[kgr]

I would say that the best programmers don't change the multiplier, but the exponent, so that the factor by which they're better than the average increases with the size of the problem. This is why you might not see as much of a big difference with a test lasting a few hours, as you would with one lasting weeks, months or years.

I explain this a bit by means of a simulation in this video: https://www.youtube.com/watch?v=3Ea3pkTCYx4