

The choice problem can be formulated as follows:

How many possible options do you need to go through to get the best result? From the point of view of mathematics, the answer will be **37%**.

**The original rule reads as follows:**

Of all the people you can meet, meet the first 37%, then choose the first one who is better than everyone you’ve seen before (or wait for the last one, if no such person is available).

*Fig. 1*

In practice, you have no way of knowing how many women you can potentially date, so you will have to set reasonable expectations.

Here is a sample algorithm:

- Let’s say you want to find a partner within a year (a long-term partner, an ideal girl to start a family with).

- You can date a maximum of two girls per week.

- So the total pool to choose from is 104.3 girls.

*Note*. Theoretically, the total pool can be calculated according to the adapted formula of Drake, which was created to calculate the probability of meeting with an alien civilization:That is, to multiply the total number of girls in the city by the share of girls of a certain age and education, attractive appearance and ready to meet. You will get the theoretically possible maximum pool.

- So the total pool to choose from is 104.3 girls.
- We meet 38.6 girls (104.3 * 0.37) during 135 days (365 * 0.37). Obviously, we round the girls up to 39.

- After that with
*remaining pool*we choose the first girl who, according to the overall rating, surpasses all 39 candidates whom you have met before.

*Exception*. If it is not found, then select the last one in the list or increase the pool (extend the viewing time) by recalculating the formula.

The algorithm can be adapted for different situations, but the basic principle remains the same:

- View enough options.

- Then choose the first, the best, seen before.

This rule does not simply contradict everyday wisdom: if you choose too early, there is a great risk of a suboptimal choice with subsequent disappointment. And if you continue to sort through the options endlessly, then all the good options will be sorted out. That is, there must be a certain optimum.

## ▍ Calculation of the optimum

In fig. 1 general candidate pool is marked as

and the number of reviewed candidates (appointments) is accepted as

. We remind you that according to the conditions of the algorithm after

we make the final choice and settle on the best of all that we have seen so far. Cross

marked ideal candidate in the pool we are looking for.

Our task is to calculate the optimal ratio, which is necessary for finding.

For this, the opportunity to meet after reviewing the candidates is calculated.

It is obvious that if we met among , then we missed our chance and we will not find anyone better in the rest of the sample. Therefore, the ratio should not be too large and cannot be equal to 1 (the probability of meeting the maximum there is equal to ). On the other hand, it cannot be too small, because then the object found will be lower than the maximum possible estimate, which will lead to a sad situation in the long run:

*Consequences of too little *

If you add up the probability of meeting for each subsequent date, you get:

For everyone

we want to choose

with maximum probability

ie:

If we substitute the expressions from the previous equations here, we get:

Similarly, for the next person

it turns out:

These formulas can be substituted with real numbers. Let’s say we have a sample of four girls (

). In this case, there is only one value

which satisfies the equality, this

.

That is, according to the algorithm, we should ignore the first girl (25% of the sample), and choose the first girl who will be better than her from the remaining ones.

If you make such a calculation for other sample sizes, you can find out that from a certain sample size, the value approaches , namely . In particular, for a sample of one hundred girls, this will be the value, hence the “37% rule”. Everything is simple.

*Note*. The discerning bride problem (is it the secretary problem or the choice-stopping problem) was probably first formulated by the mathematician Merrill M. Flood in 1949. The first publication was by Martin Gardner in the February 1960 issue of Scientific American (*Who Solved the Secretary Problem?*, Thomas S. Ferguson, Statist. SCI. 4 (3): 282-289 (August, 1989). doi: 10.1214/ss/1177012493).

## ▍ Selection optimization

From the above calculations, you can understand that we need to choose the optimal object

*selection criteria*

by which we can compare and evaluate them.

It is clear that each person has their own criteria, it can be physical attractiveness, intelligence, charm, sense of humor, cheerful character, good nature, material capacity, self-confidence, happiness, health, height/weight, etc. Whatever criteria you are guided by, they can be formalized and evaluated numerically to compare objects with each other according to the overall score, but this is not essential, because a person still subconsciously derives the overall score without resorting to subcategories. A simple rating from 1 to 10 is enough to choose the optimal result according to the formula above.

If you don’t want to use dating sites/apps, there is an alternative to asking out girls *themselves* fill out a form if they want to meet you. For example, the writer and rationalist Jacob Falkovich (Jacob Falkovich) published the following questionnaires for all applicants who want to enter into a platonic or romantic relationship with him:

You can go even further. Another intellectual and rationalist, Damon Sasi, not only posted a questionnaire on his personal website, but also opened a page with reviews from former partners (both respond positively, but note certain shortcomings of the subject).

Posting a dating profile on your personal website (or on Twitter) is a great alternative to dating sites. Such a questionnaire shows that you are open to a relationship. According to the idea, among the visitors of your site, only those girls who are attracted to your person and/or work, that is, the initial selection has been carried out.

Even if you don’t have a personal site, you can create a Date Me document in Google Docs or Dropbox Paper with a description of yourself, photos, and a list of what you want from a relationship. Here is an example of such a document from the lovely programmer Catherine Olsson of AnthropicAI:

The link is published on Twitter and social networks, in the profile on Habra, etc. Such a questionnaire with self-presentation itself weeds out unsuitable candidates. Only those who really like you remain. Catherine herself calls this strategy “optimization for partnerships.”

And then dating sites are not needed at all.

The Internet has become the main platform for dating. Although many still have a negative attitude towards it, statistics show that this trend is only increasing. If this continues, then meeting offline will become quite unusual.

*Share of US heterosexual couples who met in one of these ways (the table does not include other answer options: in the church, in the neighborhood). Stanford University survey, 5,421 respondents*

Even now, people in the subway and on the street are mostly sitting on their phones and are not too eager to look in all directions. The only exception is bars/restaurants, where people come purposefully to communicate. So dating methods are gradually evolving. The pickup methods that worked in our youth now look a little corny.

In any case, the “37% rule” applies not only to dating, but also to any objects that you consider in turn and want to choose the optimal one.

Telegram channel with prize draws, IT news and posts about retro games 🕹️