|
On April 12 2011 14:29 nodule wrote:Show nested quote +On April 12 2011 13:22 Rabiator wrote:On April 12 2011 13:07 Spammish wrote:
Everything should be fine now. ... and still it doesnt matter, because you need to manage your units on the battlefield and can waste the awesome upgrades. Also offensively upgrading big hits is only important if you need less hits to kill your opponents stuff with that (Thor needs 3 shots of 30*2 for Zealots and Stalkers no matter what) and could get a lot of overkill. Percentages simply dont matter in SC2, only the real values. Sure they matter. It's useful to know which units scale well with upgrades and which ones don't, even if they don't tell you the whole story. There will always be important edge cases that this kind of analysis doesn't cover. Percentages as also useful because they give you a rough idea how powerful the upgrade is (e.g., you can assume your army fights like it is ~10% bigger if you have a 10% attack up). One interesting thing that this reminded me is that attack upgrades scale worse as they get higher and more expensive (less % increase in damage), while armour ups scale better the higher they are (more % damage reduction for each level).
A lot of the "top increases" are just +1 damage per attack upgrades which are equalized by any defensive upgrade. There are upgrades which increase the damage by more than 1 damage per shot (Marauder vs armored, Siege Tank, ...) and these units get better against targets with more upgrades and defensive upgrades only negate a fraction of the offensive upgrades. So we are back at "percentages dont matter".
|
On April 12 2011 14:31 village_idiot wrote:He has corruption, though.
He has corruption with or without attack upgrades though.
IMO corruptors should get stim.
|
On April 12 2011 15:37 Tektos wrote:Show nested quote +On April 12 2011 14:31 village_idiot wrote:On April 12 2011 12:07 Nevy wrote:
Corruptor dead last.
poor guy. He has corruption, though. He has corruption with or without attack upgrades though. IMO corruptors should get stim.
No need for balance discussions.
On Topic: Good Work OP
|
I don't think calculating the percentage increase while including the previous upgrade is meaningful. By definition, attack upgrades add 10% of the original damage, minimum 1, usually rounded to the nearest integer. They shouldn't be compared with a value that is already upgraded.
The first column of the chart makes sense, but the next two don't have any practical meaning. Upgrading from level 2 to level 3 is exactly as powerful (though more expensive) than upgrading to level 1, while this chart would let people think it isn't.
One interesting thing that this reminded me is that attack upgrades scale worse as they get higher and more expensive (less % increase in damage), while armour ups scale better the higher they are (more % damage reduction for each level). As Rabiator said, this is why percentages shouldn't be used. If you make these calculations, you conclude that armor is more powerful than weapons, or at least scale better. But it doesn't: 10 marines, tanks, zealots or whatever unit with 3 weapon upgrades are equal or better than the same group with 3 armor upgrades.
And of course, in practice, the choice between armor and weapons often depends on the unit you have and those that your opponents have.
|
I really thought that the battlecruiser would be higher on this. I guess it doesn't really show the full potential though with its firerate. Interesting stuff, nice work.
|
On April 12 2011 15:52 Telenil wrote:
I don't think calculating the percentage increase while including the previous upgrade is meaningful. By definition, attack upgrades add 10% of the original damage, minimum 1, usually rounded to the nearest integer. They shouldn't be compared with a value that is already upgraded.
The first column of the chart makes sense, but the next two don't have any practical meaning. Upgrading from level 2 to level 3 is exactly as powerful (though more expensive) than upgrading to level 1, while this chart would let people think it isn't.
It's meaningful in the same way that the first column is meaningful: it tells you the % decrease in the # of shots required to take down a target (modulo rounding). It is the increase in DPS you get by purchasing the upgrade. That is absolutely relevant.
Show nested quote +One interesting thing that this reminded me is that attack upgrades scale worse as they get higher and more expensive (less % increase in damage), while armour ups scale better the higher they are (more % damage reduction for each level). As Rabiator said, this is why percentages shouldn't be used. If you make these calculations, you conclude that armor is more powerful than weapons, or at least scale better. But it doesn't: 10 marines, tanks, zealots or whatever unit with 3 weapon upgrades are equal or better than the same group with 3 armor upgrades. And of course, in practice, the choice between armor and weapons often depends on the unit you have and those that your opponents have.
The fact that armour upgrades scale better than attack upgrades does not mean they are preferable to acquire in an in-game situation or "better" in some overall sense (nor did I say/imply that).
Armour does better when the opposing units have low base attack (obviously), and percentages are a ogod way to see this (with GS, armour ups provide -25%, -33%, -50% cumulative armour damage reduction when facing marines, for instances)
|
clearly all this time ive spent not making fully upgraded 3/3 queens has been wasted
|
This is kind of redundant without the armour of the enemy unit. For example, a zealot with +3 armour upgrade has a total of 4 armour. A zergling does, by default, 5 damage per hit, so against a fully upgraded zealot he will only do 1 damage per hit (5dmg - 4 armour). If you were then to get a +1 attack upgrade for your zergling he would be doing 2 damager per hit - a 100% increase in DPS. An extreme case, but armour really matter to units that have low damage attacks.
EDIT:
On April 12 2011 16:51 fire_brand wrote:
I really thought that the battlecruiser would be higher on this. I guess it doesn't really show the full potential though with its firerate. Interesting stuff, nice work.
The battlecruiser WOULD be higher on this with armour taken into account. It's not the fire rate that matters but the fact that a battlecruiser has such low damage attacks, and by the time a terran might be making battlecruisers it is likely that his opponent will have been getting upgrades. The most extreme example: a 0/0 battlecruiser (8 dam) vs a fully upgraded (6 armour) ultralisk will only do 2 damage per shot. With a +1 air upgrade the battlecruiser would do 50% more damage to the ultralisk.
The thing that's generally far more important are the so-called "critical upgrades" - those upgrades that dramatically change how many hits are required to kill an enemy unit. You say a +1 attack on a zealot increases its damage by 12.5%, but against a zerglings that zealot will be 50% more effective (assuming you are out-upgrading the zerg). Tanks without +1 attack may not 1shot zerglings; the same applies to collosus but with +2. You get the idea.
|
It's meaningful in the same way that the first column is meaningful: it tells you the % decrease in the # of shots required to take down a target (modulo rounding). It is the increase in DPS you get by purchasing the upgrade. That is absolutely relevant.
The fact that armour upgrades scale better than attack upgrades does not mean they are preferable to acquire in an in-game situation or "better" in some overall sense (nor did I say/imply that).
Armour does better when the opposing units have low base attack (obviously), and percentages are a ogod way to see this (with GS, armour ups provide -25%, -33%, -50% cumulative armour damage reduction when facing marines, for instances) DPS is not significant because most units kill each other in less than 10 attacks - battles are too short for the damage to be "averaged".
Let's phrase it this way: these cumulative percentages figures do not represent a real, in-game increase of the unit's efficiency. You personnally didn't imply that armor upgrades were better or anything of the sort, but presented that way, the figures do - which is why they they are irrelevant.
Take the immortal. The table indicates 10% more dps with a weapon upgrade, which is less than most other units. Now if you look at actual in-game battles, a weapon upgrade results in pure immortals killing zerglings 0% faster, ultralisks 12% faster... and stalkers and tanks 33% faster. Immortals are actually one of the units that benefits the most from 1 attack upgrade.
It is clearer to say than immortals kill tanks in 3 hits instead of 4 and ultralisks in 8 shots instead of 9 than to say they do 33% and 12% better against them.
The first column of the % may have some use - though it has to be considered carefully.. But a calculation which considers the relative increase in DPS with respect to a previously upgraded attack is practically meaningless, due to the huge difference between the calculated DPS value and the "real" benefit the unit gains from it. It is also misleading, since the upgrade doesn't get worse at each level - it has exactly the same effect, which is to add X damage per attack. This upgrade may or may not have a measurable effect depending on the unit you face, and this effect is not related to the DPS increase.
See what I mean?
|
I don't think these calculations are usefull, most of the time +X% will not mean anything you can actually use, however knowing stuff like +1 atk means zealots 2 shot zerglings and collosi roast like 10 of them in a single beam(+1) and immortals 2 shot roaches on +2 (IIRC).
Maybe someone will find this useful but most people would rather the "x less shots to kill y"
|
On April 12 2011 22:56 Telenil wrote:Show nested quote +It's meaningful in the same way that the first column is meaningful: it tells you the % decrease in the # of shots required to take down a target (modulo rounding). It is the increase in DPS you get by purchasing the upgrade. That is absolutely relevant.
The fact that armour upgrades scale better than attack upgrades does not mean they are preferable to acquire in an in-game situation or "better" in some overall sense (nor did I say/imply that).
Armour does better when the opposing units have low base attack (obviously), and percentages are a ogod way to see this (with GS, armour ups provide -25%, -33%, -50% cumulative armour damage reduction when facing marines, for instances) DPS is not significant because most units kill each other in less than 10 attacks - battles are too short for the damage to be "averaged". Let's phrase it this way: these cumulative percentages figures do not represent a real, in-game increase of the unit's efficiency. You personnally didn't imply that armor upgrades were better or anything of the sort, but presented that way, the figures do - which is why they they are irrelevant. Take the immortal. The table indicates 10% more dps with a weapon upgrade, which is less than most other units. Now if you look at actual in-game battles, a weapon upgrade results in pure immortals killing zerglings 0% faster, ultralisks 12% faster... and stalkers and tanks 33% faster. Immortals are actually one of the units that benefits the most from 1 attack upgrade. It is clearer to say than immortals kill tanks in 3 hits instead of 4 and ultralisks in 8 shots instead of 9 than to say they do 33% and 12% better against them. The first column of the % may have some use - though it has to be considered carefully.. But a calculation which considers the relative increase in DPS with respect to a previously upgraded attack is practically meaningless, due to the huge difference between the calculated DPS value and the "real" benefit the unit gains from it. It is also misleading, since the upgrade doesn't get worse at each level - it has exactly the same effect, which is to add X damage per attack. This upgrade may or may not have a measurable effect depending on the unit you face, and this effect is not related to the DPS increase. See what I mean?
I see what you mean, but it sounds like you are supposing only one unit type is attacking another unit and that the unit always starts at full health.
A typical Protoss army has Zealots, Sentries and Stalkers Against roaches with no upgrades they do 14, 5 and 13 damage. Here I am quite certain that % based calculations are more appropriate than looking at # of hits a units take to kill another unit.
What makes the %-based table in the OP quite useless is that it doesn't take into account armor.
Unit that benefit the most from weapon upgrades are the units which have their attack increased by more than one. These attack upgrades are not fully mitigated by the same level of armor upgrades. Typical units are: Roaches, Tanks, Immortals, Thors.
Let's compare the Stalker and Roach
at +0/+0
Stalker damage vs Roach: 13
Roach damage vs Stalker: 15
and at +3/+3
Stalker damage vs Roach: 13 (0% increase)
Roach damage vs Stalker: 18 (20% increase)
Actually these numbers lie a bit, because the stalker gets even worse since the armor upgrade only applies to half it's health. (I am not saying the Stalker is a bad unit, my conclusion is that the attack/armor upgrades favours the Roach)
|
The biggest "vs Air" upgrade in the entire Zerg army is 11%
That is a tie between the Mutalisk and the Queen. Neither of which is particularly renowned for its anti-air capabilities.
God, Zerg Anti-air sucks.
|
I see what you mean, but it sounds like you are supposing only one unit type is attacking another unit and that the unit always starts at full health. Yes, but the % table does not become more accurate if you don't. How would you quantify a stalker+immortal vs roaches situation? It would technically come down to target selection, overkill and positionning, not to raw DPS.
Note that if you want to keep the DPS approach, the table would be much more significant if it showed the absolute increase in the DPS, in addition to the relative. The absolute variation is not affected by your target's armor, and at the very least it will put things in perspective. Example:
The biggest "vs Air" upgrade in the entire Zerg army is 11%
That is a tie between the Mutalisk and the Queen. Neither of which is particularly renowned for its anti-air capabilities.
God, Zerg Anti-air sucks. A queen may gain 11% more dps and the hydralisk by 8% more, 8% of 14.45 is greater than 11% of 9.
An hydralisk gains more dps from a ranged attack upgrade than a queen does.
|
On April 12 2011 12:07 Nevy wrote:
Corruptor dead last.
poor guy.
Along with stalkers (armoured) and vikings (armoured)
|
Good work OP. A useful guide. Apparently I need to upgrade my phoenixes more often.
On April 12 2011 23:41 Jermstuddog wrote:
God, Zerg Anti-air sucks.
I feel like hyrdas do really well against air (protoss air at least; I don't play terran enough to speak on it) even without upgrades.
|
On April 12 2011 23:08 Bulldozor wrote:
I don't think these calculations are usefull, most of the time +X% will not mean anything you can actually use, however knowing stuff like +1 atk means zealots 2 shot zerglings and collosi roast like 10 of them in a single beam(+1) and immortals 2 shot roaches on +2 (IIRC).
Maybe someone will find this useful but most people would rather the "x less shots to kill y"
There's another thread for "x less shots to kill y"
http://www.teamliquid.net/forum/viewmessage.php?topic_id=187603
This thread is useful in it's own way.
Also, +1 colossus does not 1 shot lings. +2 immortals do not 2 shot roaches. You're lucky this isn't the strategy forum.
|
On April 13 2011 00:19 Philip2110 wrote:Along with stalkers (armoured) and vikings (armoured)
Well atleast the corruptor can say he's even with them
so he isnt totally bad =[
|
you can't say upgrades arn't significant unless a unit can kill another unit in one less shot, because you are (hopefully) going to have different units mixed in your army. True, a stalker may still take 5 shots to kill a marine with +1 and a collosus still takes two shots with plus 1, but 1 stalker and 1 collosus together with + 1 can kill a a marine in 1 shot each, rather than needing an extra shot from one of them without the upgrade. you have to look at all the different combinations of units and upgrades to determine the true potential of the upgrade, rather than just the "x less shots to kill y", which is why i think the dps can be (somewhat) useful here since it is an average, and it is too hard to take into account all the different combos of units individually.
|
|
|
|
|
|
EPT
