MPP deployment - formula, efficiency

Day 3,580, 05:10 • Published in Finland Finland • by TheJuliusCaesar

Evening peeps,


During the last 36 hours or so, some of us have been crunching numbers and fitting linear models to get to the core of the MPP deployment update. It seems like we came close enough of the formula which determines how the firepower of each nation is calculated. This should help citizens to maximize their country's damage while minimizing its firepower (to make more room for others) for their own part.

Further, it is rather straightforward to compute which nations (this week) offer most "bang for the buck", which nations bring in the most damage compared to their firepower. For air battles and ground battles, respectively. These constitute the two parts of the article. Overall data and exact specs of the nations can be found from here. Feel free to use it for your own projects. Remember, they are only valid until Tuesday, after which percentages need to be calculated again and firepowers change according to the latest weekly challenge.

Following extensive testing and deducing, it seems that the firepower of each nation is pretty much just a weighted average of their weekly share of global kills on air and ground battles. In other words, and more mathematically, the formula was iterated to be, as a rule of thumb,

Firepower = ((1/3)*share of weekly global air kills) + ((2/3)*share of weekly global kills)

Exact coefficients would be 0.3389 and 0.6594, respectively, but the tiny rounding making the formula more aesthetic and easier to remember do not make it enough more inaccurate to be significant. The third - two thirds split is a bit odd, considering that the relative weights of air and ground battles could be characterised as 2/5 and 3/5 respectively, per the points they bring in battle. It might be possible they have computed that ground battles "decide" more of the battles regardless of their actual point input, but then again I doubt the share would be exactly an aesthetic 2/3. Or then they just decided it's a nice and round fraction, even though I don't really believe in that. Or we are on a completely wrong track here. Who knows?


The formula has an explanatory power (R2) of 99.5%, meaning the factors explain the firepower almost completely. The rule fo thumb formula produces such results that the standard deviation of the error between prediction and actual firepower is 0.15% points, meaning that the predictions tend to be rather close to the actual firepower and only differ fractions of percentage points, mostly.

That said, there were a few bigger errors, namely Finland (0.58% point error in predicted vs. actual firepower), Serbia (0.44😵, Poland (-0.40😵 and Hungary (-0.30😵. All other mentioned nations except Finland having a quite larger overall firepower, the biggest exception is a really odd one.

From this outset, we can further deduce that for nations to minimize their firepower while maximizing their contribution would be most efficiently done by having the maximum damage inflicted per kill, as the kills pretty much determine the firepower. If a nation has little kills compared to the damage they've done, it means the nation is "punching above its weight" and offers most damage with least firepower used.

Now I am not particularly well versed in military module, but if my memory serves me right, Guerrilla battles (yes lmao, are they still a thing? Do people use them?) offered considerable damage relative for the kills. Vice versa, bazookas are practically the worst thing for nations' firepower, as they get you a kill with just 10k damage.



Talking about punching above the weight, it is relatively easy to compute which nation is punching the most above by comparing their respective firepower and share of weekly damage. Of course, the results vary whether it comes to air battles or ground battles, so military leaders would want to deploy different nations for different rounds.

Without further ado, here are the 20 nations giving more damage in ground battles than their firepower would indicate, and vice versa those 28 nations giving less damage in ground battles than their firepower would indicate:


Here are the similar statistics for air battles, 25 nations get more damage done on air than their firepower indicates, while 23 nations get less damage done:


Complete statistics can be found from the document linked in the first paragraph if you want to play around with it. As you noticed, I'm missing 26 (small) nations' firepower from the statistics. However, they would only account for 8.49% of world's firepower. A global database of weekly firepowers would be a welcomed addition (hint hint, Plato).

Some observations to make from the above graphs:
- Italy, Poland and Republic of Macedonia (FYROM) have significantly bigger share of world damage in both air and ground battles than their firepower indicates.
- Finland, partially probably because of their exceptionally high and oddly strong firepower, has significantly lower share of world damage in both air and ground battles than their firepower would indicate.
- Russia is certainly concentrating on air battles more than on ground battles, which shows as them having considerably bigger damage share on air battles, and conversely significantly smaller on ground battles.
- Slovenia and Serbia, then, are certainly concentrating on ground battles more than on air battles, which shows as them having considerably bigger damage share on ground battles, and conversely significantly smaller on air battles.

I am indebted to Jordic69, She Is A Gun, n0s3 and inpoc1 for their insights and calculations regarding the subject. I wouldn't have been able to ponder it out without their work.

I'm writing concerningly much nowadays,
xoxo,
- Caesar