|
DPS for cost is something that gets brought up a lot, mainly in rants ending with "...and that's why marines are imba!!11!" (fun fact: Lings are actually the best dps for cost in the game even over marines--obviously the fact that marines have range lets them apply that dps in much larger numbers however). Obviously, it can be interesting to know these things to a point...but I got to wondering, what about the flip side? What units have the most health for cost?
Now obviously this is a complex question in practice--an armored unit may have substantially more health in theory, but also be much more vulnerable to damaging anti-armor attacks from Immortals or Tanks. Armor is also of variable importance depending on the attack--Ultralisk armor doesn't do a ton vs. Immortals, but is hugely important against marines. Zerg units have regen, and Protoss units tend to have more defense on health than they do on shields, except for the special case of Immortals.
I'm...not going to model any of that. Obviously its important, but there's no simple way to model it, short of breaking down how many shots every units can take from every other unit. If anyone feels like doing that, be my guest.
This is gonna be a much more simplified, reductive list. I'm not accounting for damage mitigation from armor, regen, etc. nor am I accounting for different bonus damage from different attacks. ALL I'm doing is looking at raw health totals vs. cost.
For cost, obviously minerals and gas are valuable for different reasons and open up different possibilities, so its not easy to simply say x amount of gas is worth x number of minerals. That said, we can do a very rough comparison by looking at how much gas a saturated base will harvest vs how many minerals a saturated base will harvest over the same period of time (for P and Z--MULEs screw this up of course). A fully saturated base will harvest 816 minerals and 228 gas per game minute. This means that gas will be roughly 3.58 times as rare as minerals. For purposes of making very broad generalizations, I'm therefore going to treat any gas cost as being equivalent to 3.58 that amount in minerals. For example, Stalkers are 125 minerals 50 gas, so 50*3.58 = 179, so total cost for these purposes would be 125+179, or 304. This is of course super crude, and if anyone has any better suggestions for making a broad generalization about how many minerals gas is "worth", I'm all ears.
Okay, with all the caveats/explanations out of the way, here's the list of health per cost. Numbers are just total health (counting shields as health for Protoss), divided by total cost using the gas to minerals equation outlined above.
Protoss units
Zealot 1.6
Warp Prism 1
Probe .8
Stalker .52
Immortal .49
Mothership .38
Carrier .36 (not counting additional interceptor cost--Interceptor health per cost is a crazy high 3.2)
Phoenix .35
Colossus .34
Void Ray .32
Archon .31 (cost is variable, but surprisingly any way you do it comes out to about .31 health per cost)
Dark Templar .21
Observer .2
Sentry .19
High Templar .13
Zerg
Overlord 2
Zergling 1.4
Queen 1.16
Roach .88
Drone .8
Overseer .6
Utralisk .49
Nydus Worm .44
Corruptor .39
Hydralisk .29
Mutalisk .26
Baneling .21
Brood Lord .19
Infestor .14
Terran
Marine.9 (1.1 with combat shield)
SCV .9
Hellion .9
Marauder .66
Thor .39
Battlecruiser .37
Medivac .33
Viking .3
Tank .27
Banshee .27
Reaper .21
Ghost .18
Raven .17
some interesting takeaways:
Among all combat units, Zealots have clearly the best hp for cost, and thats not even factoring in armor.
Corruptors have the best health for cost of any air unit in the game (except Interceptors, who are a niche case anyway, and Overseers who aren't really combat units). On the other hand, Brood Lord health for cost is awful--much more comparable to spellcasters rather than other capital ships.
Among the core armored tier 1.5 units (Stalkers, Marauders, Roaches), Stalkers have the worst durability for cost by a considerable margin, and thats not even including their wore armor (only half their hp is armored) or regen for Roaches or medivac support for Marauders. Considering they also have the worst dps for cost in the game, the fact that Stalkers are still such an integral unit is a testament to how powerful the combo of high range, high speed, warp-in and blink is.
Unsurprisingly, among T3 ground massive units, Colossi have pretty awful health for cost, and Ultras have the best.
Among drop tech, medivac health for cost is vastly worse than either Overlords or Prisms.
edit: some more numbers
If we’re instead looking at optimally saturated bases rather than fully saturated bases, gas is worth more like 3 times as much as minerals (3.12 if you’ve got 2 workers per mineral patch and 3 per gas, 2.9 if you put 4 on gas, I’m splitting the difference and just using 3). Then here are the numbers
Protoss
Zealot 1.6
Warp Prism 1
Probe .8
Stalker .58
Immortal .55
Mothership .44
Carrier .41
Phoenix .4
Void Ray .36
Colossus .39
Archon .36
Observer .24
Dark Templar .24
Sentry .23
High Templar .16
Zerg
Overlord 2
Zergling 1.4
Queen 1.16
Roach .97
Drone .8
Overseer .67
Utralisk .56
Nydus Worm .5
Corruptor .44
Hydralisk .32
Mutalisk .3
Baneling .24
Brood Lord .21
Infestor .16
Terran
Marine.9 (1.1 with combat shield)
SCV .9
Hellion .9
Marauder .71
Thor .44
Battlecruiser .42
Medivac .38
Viking .33
Tank .3
Banshee .31
Reaper .25
Ghost .2
Raven .2
Now, some people have requested I use a 2:1 ratio for gas value to minerals. I think in many situations this undervalues gas, but I could see how in some situations this might be closer to the truth, so here are the numbers run at a 2:1 ratio
Protoss
Zealot 1.6
Warp Prism 1
Probe .8
Stalker .71
Immortal .67
Mothership .58
Carrier .53
Phoenix .51
Void Ray .45
Colossus .5
Archon .51 (variable, with DTs its more like .48)
Observer .34
Dark Templar .32
Sentry .32
High Templar .23
Zerg
Overlord 2
Zergling 1.4
Queen 1.16
Roach 1.16
Drone .8
Overseer .8
Utralisk .71
Nydus Worm .67
Corruptor .57
Hydralisk .53
Mutalisk .4
Baneling .3
Brood Lord .28
Infestor .23
Terran
Marine.9 (1.1 with combat shield)
SCV .9
Hellion .9
Marauder .83
Thor .57
Battlecruiser .55
Medivac .5
Viking .42
Tank .4
Banshee .4
Reaper .33
Raven .28
Ghost .25
|
Interesting numbers, I feel that most of them get overshadowed by things like range, armor type, speed, etc. too bad there isn't an easy way to factory those things in, at least I can't think of one. The Brood Lord health for cost does seem like it has some real world applications, I do feel like they can get sniped rather easily compared to Battlecruisers or Carriers, although we don't see those as much as we do Brood Lords, so it's a little harder to tell. Cool numbers nonetheless.
Go zealots I guess?? 
|
what's the point? I mean can we make up a new strategy or use a certain unit in a tactical maneuver now?
|
what's the point? I mean can we make up a new strategy or use a certain unit in a tactical maneuver now?
whats the point of any SC2 post outside the strat forum? to discuss aspects of SC2 we find interesting.
|
Interesting numbers but I don't see any corollary between this Heath/cost ratio and usefulness of the units. In fact I think many people would argue that the stalker is better than both the roach and the marauder due to its blink and, range, ability to shoot up and general mobility.
And it is pretty clear that in the current metagame BLs are far superior to the other capitol ships since the others are seen as suicide while refraining from Brood Lord play is also tantamount to Zergy masochism.
|
Good news everyone, I've made up an equation unrelated to the OP.
Cost per health per damage per second per armor (is that the way you'd say this: )
((Cost/health)/(maximum upgradable damage per second/maximum upgradable armor)) = My random just made up calculation for unit combat worth.
I guess this would work differently for casters and AoE units.
ps, cool list. I like to know this kind of stuff.
|
I dont like the way you did the gas vs mineral cost.
Is ridiculous high. You should disregard saturation per base, because while u can harvest 3.58 times more mineral you also have to use 3 times more harvesters to do that. Is not a question of rarity, and more a question of investiment u got make to get the resources. Extractors needs to be contructed and harverters take less gas per trip.
Almost everyone does comparison of cost using a 1gas=2minerals or 1gas=1.5minerals. Is way more conservative and give us way more information for analysis. Doing the way u did, only tells that minerals units alway have more health per cost.
|
Interesting facts. These numbers show how good a unit can tank damage.
But the gas mineral ratio is just ridiculous. I feel like it falsify the numbers. That's why Sentry .19 << Zealot 1.6. This is way too extreme.
I think 1:1 ratio would be better. Sure you don't get as much gas as minerals but why sdould that influence these numbers? The collecting rate per minute and probe is the same.
|
Interesting numbers but I don't see any corollary between this Heath/cost ratio and usefulness of the units.
Certainly the fact that a unit has poor health to cost ratio doesn't make it a bad unit. For example most spellcasters have terrible health to cost, but no one would claim that sentries or infestors are bad units because of that.
But I do think there are some interesting takeaways here. For example...
despite their reputation for flimsiness even post patch, Warp Prism health for cost is actually pretty damn good.
The "imba" nature of marine range and dps also has to consider that without medivac support, their health for cost is easily the worst of any T1 unit even with combat shield.
Overlords have nearly 3 times the health for cost of Overseers (of course overseers have armor and more speed, but still, thats a pretty big gap), which means that if you don't need detection, saccing multiple Overlords for scouting may well be more cost effective than losing an Overseer.
|
Almost everyone does comparison of cost using a 1gas=2minerals. Is way more conservative and give us way more information for analysis.
That seems off to me. Zerg, for example, tend to be hugely gas starved, whereas minerals don't mean very much to them. protoss are similar. Terran has more uses for minerals so theyre more valuable for them, but due to MULEs they can also harvest a higher proportion of minerals to gas.
|
3.58 is way too high. On a per harvester basis, gas is roughly equivalent to minerals. Mineral heavy units generally have less utility, but overall, a roughly equivalent power to gas heavy units.
|
Neat thread. I always like looking at numbers for different situations in the game.
Ignore the haters.
Though your means of getting the numbers could admittedly use some tweaking.
|
Warp Prism health for cost is actually pretty damn good
Compared to medivacs, which are goofed up because of the conversion rate.
Overlords have nearly 3 times the health for cost of Overseers (of course overseers have armor and more speed, but still, thats a pretty big gap), which means that if you're not hurting for detection, saccing multiple Overlords may well be more cost effective than just one Overseer. Again, because of the conversion rate.
Not saying what you did is bad, just that with some minor tweaking it could be a little better and more accurate. Obviously it's going to be different for all the races though.
|
I remain skeptical of a 2:1 gas to mineral ratio as I think it underestimates just how gas-starved both P and Z tend to be (and therefore how much value gas has), but if people are interested in seeing those numbers I'd be happy to run them.
|
On March 04 2012 14:42 GreenManalishi wrote:
3.58 is way too high. On a per harvester basis, gas is roughly equivalent to minerals. Mineral heavy units generally have less utility, but overall, a roughly equivalent power to gas heavy units.
On a saturated base basis, 3.58 seems pretty accurate.
|
I'd say this is a bit skewed, because DPS/cost is much more useful considering that is what deals damage, we see here that zealots are considerably better tanks than stalkers, but this info is skewed. The fact is that since there are so many types of damage take, i.e. + to light or +armored, its not helpful at all to show this data alone with nothing to accompany it like how many hits of what kill it vs the other tiers. My point is that DPS is a standalone kind of thing where as damage must be taken as a number in context of what is dealing the damage to the unit in question.
|
I think basing the value of gas on a max saturated base isn't accurate. Take for example a zerg player on 4+ bases. You think all of those bases are fully saturated? Furthermore, how often are bases fully saturated (3 workers per patch) compared to optimally saturated (2 workers per patch)?
The flaw in this 3.58 figure should be extremely obvious when you get a unit like the roach (145hp) with cost 75/25 (3:1 min-gas ratio) being only .88, and the zealot (150hp) being 1.6....when for 25 more minerals the zealot has 5 more HP (shield + HP). I can understand gas being a more valuable resource, but to this extent is obviously wrong.
Or is the roach really that much lower/worse health for the cost than a unit like the zealot.
|
Gas is worth a little less than in your model in most games because when you aren't saturated, gas income will stay the same (assuming you want the gas, so you mine it) while mineral income will suffer. You can take an expansion just for the two gas with only 6 harvesters but you can't get a fully mineral-mining base as easily.
|
My point is that DPS is a standalone kind of thing where as damage must be taken as a number in context of what is dealing the damage to the unit in question.
DPS isn't standalone at all. It is hugely dependant on not only bonus damage (fairly easy to account for), but also the targetted unit's armor value. Marine dps, for example, goes waaay down when facing decent armor values, whereas Tanks are barely effected.
I'm not saying that what I've done isn't reductive, it clearly is and I freely admit it, I'm just saying that dps/cost charts tend to be equally reductive. They have to be because the nature of these things is that they're hugely variable based on what you're fighting.
|
Alright, I'm gonna head to bed, but tomorrow if people are interested I'll run the numbers again at a 2:1 ratio for value of gas to minerals. I think thats underestimating gas value, but perhaps the two sets of data, one for saturated bases and one a more generalized look, will together help paint a clearer picture.
edit: I will not however, that bases which are optimally saturated rather than completely saturated (only 2 workers per mineral patch, 3 on gas), you till get a ratio with gas being 3.12 times as valuable as minerals. Even with 2 workers per patch on minerals and 4 on gas, the ratio is about 2.9. Basically, once you take a new base, there may be a brief window in which you're harvesting, say, only twice as many minerals as gas, but thats going to fairly quickly going to change so that you're harvesting about 3 times as many minerals as gas, and potentially more. A 3:1 gas to mineral value ratio seems much more realistic than 2:1.
|
|
|
|
|
|
EPT
