Forged Alliance Forever

Re: unit analysis

2015-01-12T13:12:59+02:00

You can sort of account for range by assuming that the unit with the longer range is stationary, and the unit with shorter range is "charging" in at full speed. You know the difference between range_short and range_long, and you know the speed of the unit with shorter range, so you can easily calculate how many seconds that unit is exposed to the fire of the other unit without being able to return fire.

This isn't always accurate - some units are able to kite - but in many cases, it is accurate. A lot of units cannot fire behind them as they are running away, and other units gain a large amount of inaccuracy when moving, such as sniper bots, so kiting is a bad plan.

It would be interesting to apply this analysis to comparing the balance of this game's various battleships. The Aeon get the shortest range battleships but they have a high movement speed - would be interesting to compare them directly to the Summit class.

As for mass/energy, or unit bulkiness, etc - it isn't possible to compare balance precisely in a mathematically rigorous way with so many confounding variables, but you can at least get a first order estimate. The next thing to do would be to test your model against actual unit matchups and compare.

Statistics: Posted by BrickedKeyboard — 12 Jan 2015, 13:12

Re: unit analysis

2015-01-12T02:11:41+02:00

KrogothFTW wrote:
Party time, build resources are a function of game time (the only real resource) and the number,level of producers (limited by an exponential function of game time). At the beginning, 1 mass is equivalent to 20 energy, no?

Well, wouldn't it be better to use the mass cost a t1 pgen?

Statistics: Posted by nine2 — 12 Jan 2015, 02:11

Re: unit analysis

2015-01-12T01:48:08+02:00

ZLO_RD wrote:
NodCommander wrote:
http://spooky.github.io/unitdb not really sure how much improvement is done there, Fire cycle thing is much better for old database
I didn't read much of this thread but Fire cycle is really important, that is reason why some times hunters beat mechmarines

I use the interface trom this to pick the unit filename, then copy and paste those to the old one's URL.

Statistics: Posted by KrogothFTW — 12 Jan 2015, 01:48

Re: unit analysis

2015-01-12T01:12:24+02:00

NodCommander wrote:
speed2 wrote:you can find units stats here http://content.faforever.com/faf/unitsDB/

Sorry for the off-topic, but wasnt there an improved version of that database?

http://spooky.github.io/unitdb not really sure how much improvement is done there, Fire cycle thing is much better for old database
I didn't read much of this thread but Fire cycle is really important, that is reason why some times hunters beat mechmarines

Statistics: Posted by ZLO_RD — 12 Jan 2015, 01:12

Re: unit analysis

2015-01-11T23:21:30+02:00

speed2 wrote:
you can find units stats here http://content.faforever.com/faf/unitsDB/

Sorry for the off-topic, but wasnt there an improved version of that database?

Statistics: Posted by NodCommander — 11 Jan 2015, 23:21

Re: unit analysis

2015-01-11T20:10:32+02:00

Party time, build resources are a function of game time (the only real resource) and the number,level of producers (limited by an exponential function of game time). At the beginning, 1 mass is equivalent to 20 energy, no?

Statistics: Posted by KrogothFTW — 11 Jan 2015, 20:10

Re: unit analysis

2015-01-11T20:00:58+02:00

Vee, so you call it "strength".

Statistics: Posted by KrogothFTW — 11 Jan 2015, 20:00

Re: unit analysis

2015-01-11T14:29:39+02:00

Well there is MAX range and EFFECTIVE range... Thanks to each weapons preset firing inaccuracy.

Resin

Statistics: Posted by Resin_Smoker — 11 Jan 2015, 14:29

Re: unit analysis

2015-01-11T12:29:46+02:00

Mathematically if you ignore range and pretend that damage is applied in a continuous fashion rather than in discrete shots then the strength of a unit is dps*hp: a unit with a higher dps*hp will win against a unit with lower dps*hp. Suppose you have two units fighting each other, having dps1, dps2, and hp1, hp2 respectively. Then the lifetime of unit 1 is hp1/dps2 and the lifetime of unit 2 is hp2/dps1. So if you want to know who dies first you need to compute the minimum(hp1/dps2, hp2/dps1). We can multiply and divide by dps1*dps2 to get minimum(hp1*dps1, hp2*dps2)/(dps1*dps2). So you see that who wins depends on hp1*dps1 vs hp2*dps2. So we could say that hp*dps is the strength of a unit.

A similar argument shows that dps*hp/mass^2 is the strength of a unit if you also take cost into account. If you compute that number for various units you'll see that it's very consistent: even very different units with vastly more hp or dps still have roughly the same dps*hp/mass^2. Of course in practice range and speed and how bulky units are matters so much that this theorycrafting is useless for actually playing the game.

Statistics: Posted by Vee — 11 Jan 2015, 12:29

Re: unit analysis

2015-01-11T10:03:34+02:00

i think the short answer is 'no'.

i used to think about dps per mass and hp per mass.

but even that is pretty useless. Range is a huge factor in how useful a unit is, any maths you did would be discounting that, which is a big problem

similarly, people used to try and determine the cost of buildings by multiplying the mass and energy together to get "cost". energy was weighted less. which just doesn't work.

Statistics: Posted by nine2 — 11 Jan 2015, 10:03

Re: unit analysis

2015-01-11T09:57:19+02:00

Thank you, I know.

Statistics: Posted by KrogothFTW — 11 Jan 2015, 09:57

Re: unit analysis

2015-01-11T09:36:58+02:00

you can find units stats here http://content.faforever.com/faf/unitsDB/

Statistics: Posted by speed2 — 11 Jan 2015, 09:36

unit analysis

2015-01-11T08:28:56+02:00

Is there a special term for the mathematical product of hit points and DPs? It seems like it ought to be a useful measure of how much damage a unit will do while under fire, before it dies.

Statistics: Posted by KrogothFTW — 11 Jan 2015, 08:28