Except that unlike with runners - who are tested against a single skill, in which a 1% improvement of physical performance means a gold medal, soldiers are tested in a variety of ways, in which mere physical strength, even boosted by 10, 20% will not boost your chances of victory all that much.
Need I remind you about that little thing Samuel Colt did?
If you read the entirety of the linked article and its links(0) you would find that when males and females compete in ways that include several/many physical attributes, the deltas are larger than for one attribute. IOW, top male sprinters may only be 1 1/8x (1.125x) faster than top female sprinters, but male top basketball players(1)
wipe the floor with top female basketball players. Basketball calls on several attributes where females are lacking and the total task disadvantage is greater than the sum of the disadvantages.
We don't have "Top Men's Basketball Team vs Top Women's Basketball Team" or "Men's All-Star Team vs Women's All-Star Team" games because it would not be fun to watch a one-sided blowout and we, as a society, generally do not like watching women get the crap kicked out of them on the teevee(2). Also, such a spectacle would be Bad for the (Feminist) Movement.
Then, there are issues of unit cohesion, elevated cost to train to the same standard(3), etc., etc. Frankly, it is just a mess and no number of movies by James Cameron or Joss Whedon will make sense of it.
"God created man, Sam Colt made them equal."
Nice sentiment and very helpful, indeed, but not 100% backed by empirical evidence.
(0) Yes, kinda onerous and two links deep, so this is not "Why didn't you RTFA?" snark, but, "Hey, RTA because it is full of tasty factualness."
(1) Basketball using acceleration, top speed, mass, height, among other attributes
(2) But many seem to be just fine with them getting killed in far off lands for the benefit of illiterate pedophile goat herders.
(3) Yes, it costs more to train women to the same standard as men.