|
Hey guys, i already posted this in the simple questions simple answers thread from the strategy section, but didn't get a definitive answer so i'll try here
So i was messing around in the LOTV unit test map from the arcade, and simply tried to kill a hydra with 5 phoenixes, one lifting up and the other 4 shooting at once. Hydralisks having no armor and 80hp, and the phoenixes dealing 10x2 damage to light units per shots, i expected the hydralisk to die. NOPE, it survived with 1 hp.
Now you're gonna tell me that the zerg regen let him survive, but simple testing like shooting once at the hydra (or any zerg unit) shows that it doesn't start to regen until around 2 seconds after first being damaged. So why would a full health hydralisk have time to regenerate 1hp if all the phoenixes shoot almost instantly? I've had the same result with every other zerg units that should have taken exact lethal damage, like 4 immortals vs a hydra, or 5 vikings with +2 vs viper (granted that roaches and mutalisk regen faster).
After further testing i found out that sometimes 4 immortals will actually one shot the hydra, but most of the time it will survive with 1hp... seems like theres some kind of rng in starcraft 2 after all -_-
The only constant thing is that the zerg unit gets shot by more than one unit but i don't see why that would matter. I figured it was still due somehow to the zerg regeneration mechanic, because this did not happen for protoss or terran units. Am i missing something very obvious?
TLDR : Zerg units that should take exact lethal damage seem to randomly (almost every time) survive with one hp, even though they should only start regenerating ~2 seconds after first being hit.
edit : it was pointed out that zergs units do regenerate instantly after being hit, but at a rate of 0.37hp/s, hence why it only displays it if the unit should die, rounding up 0.37hp to 1. Seems like there's still an inconsistency, as it will sometimes one shot the unit anyway Bottom line is, protoss players : you need 6 pheonixes to pick-up and one shot a hydralisk, not 5 :p
|
That's litterally the oldest thing in the book. It's what make drones regen several hp that make any attack dealing less than 40 damage a chance to regen. For instance, a SCV vs drone fight where the drone landed the first poke results in a dead SC2. Been this way for almost 6 years I think, and really I don't see the issue. It's not like regen is a big factor in any situation but muta harass.
|
Zerg units regen 1 hp the INSTANT they are hit. If all the damage lands at the same time unit insta dies if even a slight delay it lives with 1 hp
|
I've always thought that the first HP regenerated occurs nearly instantly. Either the first HP is always done nearly instantly (regeneration simply starts with 1 tick) or the 2 second regeneration cycle continues when at full HP, which means you might've started damaging a Zerg unit just as the cycle ended and regenerating 1 HP on the Zerg unit as the next cycle started.
In any case I've always made sure when targeting Drones in a situation where you end up dealing exactly enough damage that all the shots are fired at the same instant, to ensure that the Drone cannot regenerate the 1 HP.
|
On February 09 2016 03:50 JackONeill wrote: That's litterally the oldest thing in the book. It's what make drones regen several hp that make any attack dealing less than 40 damage a chance to regen. For instance, a SCV vs drone fight where the drone landed the first poke results in a dead SC2. Been this way for almost 6 years I think, and really I don't see the issue. It's not like regen is a big factor in any situation but muta harass.
I don't think you read this properly, how do you explain that 4 immortal vs hydra example?
|
Zerg units start regenerating as soon as they are damaged, so everything is working as intended.
I think you may have gotten the impression that it takes two seconds before it starts to regen. due to the fact that zerg units regen at 0.37 hp/s so it might not look like it's regenerating until after two seconds.
|
On February 09 2016 03:56 ZigguratOfUr wrote: Zerg units start regenerating as soon as they are damaged, so everything is working as intended.
I think you may have gotten the impression that it takes two seconds before it starts to regen. due to the fact that zerg units regen at 0.37 hp/s so it might not look like it's regenerating until after two seconds.
Alright this does make sense, but again why do the immortals sometimes kill the hydra in one shot then. With that 0.37hp/s regen it should never happen right? That a storm doesn't one shot is alright because it takes 4 seconds, giving the hydra time to regen, but the fact that you need an extra volley of shot wasted from your 4 phoenixes every time you pick up seems ridiculous
|
On February 09 2016 03:50 JackONeill wrote: For instance, a SCV vs drone fight where the drone landed the first poke results in a dead SC2..
THIS THING IS KILLING ESPORTS
|
On February 09 2016 03:58 ArtyK wrote:Show nested quote +On February 09 2016 03:56 ZigguratOfUr wrote: Zerg units start regenerating as soon as they are damaged, so everything is working as intended.
I think you may have gotten the impression that it takes two seconds before it starts to regen. due to the fact that zerg units regen at 0.37 hp/s so it might not look like it's regenerating until after two seconds. Alright this does make sense, but again why do the immortals sometimes kill the hydra in one shot then. With that 0.37hp/s regen it should never happen right? That a storm doesn't one shot is alright because it takes 4 seconds, giving the hydra time to regen, but the fact that you need an extra volley of shot wasted from your 4 phoenixes every time you pick up seems ridiculous
My guess is that you sometimes get really lucky and the shots actually are within the same cycle and kill the hydra before it regens.
|
So the regeneration only displays instantly if the unit should take lethal damage rounding up 0.37hp to 1? Because it obviously doesn't when the unit survives.
This seems a bit unfair although obviously specific to non-game ending scenarios :p That one hp can still lose you a phoenix next time you try to kill a hydra with only 4 TT
|
On February 09 2016 04:06 ArtyK wrote: So the regeneration only displays instantly if the unit should take lethal damage rounding up 0.37hp to 1? Because it obviously doesn't when the unit survives.
Well yeah it would look bad if a unit had 0 hp and survived. But in general I believe it rounds to the closest integer.
|
5/5 for comprehensive TL;DR.
|
The fuck is wrong with all those bug threads?
|
So as far as I know, the game uses "game ticks" of 1/16 seconds to handle events. (Maybe old game seconds, I don't know. Also, 1/16 seems a bit too much, but that's what I remember.) So my guess is if the 4 Immortals shoot all exactly in one game tick, then the Hydra dies, if not, it survives.
That raises the question, though, why would 4 Immortals, all in range, not fire all at the same time (within the same game tick)? I guess the answer could be either not all are facing the Hydra, so some have to rotate first, some not, which leads to a delay between attacks, or maybe some internal working mechanism of how time is sectioned into game ticks and how the computer exactly processes the actions. I'm no expert of this, though, so this second assumption may be totally wrong.
|
On February 09 2016 05:04 Sholip wrote:That raises the question, though, why would 4 Immortals, all in range, not fire all at the same time (within the same game tick)? I guess the answer could be either not all are facing the Hydra, so some have to rotate first, some not, which leads to a delay between attacks, or maybe some internal working mechanism of how time is sectioned into game ticks and how the computer exactly processes the actions. I'm no expert of this, though, so this second assumption may be totally wrong. Actually all units have a small amount of RNG timing on their shots.
|
On February 09 2016 03:36 ArtyK wrote:seems like theres some kind of rng in starcraft 2 after all -_- That does not necessarily mean there is RNG. It might be completely deterministic but way too complicated for a human being to understand. Its practically RNG because we can not grasp it buts its still technically deterministic because it adheres to well defined rules.
|
On February 09 2016 05:15 RoomOfMush wrote:Show nested quote +On February 09 2016 03:36 ArtyK wrote:seems like theres some kind of rng in starcraft 2 after all -_- That does not necessarily mean there is RNG. It might be completely deterministic but way too complicated for a human being to understand. Its practically RNG because we can not grasp it buts its still technically deterministic because it adheres to well defined rules.
With that argument, nothing a computer does is ever RNG, since computers never really generate random numbers. I don't really think that is a good point of view unless you are an IT guy discussing the problems of random number generation on PCs, because realistically the numbers the PC produces might as well be random for you in most other cases.
|
Not quite. Event though it is true that computers do not generate real random numbers there is still a difference between a very complicated but deterministic procedure and a (still deterministic) random number generator. The difference is that if you save the game and let it run, then load the save game and let that run, you will get the same outcome both times, if the game is truly deterministic. With a random number generator you might get a different outcome after loading the game. Unless, of course, we assume that the RNG is also reset when loading, which I doubt.
|
I think the error here is that you assume that the damage causes the regen to kick in, when it's more likely to be incremented according to the game timer? I doubt very much it's RNG.
|
It is indeed the case that all units by default have some 'random' delay in the timing of their attacks. In blizzard seconds: a unit's weapon can begin its firing phase up to 0.125 seconds late, or conversely 0.0625 seconds early.
If I recall, the reason for this programmed variance is to account for the travel time of missile-type attacks (such as the marauder's shells) that have to physically travel a distance before impacting and doing damage. A melee-range marauder shot actually will hit earlier than a range 6 shot.
Adding in this random delay to hitscan weapons (like the Immortal's) I suppose is in the interest of spreading around the variance.
|
|
|
|