you dont trust AI results, but you trust modelling software that was based on same formulas. only difference between AI calculations and modeling software is that modeling software also includes probabilities for different fragment weight, different initial velocities for individual fragments, multiple vectors and angles. AI calculations just happen to be on extreme end for what is possible for that bomb.
also that blue color at end represent anything from 0-10% lethal zone and not 0%.
that chart is oversimplification. it only shows most likely outcome, not that there are no lethal fragments at that distance.
HFD distance shows chance of fragment hitting you is ~1% and good percentage of those fragments are non lethal(e.g. 90% for ease of calculation). that means that at 300m there is 0.1% chance of you dying from that fragment if you are hit, not that there is no danger at all.
Dunno which chart you cherry picked this time from your book.
Anyway, requirement to penetrate 6mm of mildsteel to be considered lethal for humans is ridiculous, especiatly when your book states that penetration of 6mm steel is considered effective against trucks.
If you bother to scroll down your bible, you will find the fragments do penetrate 2mm of steel at fairly long distance.
And id say if the fragments pen 2mm of steel they do quite nasty stuff on squishy humans that are no were near resistant as 2mm steel.
from same chart that shows HFD for 100kg explosives at 300m. so at 300m you get fragments that have low chance of hitting you(~1%) and if it hits you it will likely not be lethal.
anyways multiple times from what he is showing as lethal zone.