|
I just wrapped up another afternoon of research and blog posting, this time about Terran Wall-Ins and what kind of SCV numbers you need to keep a wall up if it's under attack.
If you're interested in seeing images of how to make walls on most of the current maps, check out the post on Terran Advantage called Terran Feng Shui.
The interesting finding was that each Terran building and unit has a set repair rate, and there's no specific correlation between that rate and the building or units hitpoints. For instance, a SCV repairs a Command Center for about 14 Health Per Second (HPS) on Faster, but a Planetary Fortress is repaired at a rate of closer to 9 HPS.
I did some testing against small groups of Zerglings/Zealots as well to figure out how many SCVs would be needed, in case they try and test your wall. The results make sense when compared with the Unit DPS chart that was posted recently.
SCVs repair Supply Depots (the weak link in a Terran Wall) at about a rate of 10 Health Per Second.
Zerglings do just shy of 10 DPS each on Faster, and Zealots do just under 20 DPS each on Faster. That means you need about 4-5 SCVs for every 6 Zerglings, and 1.5x to 2x SCVs for each Zealot.
Realistically you can get by with less because you'll probably have a few marines picking them off while they try to break through. Still, it's nice to know. 3-5 Zealots can show up reasonably early in a game and actually break the wall unless the Terran brings a significant amount of SCVs to keep it up.
There's more on the post if you're interested.
|
|
|
I went ahead and added a replay of CowGoMoo vs Nuke to the post that shows good wall placement on Kulas Ravine, along with what happens when the wall gets broken.
I found it over on sc2rc.com, direct link is here.
+ Show Spoiler +The remainder of the game is ridiculous and I wouldn't expect to pull that off unless you were CGM.
|
I really like that you did this research.
But when you calcule DPS you should use game seconds so it wouldn't matter which game speed it was tested on.
|
wow that's really surprising.. planetary fortress repair rate is less by 5 points/s? Why 5? PF has an armor of 3 compared to the command center's 1, right? Unless the increase by 2 armor reduces damage by 5 per second I don't see how that repair rate makes any sense^^;. Maybe blizzard actually hand-picked repair rates for each building to "balance" things? If a PF gets repaired at 9points/s imagine how overpowered it'd be at 14 =P
|
i think the repair just like in sc1 has a simple function that they use for anything that can be repaired
they simply look at the hp and the build time
id assume PF had more repair time because they take the CC time + PF time together because if they didnt the PF would be repaired quicker than a cc which wouldnt really make sense :p
|
That makes sense Morrow - it was weird to me while I was testing it because some of the things with similar costs would have largely different repair rates, and same goes for building times. It's just a weird combination I guess.
In any case, the numbers I posted on the blog are repair rates on "Faster" for some of the most common things you would be likely to repair.
|
So what's the repair rate of a MULE compared to an SCV?
|
|
|
This is an excellent write up. Good for other races too. As a Z player I know what Ling tight walls look like 
|
I don't fully understand the advantage of not allowing the enemy a single green high level space to attack your wall in from. Since you only need ONE unit to be at the BOTTOM of the ramp to gain sight to the upper level, it would seem that there is no reason to care how much space is up there, since there is no accuracy penalty now.
A better way to compare wall-ins would be to look at the total amount of surface area presented to the enemy. If 4-6 zealots can attack one of your supply depots at once, there is a good chance they could focus fire it down before you can get enough SCV's repairing. If only 2 zealots can attack, its no problem at all.
|
On March 16 2010 06:19 eNyoron wrote:
So what's the repair rate of a MULE compared to an SCV?
I was wondering about this so I tested a while ago, they are the same. 
|
On March 16 2010 06:42 Cyclon wrote:
I don't fully understand the advantage of not allowing the enemy a single green high level space to attack your wall in from. Since you only need ONE unit to be at the BOTTOM of the ramp to gain sight to the upper level, it would seem that there is no reason to care how much space is up there, since there is no accuracy penalty now.
A better way to compare wall-ins would be to look at the total amount of surface area presented to the enemy. If 4-6 zealots can attack one of your supply depots at once, there is a good chance they could focus fire it down before you can get enough SCV's repairing. If only 2 zealots can attack, its no problem at all.
It's also important to note that if you put your 2 supply depots together (on the common wall-in) you can put all your defending troops on that side (breaking the barracks won't happen since it has way more HP and repairs faster than depots) and it makes repairing the depots a bit easier since you don't have to split your SCVs to either side of the barracks.
I think doing the "alternate" wall (depot-depot-barracks) is better than depot-barracks-depot.
|
Terran Advantage has such amazing writing. It doesn't feel elitist or obscure, but it still packs in lots of insightful information.
I wonder why Bunkers aren't used as walls though. You can't lower or raise them, but they can make an enemy group think twice before attacking, and you can salvage them when you're done with them
|
On March 16 2010 06:42 Cyclon wrote:
I don't fully understand the advantage of not allowing the enemy a single green high level space to attack your wall in from. Since you only need ONE unit to be at the BOTTOM of the ramp to gain sight to the upper level, it would seem that there is no reason to care how much space is up there, since there is no accuracy penalty now.
A better way to compare wall-ins would be to look at the total amount of surface area presented to the enemy. If 4-6 zealots can attack one of your supply depots at once, there is a good chance they could focus fire it down before you can get enough SCV's repairing. If only 2 zealots can attack, its no problem at all.
I did some quick testing and found that filling the 1 green high level space effectively removes surface area from the equation by filling it with Barracks.
Here's the breakdown against Zerglings and Zealots -
Depot Rax Depot
2 Zealots can attack the top depot
2 Zealots can attack the bottom depot
3 Zerglings can attack the top depot
3 Zerglings can attack the bottom depot
Rax Depot Depot
3 Zeals can attack the top depot
2 Zeals can attack the bottom depot
4 Zerglings can attack the top depot
2 Zerglings can attack the bottom depot
The Rax Depot Depot setup (which leaves the open green space) allows an extra Zealot and Zergling to attack the top depot. It's also a little more difficult to repair the top depot for some odd reason.
In both setups, you can fit about 6 SCVs around each depot to repair it.
Filling that green space also lets them see a little less into your base, always a plus. Your SCVs may need to run around a little to repair if they switch up damage, but honestly they won't have the luxury if you have a few marines there. Your SCVs should easily outrepair 2 zealots or 3 zerglings on either depot.
On March 16 2010 07:53 Wolfpox wrote:
Terran Advantage has such amazing writing. It doesn't feel elitist or obscure, but it still packs in lots of insightful information.
I wonder why Bunkers aren't used as walls though. You can't lower or raise them, but they can make an enemy group think twice before attacking, and you can salvage them when you're done with them
Also thanks for the kind word! I'm glad it's proving to be useful to people. I think bunkers aren't used as walls because they take a bit to build and represent additional expense to successfully wall. In the case of using Supply Depots and a Barracks, you were going to build those anyway (and early in the game!)
|
On March 15 2010 23:59 MorroW wrote:
i think the repair just like in sc1 has a simple function that they use for anything that can be repaired
they simply look at the hp and the build time
id assume PF had more repair time because they take the CC time + PF time together because if they didnt the PF would be repaired quicker than a cc which wouldnt really make sense :p
To those that understand these terms, I think this would be a perfect example of where eigen values and eigen vectors would come into play. Just do a simple component analysis and see which sets of variables led to the output.
|
|
|
|
|
|
EPT
