|
On December 04 2010 06:32 DeBurd wrote:
Sorry, the /150 is the gas cost. It's a bit confusing because I don't mention gas anywhere else.
And yeah, I understand that the point of the OP is about cutting out depots for the mid-late game. I just see a lot of bad economics in this thread and others like it, and wanted to show the correct way to calculate.
Another good example that you see floating around is the obviously silly claim--but not obvious to disprove--that the cost of one zerg building is infinite. The logic says, "You sacrifice a drone to build the building, and that drone could have theoretically mined forever. Thus the cost is infinite." But the correct calculation is actually 50 minerals plus whatever one drone can mine in the time it takes to produce another drone, which is only about 12 minerals. 62 is a lot less than infinity.
(For other math nerds: the return on constantly producing workers is a function in sum notation. I can post an image of it if anyone is weird enough to care.)
OC's don't cost gas.
|
I'm defiantly going to test this out now, I was getting a little board with zerg anyhow. This will give me a reason to play with terran. I think there is problably a happy medium in here somewhere for the early game to get scvs, maybe enough for a base and a half or something and then have orbital's and still be safe. Although a Orbital wall in would be funny.
The other thing i've been waiting for a long time is to see a thor drop with mules repairing, could you imagine the power that would have with this build? Since I don't play terran I could be wrong but don't mules repair 2x as fast as scvs along with having the ability to stack on top of each other? Think of the nerf hammer this may cause...
|
On December 04 2010 06:32 DeBurd wrote:
(For other math nerds: the return on constantly producing workers is a function in sum notation. I can post an image of it if anyone is weird enough to care.)
I've actually been reviewing this in preparation for going back to school. Post please?
|
On December 04 2010 06:37 Ekko wrote:
Since I don't play terran I could be wrong but don't mules repair 2x as fast as scvs along with having the ability to stack on top of each other?
Mule's repair at the same rate as SCVs.
|
On December 04 2010 06:10 DeBurd wrote:
Sorry if this has already been said, I don't quite have time to read 16 pages of posts.
The idea is very cool and may be viable in any number of ways. But I have to make a note about the errors in the theory because I'm really sick of seeing incorrect economics applied to SC. Almost all the SC economics (especially re mules) that I see says, "If I spend x minerals now, and it results in >x minerals at some later point, then this is good!"
That is wrong. That assumes that your alternative is doing NOTHING with your x minerals. But doing nothing with your minerals is horrible. Just like a prospective investor calculates his profit over business interest, you have to calculate the return on your investment compared to the return of simply building workers steadily.
Example: An OC costs 550/150 to build and takes 155 seconds. Mules mine about 170 minerals per minute. Assuming one mule constantly (I know that's off by a few seconds), it's easy to do the math and say, "After 195 seconds, an OC will have paid itself off."
Wrong. After 195 seconds, an OC has paid itself off compared to doing NOTHING. But actually, if you had just steadily built workers, you would break even (including the cost of supply) only seconds later, and with greater ongoing income (8 SCVs > 1 mule). So an OC isn't per se a better investment than just building workers. It's just a better investment than doing nothing.
Obviously this doesn't account for island expos, saturation issues, 200-food encounters, etc. But that's why SC2 is a game and not a math problem.
This is only correct to the extent that you could have used those minerals to build more SCVs. Production capacity is always an issue as well. If you're a one or two base Terran you are limited in the number of scvs you can produce at any instant in time. In the early to mid game the choice isn't really going to be between the two forms of economic expansion as you'll likely be producing SCVs constantly in addition to producing the OC. You are balancing economic expansion vs military/tech expansion. Mule production is going to be in addition to your SCV production, not instead of it. The real power here seems to be the ability to expand economically without having to defend a larger footprint and of course the saved supply cost.
As you said, it is a game, not a math problem, but math can always help.
Just wanted to say that mules need a mineral patch to go to, and just like having more hatcheries or nexus, you invest in their potential. Again, there is no discussion here, other than, "Hey guys, Terran has a macro mechanic! Let's build more command centers!" It's the same as saying, "Hey guys Protoss can chrono-boost more probes, and with more nexus, have an incredible economy." Or, "Hey guys, Zerg can build more queens and hatcheries, and use spawn larvae to make more drones, and have an incredible economy!"
Discussion of unique benefits of the macro mechanic that haven't been fully explored is certainly discussion. Mules need a mineral patch, but unlike other workers they don't need a new mineral patch and they don't require supply. These are two benefits completely unique to the Terran macro mechanic. Discussion of how best to take advantage of a macro mechanic is certainly discussion suitable for a strategy forum.
The "but it's not the same!" arguments revolve around the basis that it actually is the same. For example, "Hey, that candy is just the same as that one!", "No, it's not the same!" repeat...
What?
"Terran macro mechanic isn't limited in this way - a Terran using their macro mechanic this way can have a minimal number of SCVs and a larger percentage of their supply in army." Almost as good as Zerg can!
Zerg can't have a larger percentage of their supply in army because they need drones for their economy. Are you suggesting that they could sack all their drones and then instantly rebuild after an attack? I don't understand your point.
|
It will be interesting when a terran BO optimizer comes out, if this turns out to be optimal or not
|
This thread just got real weird.
The really revolutionary part of this massing OC's idea is that "you don't have to build scvs!" which I'm pretty sure has been disproven. I've tried this sort of style many times now and I can authoritatively say you will NEVER have enough orbital commands that you will not need a good number of scvs. In fact you probably will never even have enough that you can justify cutting any scv production, although they can keep you afloat in the event that one of your bases gets popped and you lose a base's worth of workers.
However the idea that an orbital command is such a strong asset that it's worth building at most stages of the game is quite true. Just use it to take a base and mine from it like normal, and mule whichever of your bases you feel is the most exposed, and put fewer scv's on that base.
|
From just skimming through this thread, I'm amused about how the idea started out with "rush for mass OCs and get few SCVs and supply depots" and evolved into "it probably doesn't hurt to get an additional OC or two in the mid to late game" 
Anyway, props for a potentially revolutionary idea. I'm interested to see if a moderate approach actually finds its way into Terran's standard repertoire.
|
i actually went up against this the other day. the guy had a couple extra OC's in his base and i thought he was an idiot or something.
and then i saw the stream of marines....
|
On December 04 2010 06:35 Stoids wrote:
OC's don't cost gas.
Sorry, I'm dumb. I was thinking of PF. Obviously orbitals don't cost gas.
On December 04 2010 06:41 dahorns wrote:
This is only correct to the extent that you could have used those minerals to build more SCVs. ... The choice isn't really going to be between the two forms of economic expansion as you'll likely be producing SCVs constantly in addition to producing the OC.
This is totally correct. I'm not saying it's always wrong to build OCs, obviously that would be ridiculous. I'm just generally trying to encourage mule theorycrafters to calculate returns minus the cost of constantly producing workers.
On December 04 2010 06:37 richter wrote:
I've actually been reviewing this in preparation for going back to school. Post please?
Reveal the spoiler to see this. I don't want to clog the strat forum with math junk.
+ Show Spoiler +I'm at work so I can't upload the image but here's the idea. The return on constantly building workers for T seconds is everything they mine minus their total cost including supply. Equation below, with T for time and W for number of workers you've produced. Aka, W = T/17 rounded down to the nearest integer.
t
SUM [(i-17).66] - 62.5(w)
i=17
This has no perfect algebraic notation. However, if you're willing to substitute T/17 for W without rounding, you can get an algebraic equation (helpful for making graphs etc). Unfortunately it produces a funny effect where workers are paid for "gradually" while they build rather than in realistic chunks of 50. See equations below.
f(t) = (t-17)(t-18)(.66)/2 - 62.5(t/17)
f(t) =.33t^2 - 15.23t + 100.1
The numbers are rounded and so results will be off by a couple of minerals, but it's negligible. I can provide proofs for this and/or other tutoring help if that's what people want, but please PM me as this is no longer a SC discussion and I don't want to clog the forums.
|
that doesm't even mention the fact you can slap them anywhere on the map to really cause some chaos with siege tanks and an infantry ball. the splash is pretty lulzy.
|
On December 04 2010 06:32 DeBurd wrote:
Another good example that you see floating around is the obviously silly claim--but not obvious to disprove--that the cost of one zerg building is infinite. The logic says, "You sacrifice a drone to build the building, and that drone could have theoretically mined forever. Thus the cost is infinite." But the correct calculation is actually 50 minerals plus whatever one drone can mine in the time it takes to produce another drone, which is only about 12 minerals. 62 is a lot less than infinity.
You're ignoring the opportunity cost of making the building. The reason the cost is not infinite is not that you can ignore the opportunity cost, it's that you need to discount future income (also that the drone can't mine forever due to limited minerals, but for a first approximation you could ignore that).
|
its a great idea, but i feel like if scouted any good timing push would crush it..
|
On December 04 2010 06:32 DeBurd wrote:
Another good example that you see floating around is the obviously silly claim--but not obvious to disprove--that the cost of one zerg building is infinite. The logic says, "You sacrifice a drone to build the building, and that drone could have theoretically mined forever. Thus the cost is infinite." But the correct calculation is actually 50 minerals plus whatever one drone can mine in the time it takes to produce another drone, which is only about 12 minerals. 62 is a lot less than infinity.
What? While it may not be infinite, you would had been N+1 drones if you did not build a building and therefore the drone would continue to mine till the map was empty.
|
Can't you also get free supply depots from the OC's allowing you to gain another 100 minerals towards the initial 550 mineral investment? Talk about paying for itself...
|
The costs look very similar to those of SCVs when purely considering the supply and minerals but time is also important and the return on investment for a CC is far slower than that of an SCV. You also seem to have neglected the minerals an SCV could have mined if it weren't building a CC. An SCV pays for itself in something like a minute and 20 seconds after you purchase it whereas the orbital command takes far longer. If you consider the money based on time via typical interest calculations, it shows that orbitals commands are not nearly as profitable as SCVs and supply depots.
Naturally, your ability to build SCVs is tied directly to your number of CCs and supply-free mining is a pretty big deal when at the food cap so for those reasons it's a great idea to build more of them.
This seems like a logical way for terran to do a more econ oriented opening without the usual vulnerabilities of expansions, much the same way as zerg can devote all their production to drones but it doesn't seem like a replacement for constant SCV production or supply depots.
|
The OP has a lot of trouble confusing "per second" and "per minute." Could use an edit I think.
|
On December 04 2010 06:37 richter wrote:Show nested quote +On December 04 2010 06:32 DeBurd wrote:
(For other math nerds: the return on constantly producing workers is a function in sum notation. I can post an image of it if anyone is weird enough to care.) I've actually been reviewing this in preparation for going back to school. Post please?
maybe you should derive it yourself then 
edit:
This thread could REALLY REALLY use a FAQ section attached to the first post. Every 2-3 pages now just repeats what the previous 2-3 pages said, minus one or two posts.
|
What would be hilarious is floating several CC's to the enemy choke at midgame and PFing right outside their Natural.
I'm going to experiment with doing a Supply>CC>Enginering>PF opening as a means to block off your Ramp and beat back early pushes before going for the OC's.
|
On December 04 2010 09:56 Conrose wrote:
What would be hilarious is floating several CC's to the enemy choke at midgame and PFing right outside their Natural.
I'm going to experiment with doing a Supply>CC>Enginering>PF opening as a means to block off your Ramp and beat back early pushes before going for the OC's.
You can't convert an OC into a PF, so what you're proposing is actually completely different from what's being discussed in this thread.
|
|
|
|
|
|
EPT
