> If you let 1000 people flip coins 1000 times and count the number of heads, you'll see an extraordinarily high number of heads among the most successful people.

Not really. If you do this experiment you are virtually assured that all the people will get between 400 and 600 heads. Not a single one will get 601 heads. Not a single one will get only 399 heads.

Zuckerberg is not one or two or ten standard deviations away from me. He's thousands of standard deviations away from me. Luck alone cannot explain his success.

Edit: For fun, here's a code snippet that shows the range of the number of heads that 1000 people get when flipping a coin 1000 times. I ran the experiment 10 times.

  import numpy as np

  for seed in range(10):
      flips = np.random.RandomState(seed).binomial([1000]*1000,p=0.5)
      print(min(flips),max(flips))

This is what you get:

  442 556
  455 546
  453 550
  452 551
  446 557
  444 552
  435 549
  446 564
  454 562
  452 548

You need to run this experiment many millions of times before you see a minimum less than 400 or a maximum higher than 600. As you can see most often the range is much tighter.