Protection/Firepower values inquiry

[Companion]

No offense / whine intended :redface:

First, I assume all armor protection / round penetration values in SB are all estimations and the estimation results would at first actually give a range of values rather than a single, solid number.

If it is indeed the case, then what are the criteria for choosing a specific value from the range of values to use in the game?

for example: deliberately over/under estimating blue (contracted nation) / red (threat) weapon systems.

Oh, and who did the most of the maths, especially that of 125mm rounds? Mr. Lakowski I guess? :smilelove-1:

Best,

[Ssnake]

The figures I am using are indeed mostly based on Paul's work as described in, among other papers, the ArmorBasics.PDF. Other notable contributors were Andrew Jaremkow, and lately Jeff Duquette, all of them respected members of the TankNet.org forum.

Their support is essentially limited to two areas (errors still are all eSim's fault),

1. performance estimations of kinetic energy and HEAT projectiles as well as
2. estimates about armor protection of various armored fighting vehicles.

A: Ammo performance

• Usually we are using the ammo performance data directly as they can easily be parameterized or already are available as parameters. Of course, the actual figures are the "best guess" out of a spectrum of possible data. There are however too many parameter sets to perform a quantitative statistical analysis to accompany these best estimates with a confidence rating. Insofar we are decidedly unscientific (but not necessarily inaccurate).
• Then there are cases where we have access to official tabulated ballistic data.
• Sometimes we need to resort to linear interpolation among "comparable" or similar projectile data that are known with better confidence.
• Finally we have had at least one case where the direct use of official, tabulated ballistic data resulted in a systematic deviation of Steel Beasts' shot predictions and live fire range results. In this case we adjusted the ballistic parameters within SB to match the live fire results and accepted the deviation from the official data (which were based on a computer simulation anyway).
• Our model seems to work quite well for KE and HEAT rounds.
• As far as HE target effects are concerned, our model is not very good, but we don't have anything better to work with. We are assigning an abstracted "HE power" value to the projectile which we scale with the cubic root of the projectile caliber, ranging from a "HE power 25" for a 20mm round or a hand grenade to a "HE power 300" for a 105mm HESH projectile.
  The model isn't very good because Steel Beasts treats the HE explosion just like any other impact vector when in reality we would have a combination of spherical shockwave propagation, fragmentation, and incendiary effect, with wave reflection and deflection effects for all the different materials involved.
  We try to compensate for this inherent deficiency by assigning pretty high HE resistance figures to the outside surfaces of armored vehicles, and rather low resistance figures for internal components. In addition we have a category for "non-penetrating impacts" where we can simulate spalling effects and eventually assign fragmentation or incendiary effects to component categories that we know or suspect in the vicinity of the individual material that we are specifying.

B: Armor protection

• The estimates that I have are, by nature, more global than the level at which I have to work with a 3D model. So I need to break down the total protection level of certain areas of each individual vehicle which I have received - usually from Paul.
• Paul bases his estimates by screening a number of journals that are discussing the very topics of armor protection, e.g. the Journal of Impact Engineering and others (as detailed in his paper ArmoreBasics.PDF).
• In addition Paul also uses measurements of material or component strengthes as reported by contributors at TankNet.
• He then breaks down his sector analysis to the various components that are contributing to the protection value, e.g. the different layers of composite armor, fuel cells, reactive armor, and whatnot. I then pry these figures apart and apply them to the different surfaces of the LOD3 model and its internal components. It's not a perfect method especially for large volume components, but it works reasonably well for thin, hard elements like armor plate.

C: Purpose
After this little description of our methodology, let's step back a little. It's easy to get lost in the details. The fundamental question to judge the quality of a model is whether it provides useful predictions for its purpose. Our model of component damages and armor protection must suit the following needs,

1. Provide crews of AFVs with feedback about their vulnerabilities, the limits of armor protection
2. Provide gunners of AFVs with feedback about the vulnerabilities and strengthes of targets, where they may be able to resist impacting projectiles and where they cannot
3. Challenge the crews of AFVs with component failures to train not only to fight with a perfectly working vehicle but also to adapt and overcome the failure or even the error induction of malfunctions
4. Provide a damage assessment at constructive simulation level to allow wargaming with unpredictable results at the lowest tactical level (the duel situation) but still give a somewhat predictable result at the aggregated level that is in line with accepted other models

What SB Pro is not supposed to deliver is to make detailed predictions of the amount and the kind of spare parts needed after performing a certain attack as a tool of planning real-life military operations. Steel Beasts Pro is not designed as a tool that would let you run several thousand iterations of a hypothetical battle to filter promising tactical solutions from bad plans of action. SB Pro is not suited for quantitative statistical analysis to predict battle damages or the logistical footprint of operations, or to predict casualty rates.

In short, SB Pro's damage model is very detailed but not necessarily accurate. We're trying our best, but we give no guarantees.

D: Bias

Undoubtedly the model is biased. But there is no conscious bias in our model. I try to be as neutral as I can when applying estimates. There is but one rule, that I tend to err on the "worse case" when estimating the likelihood of component failures because failing components are good training. Crews must prepare for things going wrong, therefore things will go wrong in SB Pro, eventually they go wrong more often in the simulation than in real life. This is not necessarily a bad thing unless you want to use SB Pro for purposes for which it was not designed. In that case, it'd be a "you problem", not an "us problem" because you chose the wrong tool for your task.

[Companion]

Wow, this is more than what I expected, I appreciate it.

As for the worse case bias, regarding the Murphy's law , I guess it fairly accurately shows what would happen in wartime, not peacetime. (Nothing goes well when it's really needed)

Cheers,