EPTUnit Random Attack Delay - Page 3
| Forum Index > SC2 General |
|
Error Ash
Germany177 Posts
| ||
|
101TFP
420 Posts
On May 25 2010 00:01 Error Ash wrote: [...] this has no real effect on games. this is true. but what about replays? i thought that there is no randomness at all in the game, so that only the actual player commands would be saved in the replay file and not the outcome of them. which leads to the small filesize and rewinding capability of replays. if there is some kind of random effect involved, wouldn't that potentially alter the outcome of some situations in replays? | ||
|
MasterReY
Germany2708 Posts
On May 25 2010 02:38 101TFP wrote: this is true. but what about replays? i thought that there is no randomness at all in the game, so that only the actual player commands would be saved in the replay file and not the outcome of them. which leads to the small filesize and rewinding capability of replays. if there is some kind of random effect involved, wouldn't that potentially alter the outcome of some situations in replays? think about high ground miss chance in starcraft1. its like ~50% and if those results of every shot wasn't saved in sc1 replays, they would have a totally different outcome. Im pretty sure such things are also saved in replays. | ||
|
Triscuit
United States722 Posts
| ||
|
yoshi_yoshi
United States440 Posts
Regarding replays, I'd guess they have a random seed for each game to generate all the random values needed. Then they only need to store this seed in the replay file. Everything will turn out the same since each pseudorandom value is accessed in the same order. | ||
Excalibur_Z
United States12237 Posts
On May 25 2010 02:50 yoshi_yoshi wrote: Sweet, I'm going to hold a tournament to see which of my zerglings are the luckiest. Regarding replays, I'd guess they have a random seed for each game to generate all the random values needed. Then they only need to store this seed in the replay file. Everything will turn out the same since each pseudorandom value is accessed in the same order. Yep it has to store the random seed value. The attack delay phenomenon also appeared in BW and isn't statistically relevant, especially to the degree that BW's high ground mechanic was. | ||
|
CowGoMoo
United States428 Posts
On May 25 2010 00:01 Error Ash wrote: This was exactly the same in StarCraft 1, just watch a marine ball shooting at a building. The first shot will be perfectly simultanious from all marines, after 3 shots they all shoot at different times. Because the difference between these times is very very small and StarCraft units shoot relativley fast, this has no real effect on games. qft I think the randomness is just for aesthetics. It looks a lot better when 20 Marines stagger their attacks by a very small fraction of a second than if all their attacks are 100% in sync. The effects on gameplay are pretty much non-existent. | ||
|
glassmazarin
Sweden158 Posts
| ||
|
Spidinko
Slovakia1174 Posts
| ||
|
Lemure
189 Posts
| ||
Derby
Sweden31 Posts
On May 25 2010 02:38 101TFP wrote: but what about replays? i thought that there is no randomness at all in the game, so that only the actual player commands would be saved in the replay file and not the outcome of them. which leads to the small filesize and rewinding capability of replays. if there is some kind of random effect involved, wouldn't that potentially alter the outcome of some situations in replays? There is no true randomness. I'm no expert but as far as I know every "random" thing that happens in a game (and all other computer applications) is based on some kind of formula which can be based on a lot of factors. A replay mimics everything that has happened during a game and is dependent on the same formula and therefore every battle has the exact same outcome. As I said, I'm no expert and I'm sure someone can give a better explanation but I hope you understand the basics of it. | ||
|
Jonoman92
United States9104 Posts
In sc1 a unit who strikes first would never lose in a 1-on-1 battle. That is assuming the report in the OP is true and both zeals began with the same hp/upgrades. | ||
|
Roblin
Sweden948 Posts
both units have the same kind of randomness, its not like your unit will always lose, also, if you attack first its a ≈99% chance that your unit will win anyway, what the complaining people here are complaining about is that last 1% but honestly, lets face it, that unit will be 1 hit from dying anyway, i dont think it matters, plus, it evens out in the long run. | ||
|
Jyvblamo
Canada13788 Posts
On May 25 2010 03:55 Jonoman92 wrote: The comparisons people are making to sc1 aren't completely accurate. In sc1 a unit who strikes first would never lose in a 1-on-1 battle. That is assuming the report in the OP is true and both zeals began with the same hp/upgrades. *Not actually true | ||
|
dimfish
United States663 Posts
1) The randomness is in the attack delay which implies you can consistently count on your first attack to execute when in range. So when you are dancing around trying to get in the best position to engage there will be no randomness in when your units first fire when they are finally in range. 2) Even when you let a unit continuously fire and the delay between attacks is slightly random, there are many other factors like positioning that create random noise. If two players have the exact same number of zealots and engage with a-move, just the shape of the group and which zealots have to walk further to start attacking is going to drown out the random attack delay. A skilled player should not be frustrated by these minimal random factors because micromanagement of units improves their efficiency over an a-move blob and can win even if the random factors are stacking up against him. On May 24 2010 23:20 lolaloc wrote: ...and they defended their high ground mechanic because it's not random. I don't think that they are defending their high ground just because it's not random. They are saying the random chance to miss isn't as intuitive as "if you can see, you can shoot." So their goal isn't to eliminate all random factors from the game entirely (see point 2 above--if you wanted all random factors removed then you should be able to stack all units so their positioning isn't random, etc) Q: Are there any plans to change the current high ground advantage to the Wacraft III/StarCraft mechanic or other alternatives? A: No. We like the high ground rules and we think they are cool for StarCraft II. The random high ground from StarCraft just didn’t seem right for a such a skill based game. The clarity of “if you can see, you can shoot” makes a lot more sense to us. source Suppose in BW player A has enough dragoons to take a group of player B's tanks in an open space, but only just enough that the tanks would on paper be able to beat the goons from a high ground advantage. It's still possible for the goons to win against high-ground tanks if they "roll well" and land most of their shots. For a player in the game or an observer it may not be apparent what role the high ground played in the battle, or if A's goons were better micro'ed or what. Even though the skilled player usually wins in BW, it may not be clear if high ground is helping a lot or a little or what. Starcraft 2 has more intuitive high ground--if you can see, you can shoot. So players and observers can intuitively see the high ground mechanic's effect on a battle. I don't know if this means the new high ground is better, but I'm pretty sure this is Blizzard's argument. TL;DR: If you are skillful you will still win in Starcraft 2. | ||
|
guitarizt
United States1492 Posts
| ||
|
lololol
5198 Posts
On May 25 2010 03:49 Derby wrote: There is no true randomness. I'm no expert but as far as I know every "random" thing that happens in a game (and all other computer applications) is based on some kind of formula which can be based on a lot of factors. A replay mimics everything that has happened during a game and is dependent on the same formula and therefore every battle has the exact same outcome. As I said, I'm no expert and I'm sure someone can give a better explanation but I hope you understand the basics of it. As Excalibur_Z posted they store a number called a "random seed". Using the same random seed always produces the same sequence of numbers, so a replay will always match the events that occured in the game. Wikipedia Article | ||
|
iiomega
Romania94 Posts
 Here's a patch 13 replay that shows this: http://ul.to/jm2x75 Here's the fight recorded and uploaded to Ustream (excuse the choppiness, this is my very first attempt at recording a SC2 video): http://www.ustream.tv/recorded/7184942 In the process of uploading this to Youtube too. You can see that the zealots aren't upgraded in any way (they are both the first zealots me and the computer make). You can see that the Computer's zealot hits first (I think you can make out the shield numbers in the UI, again sorry for the low res). Having thought about it I agree it's not such a big deal, especially later on when there are more units fighting against each other. I'm not that bothered about it anymore. However, I think this is something players need to keep in mind, especially in low food fights (early game). It definitely wasn't something I was aware of before  I have edited the OP to include the replay and video. | ||
|
kli6891
United States143 Posts
| ||
|
Archerofaiur
United States4101 Posts
I think Blizzard is just trying to make the competive crowd angry at this point. Dustin: "Hey Bob what do you wanna do today to piss of the esports crowd?" Bob: "Eh i dont know nerf ultras maybe? Maybe make them choose between facebook and giving out privacy info?" Dustin: "How about we say that we cant do high ground cause its random and then include randomness in every unit battle anyway?" Bob: "Oh thats good..." | ||
| ||
