If you use UnitTester, make two stalkers, select both and move them towards each other and press S or H to stop, the first shots by both will hit at the exact same times, but there are several times when one will win over the other even with 1-2 unit distance between them.
Distance between them: 0 stalker width Amount of tests both died: 6 Amount of tests one died: 4
Distance between them: 1 stalker width Amount of tests both died: 8 Amount of tests one died: 2
Distance between them: 2 stalker width Amount of tests both died: 8 Amount of tests one died: 2
On May 24 2010 21:19 spinesheath wrote: BW seemed to have something like this too though it probably wasn't intentional. Have you ever tried to kill all your SCVs as an obs? Even if your last 2 SCVs started attacking at the same time, one SCV would often randomly attack faster.
That being said, I don't like this randomness. There is no need for it. It doesn't make the game any better.
Oh and good job finding this.
With the unclear attack animation of the SCV, I think, that the SCV double kill only was this difficult due to poor unit pathing. If they did start attacking at the same tick, they should die both 10 out of 10 times.
I am pretty sure that I saw cases where both SCVs lost HP in the same frame (both wireframes changed in the same frame), but at the end of the battle the attacks were quite a bit out of sync. Some people (including me) also had the impression that they were almost never successful at killing both SCVs if the SCVs started attacking at the same time.
On May 24 2010 21:20 shalafi wrote: I couldn't find it in the editor. Which tab should I be looking at?
It's in the "Stats" tab of the weapons.
I can't take all the credit, Sickstee.nine found it, I just thought about posting this on TL.
As if it was hidden somewhere haha
BW has it too and that's very noticeable especially when you watch m&m force. The problem here's not the random delay itself, but blizzard's inconsistency in leaving this random shit in the game, that is good ONLY for aesthetics, and refusal to put actually important high-ground advantage, cause it has randomness in it.
I'm pretty sure this was not in broodwar. Muta stack? If they fired at different times it just wouldn't work.
This is stupid since it ruins the game. I'm not talking about the randomness... I'm talking about the fact you can't time your army'a attack (attack then move after attack, you have to wait until other units finishes attacking, sighs)
Yeah, it's there, so that all Units dont attack in sync. While somehow understandable, I think this delay is totally unacceptable for the early game units or melee units. (melee units usually dont attack in sync, because they have to move all the time).
On May 24 2010 21:53 Slunk wrote: This should be easy to test. Just make a map with 2 zerglings, wait for them to attack each other one or two times, save the game and reload it a bunch of times to see if the outcome is allways the same. (If it's even possible, I am not sure if SC" even has a save feature o0). I don't think this is actually true. I remember a lot of times seeing let's say reapers tossing they grenades allways at the exact same time without any guy missing the rhythm.
Just tested with reapers against 5 command centers, moving 10 reapers as a group with all within range, and then the first attack on the command center at the exact same time. On all 5 command centers there were some that went out of sync.
In other news, 11 reapers kill a command center in 5 volleys.
I guess you people hit it right on the spot. It's the inconsistency that is worrying
I guess I wouldn't be as annoyed about this if it was stated right from the start. Unit X doesn't do Y damage. It does Y-Z damage (because of attack speed randomness). That way you can sort of account for that in your play.
I always thought that a unit does the same damage all the time and always planned around that. I have a zergling and my enemy has a zergling? I'll try to hit his first and I should win! Seems I was wrong.
The reason I had this assumption was because of stuff Blizz has said - they don't want high ground advantage because it is random and they don't want randomness. But then there is randomness hidden all over the place.
I understand the need to vary the animations but the fact that also varies the outcome of the battle should be like clearly stated everywhere. I'd prefer there was none of this randomness (I think it would make for a better game IMHO) but if we have to live with it, at least make it clear as day so people can expect it.
I agree with the randomness. As long as its not huge its ok. Is there randomness in the amount of damage done? I believe not. What about how fast a unit moves? Nope. So the attack speed is a bit random. If there wasn't any randomness in the game every engagement should just insta-kill based on the algorithm they have for it. I've seen games who took this approach. It efficient though not engaging. I know that Y should always behave like this. Blizzard probably wanted to put an element of luck into the game without it being overpowered. Honestly wouldn't be surprised if this was in all there RTS's but very hidden. Sort of like in other games there are internal stats for your characters you never get to see.
On May 24 2010 22:20 Ceric wrote: I agree with the randomness. As long as its not huge its ok. Is there randomness in the amount of damage done? I believe not. What about how fast a unit moves? Nope. So the attack speed is a bit random. If there wasn't any randomness in the game every engagement should just insta-kill based on the algorithm they have for it. I've seen games who took this approach. It efficient though not engaging. I know that Y should always behave like this. Blizzard probably wanted to put an element of luck into the game without it being overpowered. Honestly wouldn't be surprised if this was in all there RTS's but very hidden. Sort of like in other games there are internal stats for your characters you never get to see.
Look at my previous post. In this situation it does ruin games.
I agree randomness can be a factor but not when it ruins the "game play". I can see how random damage output can affect the outcome, that's ok because the margin is small and why not have this luck factor in games. However this is about the random "attack delay", how can you know when to withdraw after attacking? You don't! you just have to guess. If there won't this random attack delay feature then how can you know when to withdraw after attacking? Through experience and playing a lot. Even having minimum and maximum damage would be a better solution if you want this randomness in RTS. This better be removed when sc2 comes out.
On May 24 2010 21:19 spinesheath wrote: BW seemed to have something like this too though it probably wasn't intentional. Have you ever tried to kill all your SCVs as an obs? Even if your last 2 SCVs started attacking at the same time, one SCV would often randomly attack faster.
That being said, I don't like this randomness. There is no need for it. It doesn't make the game any better.
Oh and good job finding this.
With the unclear attack animation of the SCV, I think, that the SCV double kill only was this difficult due to poor unit pathing. If they did start attacking at the same tick, they should die both 10 out of 10 times.
I am pretty sure that I saw cases where both SCVs lost HP in the same frame (both wireframes changed in the same frame), but at the end of the battle the attacks were quite a bit out of sync. Some people (including me) also had the impression that they were almost never successful at killing both SCVs if the SCVs started attacking at the same time.
This is actually the case, I am surprised.
I tested this using this savegame and I've had all sorts of outcomes. Just test this save a few times, the outcome is allways different. I have had each one survive and both of them explode.
Wow this topic was really helpful. I was wondering why I would lose when my zealot/probe attacked their zealot/probe first. Really, if your unit attacks first, it should win. In 1v1 situations, losing a zealot because their zealot just sat there while yours was actively attacking it is a big psychological blow. They definitely need to fix this.
On May 24 2010 21:49 HubertFelix wrote: So zergling fights are random?
Obviously they are. But mostly because of unit AI
On May 24 2010 21:21 iiomega wrote:
On May 24 2010 21:20 shalafi wrote: I couldn't find it in the editor. Which tab should I be looking at?
It's in the "Stats" tab of the weapons.
I can't take all the credit, Sickstee.nine found it, I just thought about posting this on TL.
As if it was hidden somewhere haha
BW has it too and that's very noticeable especially when you watch m&m force. The problem here's not the random delay itself, but blizzard's inconsistency in leaving this random shit in the game, that is good ONLY for aesthetics, and refusal to put actually important high-ground advantage, cause it has randomness in it.
I'm pretty sure this was not in broodwar. Muta stack? If they fired at different times it just wouldn't work.
Muta micro works because this delay doesn't affect first shot
The randomness is potentially for gameplay reasons; there's a possibility that it's there to prevent large amounts of overkill.
Imagine something like 10 zealots all attacking a thor who gets down to 5 hp. If they were in sync then all 10 would attack at once, killing the thor and and doing 9 overkill attacks. If the zealots are ever so slightly staggered then only 1 zealot will kill the thor and the other 9 can turn on other units.
Even if the other 9 zealots wouldn't overkill in a synced attack it still can help to have the staggered attacks. Every second a zealot's attack cooldown is up, but he's not attacking, is wasted dps. In synced attacks, 9 of the zealots would be wasting DPS once the thor dies as they go to find new targets. In an unsynced setup it's likely that most of the zealots will be able to retarget and attack while losing less dps because they'll retarget while their attack is on cooldown.
Yeah it sucks for low food fights, but in large food fights it might actually be necessary for units to operate a little more efficiently in large battles.
I be of a mind set to have a first strike advantage. It would work like this.
The first hit of a target who is not in combat would have something like a 1.25. The in combat is there so you can't just attack stop attack stop for the bonus.
So your units first hit would pack a little more umph and put the enemy unit in combat for say 3 seconds (or about an average of 2 attacks worth of time.)
I would also consider you to be in combat if the enemy unit is in the process of actively engaging you (A zealot actively going to attack a seige tank for example, the siege will obviously hit first but the Zealot would get the advantage because it engaged first in this example.)
There some different things they could do. The other road is random damage and synchronized swimming looking armies. That would probably bother everyone very quickly.
Take the different races. I expect my Thor team to strike almost as one. Why because they are highly trained individuals. Same with Battleships.
I expect my Zerglings or Ultralisks on the other hand to hit almost whenever because they are not highly trained to be organized. They're made to get the job done however possible.
On May 24 2010 23:25 Logo wrote: The randomness is potentially for gameplay reasons; there's a possibility that it's there to prevent large amounts of overkill.
Imagine something like 10 zealots all attacking a thor who gets down to 5 hp. If they were in sync then all 10 would attack at once, killing the thor and and doing 9 overkill attacks. If the zealots are ever so slightly staggered then only 1 zealot will kill the thor and the other 9 can turn on other units.
Even if the other 9 zealots wouldn't overkill in a synced attack it still can help to have the staggered attacks. Every second a zealot's attack cooldown is up, but he's not attacking, is wasted dps. In synced attacks, 9 of the zealots would be wasting DPS once the thor dies as they go to find new targets. In an unsynced setup it's likely that most of the zealots will be able to retarget and attack while losing less dps because they'll retarget while their attack is on cooldown.
Yeah it sucks for low food fights, but in large food fights it might actually be necessary for units to operate a little more efficiently in large battles.
This would be a really nifty way to prevent overkilling. It seems plausible.
I went through all attacks, and the only non-standard (-0.0625 to 0.125) delays are these:
Battlecruiser: min -0.0625, max 0.1875 Interceptor: no randomness (Interceptor firing rate on Carrier has standard delays though) Reaper grenades: min 0.1, max 0.5 (yep - attack is always delayed at least 0.1)