Understanding the Monte Carlo Method

I personally think it's a load of BS until we start having sprints in at least in the high 20s because the equivalent bounding area will vary over time, based on staff availability, overtime, skill set required, complexity, and tech debt etc etc which can only be accurately estimated using many actual samples.

With a high number of sprints the stats can effectively be collapsed into a bounded "box and whisker" average and confidence intervals anyway.

This means there is no need for Monte Carlo in this field since it too only approximates using probability.

The non technical scrum master is unable to explain to me how this works as it's black magic to them and all they do is fill in fields in a Jira plugin, and doesn't understand it at all (or my attempt to find how it's lack of data for the min/max - the bounding box - is somehow magically offset by 10,000 simulations) whilst he's trying to sing it's virtues to us. :shrug:

I'd be interested to hear if anybody else can tell me why it would be any better if all it's doing is basing it's artificially generated "within area" values against the probability determined using only a small set of real samples (sprints).