If experience teaches us anything, making predictions is a fool’s game. But that doesn’t stop regular predictions that top women athletes will soon be competing with the best men, and may even outperform them someday.

Is there any truth to this idea, and how do intersex athletes, like Caster Semenya, complicate the discussion?

In 1992, two eminent physiologists published a paper in Nature with the provocative title Will women soon outrun men? They concluded that if women’s running performance continued to improve as rapidly as it had since the 1920s, top women athletes would soon be running as quickly as the best men and might even be faster in the future.

The public agreed; a 1996 poll reported that two-thirds of Americans believed “the day is coming when top female athletes will beat top males.”

Is that day coming and, if not, why not?

Flawed science

Let’s start with the original 1992 predictions. The authors of the Nature paper placed a straight line over the past performances of men and women, extended these straight lines into the future, and calculated that the “projected intersection (between male and female records) for the marathon is 1998”.

How did their projection fare?

By the end of 1998, the women’s world record for the marathon was still more than 10 minutes behind the men’s. In 2016, that gap has increased. The record marathon time for the men (2:02:57) is now more than 12 minutes faster than for the women (2:15:25).

So, what went wrong?

The main error stemmed from the assumption that linear improvements in the past would continue into the future. It’s a common mistake.

Using the same flawed method, another group of scientists wrote in 2004 that if current trends continued the female winner of the 100m final at the 2156 Olympics would win in a faster time than the men’s event.

They neglected to calculate that, in the same year, linear trends predicted women would be pole-vaulting heights close to 17m!

A simple example may help illustrate the dangers of assuming linear changes will continue into the future. Below is a graph showing changes in height during the teenage years (when growth is most rapid). I’ve then extended this roughly linear improvement in height 50 years into the future. This projection would have the average man more than four metres tall – and still growing – by age 70.

What actually happened

The rapid improvements in women’s athletic performance we saw in the 1980s haven’t continued for two big reasons.

The first is that discrimination had kept women from competing in many sports; it was entirely predictable that, once the sports were opened up, there would be big improvements.

The second reason is based on suspicions that many of the records from the 1980s were achieved by doped athletes. We know that the artificial addition of male hormones, such as testosterone, have especially powerful effects on female athletic performance.

This could help explain why nearly all women’s records in Olympic sprint and power events are from the 1980s.

With much greater participation by women in sport, and improved detection of doping with testosterone, there’s now quite a stable 10% to 12% difference between men’s and women’s world records from the 100m to the ultramarathon. This difference tends to be greater in more explosive events (throwing, for instance, and jumping) and less in distance swimming races.

Is there a biological explanation for this persistent difference in performance between men and women? Most scientists believe that testosterone is the most important – but not only – factor distinguishing male and female athletes.

Testrosterone’s role

Probably the best illustration of the powerful effects of testosterone is a study that compared the race times of transgender women before and after they undertook testosterone suppression to change from normal male levels to normal female levels (or even below the average).

All eight women in the study were far slower after the change than their former high-testosterone self.

High testosterone levels have also been used to explain the performance of intersex athlete Caster Semenya. But, it’s important to point out that even with elevated testosterone levels, Semenya’s best time for the 800m event (1:55:33) is far behind the men’s world record (1:40:91). And it’s also slower than the first officially recognised men’s 800m world record set in 1912.

There appears little hope of women, regardless of testosterone levels, ever being able to compete with men in most, but maybe not all, Olympic events.

So, in what events might women be able to compete with men?

Men and women currently compete head to head at the Olympics only in equestrian events (there are also some mixed events). Prior to 1992, shooting was also an “open” event, and two women have won Olympic shooting medals competing against men – Margaret Murdoch in 1976 and Zhang Shan in 1992.

In US college sports, women and men still compete against each other in shooting events and women often win. And despite it not being an Olympic event, women can compete with men in open-water swimming events.

The women’s record for swimming from Catalina Island to the California mainland (7:15:55) is more than 30 minutes faster than the male record. But these are rare exceptions to the profound genetic advantage that men have over women in most sports.

If you’re in the prediction game, it’s much safer to say there are very few Olympic events where women will ever be able to compete against men.