And if you do your homework you don't need to even start some of the failed companies. By check the references and ask the right questions you can avoid some of the failures before they even happen.
Of course you must try, but don't spend 18 months trying something that three days of research would have shown to be doomed from the start.
This is my exact definition of entrepreneur. Someone who can solve many simultaneous and interdependent problems all of which take "n(i)" attempts where "n(i)" is "much larger than what most people imagine".
A funnel.
You try 10 companies, 2 work etc.
You have 10 co-founders, each time you learn something.
Don't understate the value of luck and randomness. Even Bill Gates realizes he was just plain lucky as much as he was smart, had the tech, timing etc.
The more you try the more you can take advantage of randomness.
IE if you don;t play you can't win.