I already started to talk about this thing in the thread update that was recently posted, and i am interested to make perfect balance between units using this indicator. We just need a frame, or a foundation where we can start and balance weaponds for commanders, and even units

Curently i want to finish weaponds commanders and d gun PIV indicators

How will PIV bee ?

Examples:

Give AOE, Range, DMG, Reload time, stun dmg, slow dmg, burn time PIV indicators
so we will have something like this:
This is just an example

10 Dmg=10 Piv
reload time 1 second= -50 Piv( reload time 2 seconds 0 Piv, relod time 3 seconds 50 PIV)
10 Range= 14 PIV
30 Stun dmg=10 Piv
30 SLow dmg= 7 PIV
Burn time= 10 PIv

From these we calculate for each weapond or d gun this PIV indicator and adjust all weaponds acordingly to a maximum PIV number like 3000 for example.
THis is just an example and i repeat it will take a lot of time.

From this PIV we can adjust units to, speed, dmg, range, Hp, and so on.I hope with this program if il find the corect PIV values for all we will be able to create and adjust units easily and no more balance problems will appear.

A unit's rare strength = dmg * hp.

Ofcourse dmg when it hits, so you have to calculate range diffs, speed and accuracy - and micro.

Slow + Real damage doesn't work that simple on a sniper - either you kill it or not (just an example)

Range is more important on statics, because you need to build less to cover same space vs raiders.
HP is important for low ranges, can be made unimportant by shields, but still count early game!

Do you already know the solutions?
Maybe you can balance similar weapons this way, but you need more than just one number for a unit.

[color=grey]EDIT @Anarchid: Human brains are computers. Humans balance units. -> Thus math balances units.

EDIT2 BTW: I don't mind if we limit post-rights to neonstorms...
This thread and Forever are officially neonstorm-tagged now, too :)[/color]

NeonSturm> Human brains are computers -> Anarchid> I'm going to derail a neonthread by saying that you cannot prove it.

You can not prove the opposite also - and I don't want to start a religion-IV thread.
I think that is not what Forever wants as a thread-starter.

quote: You can not prove the opposite also

I am not required to prove opposite for your argument to be destroyed.

If your statement that human brains are turing machines is not provable, then your argument that brains

computers

maths becomes unsupported speculation and thus void.

[2up]knorke i just tried to find 2 of the most balanced weaponds for commanders.

So i took laser beam and Rocket lanuncher and tried to give them PIV to theyr DMG, AOE, Attack speed, range etc. S
o it goes like this:
Beam laser:
Range:300
Damage:155
Reload time:1 second
AOE:0

R. launcher:
Range:
430 Damage:360
Reload time: 3 seconds
So from this point of view these weaponds are verry well balanced i gues( i hope i am not mistaken).

FOr example in case Rocket lancuher has 1,5 seconds reload time it will do double the dmg.In case of Assault canon we have 360 dmg over 2 seconds. If we remove 1 second reload time we will have double dmg.

So for each second added we just decrease all the PIV with: DMG PIV /(2 X number of seconds)
Now we just need to compare the next things:
Range twards dmg
AOe twards dmg
Acuracy twards dmg
And then come up with a similar formula like i tried to make with dmg in relation with reload time.

Neddie's post, the first one, is the best overview of the topic. The equation used there is very good, but the most important thing to understand is it's assumptions, IE, that all units are doing damage to each-other uniformly.

CA got even further than this, we started to do calculations not just on the expected damage over a units lifetime in combat, but the expected damage from outranging an enemy. This bonus damage is based on effective range, which also takes into account kiting (your effective range vs an enemy you both outrange and outpace is infinite).

By this point you start realizing it needs to be broken down not just by effectiveness, but effectiveness vs a class of enemies (Performance vs skirms or vs raiders differs, as does artillery performance vs turrets, which is again, infinite). Then you realize that the most important factor in what enemies a unit faces is how popular those units are on the battlefield, and you need empirical statistics on what is actually built...

By this stage we realized this is much less effective than tracking actual performance, which is what we've always been doing.

The other way to do mathematical balancing is to 'calibrate' it against units you consider balanced (Which is what is being done here). The problem with this is that you need a huge number of very balanced units to calibrate it against, and you never really know whether the unit is unbalanced or whether your calibration is just not taking into account why that unit is good/bad.

[color=grey]Who is Neddie? Anarchid? Forever?[/color]

Quote Saktoth> Then you realize that the most important factor in what enemies a unit faces is how popular those units are on the battlefield, and you need empirical statistics on what is actually built...

No. Peoples seem to forget about some units because they are under-used. That makes a pro different from newbes or noobs.

E.g: Hammers were OP some time ago. Why did nobody build, scynthe/sniper/glaive, scorcher/ravager, kodachi/pillager, pene/scrubber, pyro/firewalker, roach/bandit, fleas/crabe, phönix, blastwings/brawler.

[color=grey]If the cheap things doesn't work, maybe not hammer's speed or range is the issue, but hp or a too big sight radius.
A low sight radius opens a possibility for scout wars (rector has a cheap jammer) and killing enemy scouts will cost them metal if you can't afford more expensive units like crabe or firewalker.

If the enemy could spam dirtboxes, he has another factory and you can do it too.
The problem is remembering and finding the best strategy in the short time you have.
To know more than just one strategy how to beat something. The enemy can't counter everything except he is better than you are - then hammers would only be a tool like any other.[/color]

But players want to use just 1 or maybe 2 units, not 5 different.
If they have to be in different formations, depending on situation, this might be too micro heavy for most peoples.
E.g: Retreat the damaged unit out of a death ball, adjust repair zones and retreat pathes (how to use this widget? :), scout with dirtboxes, but don't penetrate the front defense line to let other defense shot at them too - impossible without (not existing) widget.

So how do you weigh in air vs ground movement? Slope tolerance? Turning radius? Flat multipliers? Something nonlinear, like "it depends on the other stats of the unts"? Then you're likely to end up with (way) more variables (as in, coefficients/parameters for your PIV functions) than you can handle.

Some more things: Projectile speed? Outranging (note that this depends on stats of other units!)? Leaving a wreck? Missile arc? Leading/non-leading?

And then of course special things like dguns, buildpower (+rez), jump, puppy grey goo ability and most importantly micro gains...

More variables but they will all depend on the first variables that i make now, like the one i made with reload time.Next one is range variable that will be based on dmg and reload time if its necessary.After that il go to Aoe and so on

DMg aoe reload time, range are the foundation of the others.If i manage to give for these a clear relation to each other then i can establish the PIV value, and from the PIV value il expand more to projectile speed, commander speed, and so on.
After this il start with Hp of units, the metal cost for different PIV numbers and so on.I think it can bring a good perspective, but first i ask this from all dev comunity that reads this:

Please give me examples of perfect balanced units and perfect balanced weaponds.
And please explain to me where can i find the full detail of units regarding all the stats that are not shown in the forrum.

So this is how it will look in the future if people agree to make the twiching program integrated in zero k:



After you press that you will recive this:



And forward:

Best thing would be to derive a best fit formula given real statistics on unit success. Then you can try models with different parameters and see what comes out of it.

I have a sneaking suspicion that most variation will be accounted for by dps and health alone?

OTOH perhaps we should just balance for unit use, and let the community decide which ones are UP/OP?