More randomness

[Tazak]

When selecting units in the scenario editor there is a degree of randomness in all subunits readiness/moral/ammo.

If you select one of the sub-formations and edit its readiness/moral/ammo, if you select to apply to all sub-units they all receive the same readiness/moral/ammo.

Can there be a -/+ 2&-3% (or however much) of randomness when readiness/moral/ammo edits are applied to all sub-units

[CapnDarwin]

Sounds like a good idea. We should be able to add a check box and spinner for setting the min/max variation.

[IronMikeGolf]

A statistical refinement, if I may.

When we set a value (e.g. Readiness) of a HQ unit and propagate that to subordinate units, we are saying the unit's Readiness is x. We can take that to be the average (more precisely, the mean) of HQ and all subordinate units' readiness. I'd bet everyone who makes a scenarion thinks of this stuff in these terms.

So, if we have a standard deviation value ( the +/- spinner), we can generate a Gaussian distribution of values. What's good about doing it like that is if we average the generated values for the HQ and all subordinate units, we will get the number we originally set the HQ to. If you just use a regular random number generator and then use an offset based on the spinner, you are eventually going to get cases where you specify a battalion to have, say, 65 +/- 10 and end up with three quarters of the subordinates under 60. Probably does not matter if we don't have many subordinates and have small spinner values.

[CapnDarwin]

Ow, Jeff is using statistics on us during the weekend. A good point of course, but I believe with a small delta and the small number of subunits we should be okay.

[IronMikeGolf]

Well, there's code already out there to use that does this. Wrapper classes and I found someone who made an extension method for the .Net Random class. You don't have to invent the algorithm and implementations exist.

See this

[CapnDarwin]

Thanks, I'll check it out.