Weekly firepower rankings, Day 3591
TheJuliusCaesar
After some number crunching, it is again time to publish the weekly firepower ratings of most of the world's nations and the change in them compared to last week. If you notice your country missing, feel free to PM me your country's firepower and I'll update the lists. Or you can comment below.
The following charts are computed simply by comparing the each nations' share of global damage to their firepower. The more a nation produces damage compared to its firepower, the more efficient the nation is on battlefield as a deployment. This is because they offer more damage per firepower than some other nations. Whereas firepowers need to be manually fetched and inquired from many sources, nations' share of (air)damage can easily be obtained by dividing the amount of respective nation's (air)damage with the total amount of total global (air)damage, for ground and air battles respectively. Damage can be fetched from the weekly leaderboards.
Here lieth this week's tables:
GROUND BATTLES
|
|
AIR BATTLES
|
For the actual dataset and other miscellaneous computing, as well as sortable firepowers can be found from this sheet. Currently it has 91.62% of world's firepower accounted for, but many (small) nations are missing.
Comparing firepowers between last week and this week reveals a bit more changes than last week's comparison, probably affected by the rather widescale invasion of Albania, Italy and Turkey against Pacifica and especially Russia:
It is to note that Serbia has probably been consciously pushing their firepower down to make room for at least some allies to be deployed. Two weeks back the firepower was roughly 11%, which is just too high.
If you can't find your nation from this list, it means we are missing the firepower of your nation for at least one of the two weeks.
I am, again, indebted to Jordic69 for his help with collecting and analyzing the data.
See you again next week!
- Caesar
Comments
Voted.
Great job
о7
Nice job! Thx 😃
Did you use those new entries to check the model that you showed in your first article? Is 1/3 aircraft share - 2/3 ground share confirmed ?
Excellent question! We have been forking around with different formulas with Jordic69 ever since, but it is extremely hard to distinguish the coefficients because of the big (0.99) correlation of the variables. Currently we are leaning towards either
((3/4)*kills) + ((1/4)*airkills), which would make sense because 1/4 of battles are air rounds,
or
((3/5)*kills) + ((2/5)*airkills), which would make sense too because air rounds constitute 2/5 of the campaign points.
In any case, the predictions of those two (and some other) formulas are pretty close in overall, so it doesn't make much of a difference. The earlier 2/3-1/3 split provides good estimates too, though a bit more uncertain ones.
It would need more weeks (=more data) to determine between the exact coefficients.
I thought it was damaged, not kills ! Kills give a better idea of military activity, but not firepower... Both are acceptable
Kills and damage are too very much correlated (😃), so it is hard to be certain of that too. But so far kills give a bit more accurate predictions than damage formulas. I remembered wrong that we used kill formula already back in that article, but apparently not.
All of them give pretty accurate results.
Both formula's make sense indeed. If I were one of the admins I'd use a combination of both and differentiate between divisions therefore having 5 formula's leading to the end result. As it was abouth heath spend it likely isn't kills nor damage yet energy usege of the countries soldiers which we can't extract directly from either kills nor damage. Kills will be closer to it I guess.
I in fact formed a couple of composite variables (both 1/4 and 2/5 splits) from damage/airdamage and kills/airkills. Produced about the same results.
Yeah, it might as well be there's a formula for each division and then they're combined via some coefficients. That would make sense too, as would the health usage 😛
Problem is all this is so closely correlated it would need tons of data to differentiate and play with the errors. We'll see 😁
don't forget to account for League of Allies last week.. That would have some influence due to the amount of gold handed out.
Without deeper pondering, I'd expect it to just redirect damage from TP battles to allied battles for the week, not necessarily altering the total output of damage. At least I couldn't figure out the mechanism.
Though maybe tanks tanked more? I need to take a look at the total damage figures.
I'd say they didn't get that much damage since they fought out of to battles so legendary bonus applied quite lower. In the air, more damage was done for sure.
Checked it, apparently all figures got clearly upwards: http://prntscr.com/gn2xca
Damage output/kill might be lower since, yeah, some people couldn't use their legend bonus in allied battles. Others bypassed that by taking a different citizenship for the duration of the tournament. Share of NE damage was probably lower, too.
>Though maybe tanks tanked more?
Well I doubt gnil always tanks up four levels per average day ( https://www.erepublik.com/en/main/rankings-citizens-experience/weekly/overall ).
I think lots of people normally not fighting or just doing DO jumped in during the event to harvest daily gold which was very rewarding. Also I estimate several players just use EB's when they get them or use them in events if they have some stacked therefore increasing damage during the event and compensating loss of legend bonus.
no cyprus nor boobs
no vote
Add Ireland 🙁
o/
nice work
V+