Ivan:
Properly run simulation includes a Monte Carlo with a small—and declining
w/ the maturity of testing—probability of leaking plumbing.
In F1, strategy sims run tens of thousands of races with all the other
teams in play and w/ random but statistically validated variables. In
advance of each race….
Kinda like launches….
Bill
On Tue, Aug 31, 2021 at 5:05 PM Ivan Vuletich <ivan.vuletich@xxxxxxxxxxxx>
wrote:
At an amateur level, I think a John Carmack quote is relevant. It was
something along the lines of "In simulation your plumbing never leaks".
On 1/09/2021 3:28, Henry Spencer wrote:
On Tue, 31 Aug 2021, Ian M. Garcia wrote:
This is a complex subject with a lot of trade offs and individual
experiences and opinions. Having worked on both extremes I lean more to
hands on testing. The ideal is probably somewhere in the middle.
I should say that I do believe analysis and simulation have *value*,
particularly for things that are not readily testable. But they are the
partners of testing, not a replacement for it.
Henry