Z polarizers don't make sense (aka tooltip LIES)

[liq3]

So yeh, story time!

I pick Z polarizers again, thinking they'd be fun and strong unit. Homeworld spawns next to a Mark IV planet... I think I can probably take it out extremely quickly. So start preparing. Then I get curious. I wonder, just how much damage does a Z Polarizer do? It says "attack * armour rating". Ok, so vs a mark IV bomber, it should be 7000 * 96... I check the reference thing in stats... That's weird, Mark IV bombers BEAT them. Something ain't right. So I send one to an enemy planet, pause right as it arrives, and start checking the damage values. Turns out the "armour rating" is rather arbitrary. My findings.

Note: Armour rating = actual damage / 96 (damage of mark II Z polarizer). Armour value = displayed armour in the ship's tooltip. With AV/AR, lower = better.

Ship Name	'armour rating'	armour value	AV/AR
Fighter	20	200	10
Frigate	24	300	12.5
Zenith Bombardment Ship	24	300	12.5
Bomber	44	1000	22.72
Spec ops post	88	2000	22.72
Wormhole gaurdpost	2	0	DIV/0
Warp gate	62	1000	16.12
Orbital Command Station	76	1500	19.73
Siege Starship Mark III	28	200	7.14
AI Missile Gaurd Post	56	800	14.28

No pattern at all. They barley kill bombers faster then frigates (27 hits vs 30, 1000 armor vs 300). What I assumed would be the armor rating isn't the armor rating at all.

Conclusion: Tooltip needs to be fixed to match the ship, or the ship needs to be fixed to match the tooltip. I'd prefer the latter.

PS. Wasn't sure if I should put this in bug reports.

[keith.lamothe]

Oh, I didn't realize the tooltips said actually multiplied by, rather than simply proportional to.  Z Polarizers were dominating lots and lots more stuff than intended so I changed it to be multiplied by the square root of the target's armor rating (pre-computed square root, of course).  Also boosted z-polarizer base damage by quite a lot to compensate.  Just need to fix up the tooltip now to just say proportional-to.

As for their actual balance, I'm certainly open to suggestions, but the old linear relationship was too brutal.

[Ktoff]

Isn't 'proportional-to' just a different way of saying 'damage=basedamage*armorrating'? So multiplying by armor rating would still hold....

[superking]

as a connoisseur of Z polarizers, I beleive that atm they are are relatively effective.. twice the price of fighters of the same MK and a great deal more effective against all the well armoured gaurdians & gaurdposts. not so useful vs normal fleet ships now though

[keith.lamothe]

Isn't 'proportional-to' just a different way of saying 'damage=basedamage*armorrating'? So multiplying by armor rating would still hold....

On further thought the phrasing "does more damage against heavily armored targets" might be more to-the-point for the layman's tooltip.

But out of curiosity, I thought "x is proportional to y" simply meant "as y increases, x increases", not that they increase at the same rate.  As opposed to "x is inversely proportional to y" which I thought meant "as y increases, x decreases".

Anyway, I don't understand the term, I shouldn't use it in a tooltip

[liq3]

Isn't 'proportional-to' just a different way of saying 'damage=basedamage*armorrating'? So multiplying by armor rating would still hold....

On further thought the phrasing "does more damage against heavily armored targets" might be more to-the-point for the layman's tooltip.

But out of curiosity, I thought "x is proportional to y" simply meant "as y increases, x increases", not that they increase at the same rate.  As opposed to "x is inversely proportional to y" which I thought meant "as y increases, x decreases".

Anyway, I don't understand the term, I shouldn't use it in a tooltip

Pretty sure "proportional" means in a pattern. e.g. x = root(y), x = 2y, y = ax^2 + bx + c, etc etc.

Oh, I didn't realize the tooltips said actually multiplied by, rather than simply proportional to.  Z Polarizers were dominating lots and lots more stuff than intended so I changed it to be multiplied by the square root of the target's armor rating (pre-computed square root, of course).  Also boosted z-polarizer base damage by quite a lot to compensate.  Just need to fix up the tooltip now to just say proportional-to.

As for their actual balance, I'm certainly open to suggestions, but the old linear relationship was too brutal.

That gives me the impression the pre-computed values are old, since root(1000) = 31 not 22 and root(200) = 14, not 20. (game was on Normal, so all the 'armour ratings' were doubled). You also have 2 different values for 1000 armour.

If they actually WERE square roots of the armour rating I think they might be fine. They'd get 83* vs 7000 armor mark III bombers (they currently get 59).

Actually I think a good formula might be damage*root(armor * 2). It'd make them scale better against high armor stuff. e.g. 118 mult vs 7000 armor, but still only 20 mult vs 200 armor.

[keith.lamothe]

as a connoisseur of Z polarizers, I beleive that atm they are are relatively effective.. twice the price of fighters of the same MK and a great deal more effective against all the well armoured gaurdians & gaurdposts. not so useful vs normal fleet ships now though

Since you're the one who primarily provoked the balance change, I'm glad they still seem reasonable to you

That gives me the impression the pre-computed values are old

Well, to be clear, they are computed at runtime right after the armor value is initially assigned.

However, it is based on the "raw" armor value rather than the scaled (for high/normal/low caps), since this is basically like a damage-vs-hull-type multiplier and those aren't supposed to scale.  So basically it's computed from the armor rating that is used on High caps, and even there it can be different from what you see due to some rules about the armor of scaled types versus attacks by non-scaled types.  It had to be more complex than I liked due to the addition of armor rating after the introduction of changing-unit-scale, since there's no easy way to keep the math tidy there without having two separate numbers under the hood.

Anyway, I'll just change the tooltip to the vague "does more damage against heavily armored targets" or some such since there's not really anything more precise that is still accurate, without giving multiple lines of explanation

[o1knives]

Isn't 'proportional-to' just a different way of saying 'damage=basedamage*armorrating'? So multiplying by armor rating would still hold....

On further thought the phrasing "does more damage against heavily armored targets" might be more to-the-point for the layman's tooltip.

But out of curiosity, I thought "x is proportional to y" simply meant "as y increases, x increases", not that they increase at the same rate.  As opposed to "x is inversely proportional to y" which I thought meant "as y increases, x decreases".

Anyway, I don't understand the term, I shouldn't use it in a tooltip

Pretty sure "proportional" means in a pattern. e.g. x = root(y), x = 2y, y = ax^2 + bx + c, etc etc.

At least coming from the standpoint of mathematics, x proportional to y generally means x=a*y+b where a and b are constants.  x=root(y) would be x is proportional to the square root of y.  y=ax^2+bx+c is a bit hard to express nicely as a proportionality.  You would probably say that y is proportional to the square of x for large x (as with large values of x, the squared term vastly overpowers the linear term), but this only works for large values of x.  But, that is rather off topic!  I'd agree that the tooltip might just want to say that they do more damage versus heavily armored targets, or something to that effect as the word "proportional" is likely to just throw new players off.

[keith.lamothe]

Thank you for the clarification

[Ktoff]

At least coming from the standpoint of mathematics, x proportional to y generally means x=a*y+b where a and b are constants.

Just a final smart-assery: Technically the above statement is only true if b=0 ;-)

[o1knives]

At least coming from the standpoint of mathematics, x proportional to y generally means x=a*y+b where a and b are constants.

Just a final smart-assery: Technically the above statement is only true if b=0 ;-)

I would argue that it is true regardless of the value of b.  The relationship that x is proportional to y is the same (it is still a linear relationship), it is just shifted by the value b.  But, I can see where you are coming from as well. 

[Ktoff]

I learned that a is proportional to b means a/b is constant. So if a doubles b doubles as well etc. A linear relation breaks this.

Given the relationship y=0.1x+100000 would you say y is proportional to x? :-)

[o1knives]

I learned that a is proportional to b means a/b is constant. So if a doubles b doubles as well etc. A linear relation breaks this.

Given the relationship y=0.1x+100000 would you say y is proportional to x? :-)

Fair enough, though in your example, y is still proportional to x as x->infinity.