Stefan Stahl
Seituoda Taskforce Command
520
   |
Posted - 2014.05.18 15:09:00 -
[2] - Quote
In reaction to this thread I hacked some numbers into Matlab to come up with a linear model of relative utility for all armor and shield modules.
I've come up with two results. The relative utility of the stats armor HP, armor rep rate, speed bonus, shield HP and shield recharge bonus expressed in terms of the measure "total fitting cost" (TFC), which is CPU cost + 5 * PG. (This ratio is derived from the fact that almost all suits have about 5 times more CPU than they do have PG.)
The best linear fit for this model based off the current modules is as follows. One "total fitting cost" currently buys you:
1.05 armor HP or
0.05 armor HP/s or
0.10 % speed bonus or
0.58 shield HP or
0.46 % shield recharge bonus
These mean values feel kind of fair in regard to their relative magnitude. Would you pay 38 CPU and 2 PG (38+2*5 = 48 TFC) for 50 armor HP at no speed penalty? Or rather spend 50 CPU and 6 PG (80 TFC) for 50 shield HP? Would a std kincat be worth 15 CPU and 7 PG worth you? (50 TFC)
Using these weights as a basis I asked Matlab to show me which modules deviated the worst from the current balance. Here's the list of relative utility of each module. 0 means (comparatively) balanced, negative numbers mean (relatively) underpowered, positive numbers indicate a very good bang/buck ratio.
Matlab wrote:
STD Armor Plate: 46.311615
ENH Armor Plate: 25.379241
STD Armor Repairer: 16.569617
STD Reactive Plate: 14.482212
CPX Kinetic Catalyser: 13.557344
STD Ferroscale Plate: 8.176365
STD Shield Extender: 4.469270
CPX Armor Repairer: 3.924042
CPX Shield Extender: 3.407810
STD Shield Recharger: 2.880570
ENH Armor Repairer: 2.354425
CPX Shield Recharger: 2.065597
ENH Kinetic Catalyser: -0.961771
ENH Shield Recharger: -5.199050
ENH Shield Extender: -9.796095
CPX Armor Plate: -10.182913
ENH Ferroscale Plate: -14.605192
STD Kinetic Catalyser: -16.851107
ENH Reactive Plate: -19.929124
CPX Reactive Plate: -27.186393
CPX Ferroscale Plate: -37.907789
As expected, complex ferroscale and reactive plates come up at the lower end of the scale. Don't fit these, they are mathematically not worth it.
As expected, STD and enhanced armor plates are at the top of the board. If you fit these, you'll do better.
Technically minded readers may hit me up for a Matlab script to analyse my method. But be aware, I don't document well when I spend 30 minutes on a thought experiment.  |