SCV Cost: 50 SCV Build time: 17s SCV Collection rate: 1 min / s SCV Saturation level: 22
Mule lifespan: 90s Mule collection rate: 3 min / s
==
First let's look at a few extremes:
If there are 6 SCVs, killing one means he will have 1 less SCV from now until saturation level. He will lose the full income of that SCV plus the cost of building one extra SCV.
This equals to 21*17 + 50 mineral lost (assuming the terran does not oversaturate to prepare for expansion). That's 407 minerals, more than a Mule could bring in (270)
If there are 22 SCVs mining, then it will only take 4 more SCVs for saturation. Killing one means it now takes 5 SCVs for saturation. That's 5*17+50 = 135 minerals. Killing a brand new mule is better. In fact, this is equal to killing a half-lifespan Mule.
Now obviously this isn't the kind of math we can do during harrasment, but notice that a Mule will always be somewhere between 0 and 100% it's lifespan - average 50%. If there are 22 SCVs, the short-term mineral gain (neglecting expansions) of an SCV is equal to what the Mule is likely to bring.
In this respect, unless you are planning to deliver a one-two kill against the Terran player, it is almost always better to kill an SCV than a Mule.
This does not even account for the fact that a Mule has 60 HP while an SCV has 45.
==
My math isn't excellent on this, and there's a lot I don't account for, but I just wanted to point out that in the long-term, killing an SCV is a better idea than killing a Mule.
yea but the thing is if you kill a mule 4 mins into the game with your reaper, then he will lose 270 minerals over the next 3 minutes, instead of 1min/second for the next 8 minutes
I thought SCV saturation level is about 16? I mean, 2 SCV's on a patch is a LITTLE BIT over-saturated, assuming one SCV starts mining immediately after the other is done, you'll soon see that after 2-3 trips, 1 SCV will have to wait for the other to finish mining.
Say you kill 1 SCV, he'd stand to lose 17(minerals per a new SCV build time) + 50 = 67 minerals. But if you kill a Mule, he'd lose 3 x 45(half of the Mule's lifespan) = 135 minerals. I thought the math is clear??
On April 25 2010 02:32 duckhunt wrote: yea but the thing is if you kill a mule 4 mins into the game with your reaper, then he will lose 270 minerals over the next 3 minutes, instead of 1min/second for the next 8 minutes
This is true, but if you're not looking for a victory in the next ~5 minutes, then you're better off killing the SCV.
On April 25 2010 02:35 Sky101 wrote: I thought SCV saturation level is about 16? I mean, 2 SCV's on a patch is a LITTLE BIT over-saturated, assuming one SCV starts mining immediately after the other is done, you'll soon see that after 2-3 trips, 1 SCV will have to wait for the other to finish mining.
It's 16 SCVs for minerals, and 6 more for gas = 22. The opponent is responsible for managing how many are on Gas vs Minerals before saturation, and this means there is a direct exchange rate of Minerals vs Gas. Assuming the opponent handles this optimally, the mineral cost will be the same, but there will just be an equivalent gas cost associated with those lost minerals (if he has to pull an SCV off gas to replace the minerals he's losing.
Say you kill 1 SCV, he'd stand to lose 17(minerals per a new SCV build time) + 50 = 67 minerals. But if you kill a Mule, he'd lose 3 x 45(half of the Mule's lifespan) = 135 minerals. I thought the math is clear??
If he is under saturation though, he will be pumping SCVs constantly - this means he cannot simply 'replace' that lost SCV and get back to where he was. He will be short 1 SCV from now until he stops pumping SCVs (saturation)
you should look at this relative to your economy. It can't be like thie for example, I can do the same exact analysis assuming infinite time, and I can say I should mass BCs instead of getting marines. but there exists relativity to the opponent. mules will be stronger to kill as it gives you a relative advantage for the next 90 seconds, assuming equal economy. also there exists saturation caps with scvs, so it ruins some of your numbers (1 min/s) as you increase your scv count. Unless the mule is about to die to time, always kill the mules first. The relative advantage is much greater.
The OP erroneousness calculates the build time for the additional four SCVs when trying to find a value for killing the one SCV. In reality, it's only 17+50, plus a touch more that can't easily be calculated, but not an additional 17*4.
However, it also doesn't include that the cost to replace an SCV can impede building another building or unit, although it generally will have a negligible effect (you're generally chaining scvs anyway).
On April 25 2010 02:44 Craton wrote: The OP erroneousness calculates the build time for the additional four SCVs when trying to find a value for killing the one SCV. In reality, it's only 17+50, plus a touch more that can't easily be calculated, but not an additional 17*4.
This is assuming the opponent's CC woudl have been idle instead of replacing that SCV. The only time that should happen is if the opponent is already saturated. If he is not, then he should optimally be pumping SCVs constantly, meaning he can't replace that SCV until he builds the next 4 SCVs he was going to build anyway - meaning he is 1 SCV short that whole time.
Also don't forget that by the time you can get a unit in to kill an SCV/Mule, he should have around 10+ mining minerals and minimum 3 mining gas, say, averagely 18 SCV's.
I still believe killing a Mule is way better. You're assuming everyone will saturate their mineral line, which is not true. Most people generally have 20 or less SCV's mining per base. Not only that, after 10+ scv's on minerals, each one after that doesn't have as much an important effect on your economy as, say, the original 8 would. Whereas the Mule, because it is oblivious to saturation, adds such a huge income boost to your economy.
Yes, in long term killing SCVs is better. But SC is not about long term, you don't try to cannon up and defend till you have 12 carriers and can tear enemy apart. You want to kill him at the first possible opportunity, that's why aiming at big but short-lasting advantagesd is more liked than things that will give you advantage over large amount of time. However if you are SURE that the game will drag - then it surely may be wiser to kill SCVs
Just kill the mule. Slowing him down 270 soon means his factory or whatever will be late etc etc. Its a big chain event of slowdown and 270 short term is strong early game. LOL! Cmon ppl... CMON!!!
This depends on stage of the game in my opinion. Killing a mule gives you not just the economic advantage but it also gives you psychological advance because enemy will go: OH MAN he sniped my mule RAGEE. And if you kill scv he will think: huuuh I've saved my mule... So besides that he wasted 50 energy in CC which he could have used on extra supply (~100 min) or on scan. Other thing if you walk inside 22SCVs +2 Mules i think you'll start killing anything (maybe mule for psychological AD) and if you can kill 22 SCVs you can kill 2 mules with them. But as you showed with your mathematics SCV kill is better on longer run.Thank you for the effort to show us the difference
It depends completly on the situation. If the scv will life for 3 more minutes it will harvest as much as a mule in its full lifetime. Also the scv costs 50 minerals itself. If you want to dmg him in the long run you go for the svc cause the mule come back anyway you just denie it for some time - If you kill the mule you denie a certain ammout of minerals while an svc can mine as long as there are free mineral patches. On the other hand if you see him spawning a mule and kill it instantly it can mess up his macro in the next minute if he cannot support his rax/fac usw.
If you kill a mule while it is returning minerals you stop him from getting even more minerals since they dissapear into thin air and he no longer has a chance to mine them ~
There is another benefit to killing the scvs. Mules cannot gather gas or build, so killing scvs further limits your opponents options, and their building ability.
Think it all depends on how much time the mule has left most times i just go for the SCVs on gas first really fucks the person up and if you are lucky he forgets to put some back on gas : )
I guess it just depends on how early it is in the game and how much harass you expect to get done. If you're just trying to whittle them down for a push some distance in the future, then SCVs are probably the best investment because of how it cripples their economy in the long-distance. If you're trying to shut down some sort of early-game Marauder push, then knocking out the mule may work best.
I tend to aim for SCVs if I actually get in to harass. Terran and Zerg share the problem in that their workers are lessened while making structures. Hampering their ability to gather AND build at the same time is usually the better way to go.
starcraft is a game of cutting corners and building up small advantages, one on top of another, and eventually coming out so far ahead that theres no other option but to win
When you kill a MULE, you're basically gambling that the MULE wasn't near the end of its lifespan. Its a similar choice to targeting miners or targeting the CC. The SCV will always be a safer choice because you know you hurt him at the very least 50 minerals whereas killing the mule has an unknown (though probably negative) effect on his economy.
Most harassment aims to kill multiple units, not just one unit. Thus, total damage done needs to be taken into account. At 60 hp, a mule is 33% harder to kill than an SCV, so targetting and killing the mule is equivalent to killing 1.333 SCV's. An SCV is worth 50 mins. 1.333 SCV's is worth 67 mins.
A mule averages 240-270 mins mined during its lifespan. Their expected loss from losing a mule is 120-135 mins. Thus, killing the mule instead of the SCV costs the Terran 53-68 mins relative mins.
That single SCV can mine 53-68 mins in 53-68 seconds at 1 min/second. 1 min/second mining rate only occurs pre-16 SCV saturation. SCV production time is 20 seconds(?? - I don't remember here - it may be 25). Thus, if he has <13 SCV's, then killing an SCV costs him more minerals than a mule. At 14+ SCV's, the mule most likely will cost him more minerals.
1.33 SCV's can mine 53-68 mins in 39-51 seconds. This makes the cutoff 14 SCV's.
The 14 SCV cutoff may rise higher depending on the relative min/sec mining speed of SCV's #17-21.
Considering the expected lifespan of a killed mule is 45 seconds, and the SCV mining times we're looking at are in the same ballpark, the immediate economic impact is similar. There is a slight argument here whether the 50 SCV min cost can be immediately deducted (he's chaining SCV's anyways, and won't be forced to spend that 50 until he saturates).
To summarize:
(1) How many SCV's does he have? Less than 14? Kill the SCV.
(2) Do you want to do more long-term econ damage, or short-term econ damage where short term is <60 seconds? Short-term? kill the mule. Long-term? Kill the SCV.
On April 25 2010 12:12 majesty.k)seRapH wrote: killing mules is just so so much better.
starcraft is a game of cutting corners and building up small advantages, one on top of another, and eventually coming out so far ahead that theres no other option but to win
On April 25 2010 12:07 GreggSauce wrote: uh... you should always kill mules, you should never try to kill an SCV, seriously
this isn't right at all
Either present an argument for why you think killing a MULE is better that refutes the OP or just don't post. You are wasting space.
Excellent thread btw OP, we need more of these contributions towards figuring the game out.
Question: If you have unit bars turned on, do you see the energy on enemy mules?
Ahh math. Finally a subject in which I have some semblance of authority to comment on.
When making mathematical analysis, you should shy away from making statements like, "it is better to kill of a scv/mule in the long run" because you are using YOUR definition of what is better, but that may not be the same definition to someone else.
The mathematical fact is that in the long run the T player will lose out on more minerals from killing the scv rather than the mule. This is because an scv can continue mining for the entire game, and therefore has no cap on the number of minerals he can mine(excluding mining out the map, which is hardly relevant).
However, you have to take into account the time value of money, just like interest. Minerals NOW are better than minerals in the future. If you had the option of getting 400 minerals at the beginning of the game versus 400 minerals 5 minutes into the game, which would you choose?
The same concept applies to mule's and scvs. While an scv will eventually mine more than the mule, the huge influx of minerals in a short period from the mule is weighted by the fact that those minerals are coming in extremely fast.
So instead of trying to compare the total difference in minerals mined, you would need to come up with an interest value of sorts that says, just how valuable are those extra minerals in the near future as compared to the total minerals he will get down the road.
On April 25 2010 12:12 majesty.k)seRapH wrote: killing mules is just so so much better.
starcraft is a game of cutting corners and building up small advantages, one on top of another, and eventually coming out so far ahead that theres no other option but to win
Math can only go to a certain extent, there are a lot of factors that come into play. Do you know for sure the amount of lifespan the Mule has left? I personally don't like rolling the dice and killing the mule unless i actually see the mule come down, for all i know he could have 5 seconds left before he blows up on his own.
Another factor: how big is your harrassment army? if you're just going in early with a reaper or two and you just want to make a quick hit and run, is it worth it to take that extra bit of time to kill the Mule (it has 15 more hp than an SVC)? I would rather just get a few quick SCV kills as opposed to a slightly longer Mule kill, especially if he hasnt fully saturated his mineral line quite yet.
i like this thread. one way you can gauge MULE life early game (when there aren't many orbitals yet) is by looking at the energy of their orbital; if it's really low the MULE just came down. not very practical at all, but just something cute to think about!
Mules return minerals as often as svcs. If a mule goes to the patch and returns money to the cc as often as an scv, then a mule mines 6 minerals per second.... 30 minerals per trip. Scv only 5 minerals per trip, 1 mule= 6 scv worth of mining...
depending the harass, if he is going to pull workers off mining including the mule then its best not to snipe it since its going to die in 90 seconds and re-summoned anyway.
it depends how much time you think you will have to kill shit. if you pop in and are just gonna get one kill before two marines snipe your reaper.. then kill that mule... if you pop-in with 3-4 reapers and he is short of marines then obviously leave the mule alone and kill as many as SCVs as possible.
Kill the MULE if its an early harass, jeez. The OC just went up so the MULE has a considerable amount of time on its lifespan. Closer to mid game hit the SCVs. There is no need to break down the targeting that much, focus on getting your macro up and not losing the reaper.
When it comes to the meta game killing the mule is the better choice.
The reaction of a player who sees his scvs die instead of mules is probably increased confidence since he'll see it as a noob mistake. -> Increased confidence-> playing better. And the opposite reaction when killing his mules.
There's situations where killing the scvs is the better choice economically but I don't think it's enough to make it an overall better option.
On April 25 2010 02:35 Sky101 wrote: I thought SCV saturation level is about 16? I mean, 2 SCV's on a patch is a LITTLE BIT over-saturated, assuming one SCV starts mining immediately after the other is done, you'll soon see that after 2-3 trips, 1 SCV will have to wait for the other to finish mining. I thought the math is clear??
full saturation is 3 scvs per mineral patch. 2 workers saturates 1 mining patch perfectly (they both mine at the same speed as they would mine separately). If u add more than 3 scvs per mineral patch, mineral income won't increase.
On April 25 2010 02:34 Chairman Ray wrote: hmm, good find. I will definitely start killing SCVs from now on.
Don't be so quick - Terran works on the fact that they have mules to help them 'get ahead'. While it may seem beneficial in the long run, cutting down their resource intake earlier in the game also affects them significantly down the line.
I think there is no absolute answers for this because it only depends on the situation. killing a fresh mule in the early game will delay the opponent build and timings, wich can give you room for breathing or pushing. Killing an early SCV will give you an advantage during the long game. It only really depends on two factors : - Do I need a economy advantage right now or later ? - How old is the mule, (you'll better check that before sniping).
Interesting thread. Petered seems to have said everything that can be said on it though. That kind of mineral value/time value is going to vary game by game and so never be calculate-able I suspect? I also love some of the dogmatic proclamations (MULE IS BETTER etc) spat in the face of good reasoning.
SCV Cost: 50 SCV Build time: 17s SCV Collection rate: 1 min / s SCV Saturation level: 22
Mule lifespan: 90s Mule collection rate: 3 min / s
These are wrong, SCV's do not bring in 1min/s, they bring in 1min/1.4seconds since real seconds are not ingame seconds.
This throws off your calculations by a lot btw, and equates to it practically always being better to target mules.
Irrelevant. Whether you play the game on fast, very fast, or slower, does not affect the SCV to mule relation. If SCVs take 1.4 x longer to build, bring back 1 min per 1.4 seconds, and mules last 1.4 seconds longer, and collect 3 mins per 1.4 seconds, the relation still doesnt change. In fact, if you wanted to, you could replace the s (which stands for second) by "time units", and it would work just as well. What matters, is not the actual time, but the relation to each other.
mules > extra supply, only time that should be used is if you supply cap yourself, or of banelings bust your supplied up and you need to continue making units ASAP... its a good way to help out with that too... instead of waiting forever to build a freaking supply again when you know another atks coming..
On April 27 2010 01:16 Xursian wrote: mules > extra supply, only time that should be used is if you supply cap yourself, or of banelings bust your supplied up and you need to continue making units ASAP... its a good way to help out with that too... instead of waiting forever to build a freaking supply again when you know another atks coming..
Read the OP - it's not that hard, and it's pretty darn clear that the issue is not to chose between mule and supply drop, but which to kill when vs T, the SVC or the mules...
On April 25 2010 10:44 xnub wrote: Think it all depends on how much time the mule has left most times i just go for the SCVs on gas first really fucks the person up and if you are lucky he forgets to put some back on gas : )
I agree to this, if I knew he recently called the mules down I'd def go for them.
killing mules is better for short term while scvs is better for long term. If you harass early and kill a mule, you can delay him from teching and making more scvs, which is a lot better than killing an scv which although will mine more long term, doesn't have a huge impact on the current economy.
nice OP, good reasoning and analysis some contributing thoughts... like you say, assuming your analysis is correct a good heuristic would be to kill scvs over mules, unless we think the mule was just called down, in which case it would be better to kill the mule. let's think of a scenario where you can take advantage of this knowledge. assuming the terran has called down mules as soon as their OC gets 50 energy, then at some point theres gonna be 2 mules out, one of them is just about to break down, and the other is basically new. so in that case killing both would be like killing a 50% mule or something like that, with twice the hp of a normal mule so even then killing both would be useless since you don't know which one is new and which one is old and it's more damaging for the terran if you target the SCVS. fortunately, since one of the mules are going to break down soon (assuming t has used OC at 50 energy each time, one of them will die in a relatively short period of time, i feel its like 5-10 seconds, maybe someone can verify this, but i know its really short) so in the meantime we can kill SCVS while we wait until the old mule reveals itself by breaking down and then of course you can target the one that is still alive. anyone see a problem with this
we might also want to consider that mules cost terrans 0 supply, so while killing scvs might generally be better economically it also frees terran up of supply which might detract a bit from the effect. this seems like a minor point though, and probably isn't gonna be that useful.
On April 25 2010 21:39 Diks wrote: I think there is no absolute answers for this because it only depends on the situation. killing a fresh mule in the early game will delay the opponent build and timings, wich can give you room for breathing or pushing. Killing an early SCV will give you an advantage during the long game. It only really depends on two factors : - Do I need a economy advantage right now or later ? - How old is the mule, (you'll better check that before sniping).
OP addressed those factors already, did you not read it all? anyways, ignoring that, you're incorrectly assuming that we're trying to find something that guarantees it being the best choice across all situations--but you're missing an important principle here. it's about finding a general rule (heuristic) that is more often than not going to be more useful following it than not following it, with exceptional cases. of course, the more we discuss and think about it the more likely we'll identify those exceptional cases (admittedly there doesn't seem to be many), in which case we can ignore following the heuristic to our advantage.
On April 27 2010 02:39 niteReloaded wrote: Go for the mule. It makes a bigger shock, as it disrupts the mineral/gas ratio and makes it harder to optimize the Build Order.
Your theory is good if computers play the game.
yeah, but how exactly does it make "a bigger shock". OP brought a fairly convincing analysis to the contrary, so if you're making a claim against it, i think you should provide evidence to support it. it seems like the burden of proof is on you now.
On April 25 2010 04:42 SuperJongMan wrote: Just kill the mule. Slowing him down 270 soon means his factory or whatever will be late etc etc. Its a big chain event of slowdown and 270 short term is strong early game. LOL! Cmon ppl... CMON!!!
On April 25 2010 12:12 majesty.k)seRapH wrote: killing mules is just so so much better.
starcraft is a game of cutting corners and building up small advantages, one on top of another, and eventually coming out so far ahead that theres no other option but to win
On April 25 2010 12:07 GreggSauce wrote: uh... you should always kill mules, you should never try to kill an SCV, seriously
this isn't right at all
Either present an argument for why you think killing a MULE is better that refutes the OP or just don't post. You are wasting space.
Excellent thread btw OP, we need more of these contributions towards figuring the game out.
Question: If you have unit bars turned on, do you see the energy on enemy mules?
this is an understatement. refuting someone coming here with statistics and math with "god, you're an idiot, just kill the mule" is so mindboggling. maybe you should take a step back, realize that you're not the best player in the world, and then wonder why.
the soul of this post is: what is the true value of an scv/drone/probe? and what is the true value of a mule? how can we figure this out?
the worth of a single worker is always worth more than 50 minerals, but for each race it's different. the value of a single worker also varies over the course of the game -- as the number of CCs and active mineral patches change...
On April 25 2010 02:35 Sky101 wrote: I thought SCV saturation level is about 16? I mean, 2 SCV's on a patch is a LITTLE BIT over-saturated, assuming one SCV starts mining immediately after the other is done, you'll soon see that after 2-3 trips, 1 SCV will have to wait for the other to finish mining.
Say you kill 1 SCV, he'd stand to lose 17(minerals per a new SCV build time) + 50 = 67 minerals. But if you kill a Mule, he'd lose 3 x 45(half of the Mule's lifespan) = 135 minerals. I thought the math is clear??
Very incorrect. 16 is definitely not saturated yet, not sure how it would even be considered OVER saturated. 3 SCVs per crystal will show benefits on the further away crystals, while having a small (almost negligible, yet still apparent) effect on mineral return rates.
Many are granting the MULE his full lifespan in the equation, but we know that the MULE has even less than 90 seconds left. You could be killing the mule 5 seconds before he goes back to whatever satellite he came from. For this reason, I spit my gas at the MULE.
but when you are doing a mineral line harassment by the time they have mules that's at less than half duration left their mineral line would've been pretty much fully saturated wouldn't they? And if this is an expansion hurass they can always SCV from their other CCs that's being idle o.o
Well the math is weird considering you're never sure how long the Mule has been there. Assuming it just spawned you're looking at
Mule 1 - 90 2 - 180 3 - 270
VS SCV 1 - 60 2 - 120 3 - 180 4 - 240 5 - 280
So killing the Mule puts the Terran progressively behind by about 3-5 seconds (depending on SCV income) each minute, to 9-15 seconds by the third minute where it caps off. It takes 5 minutes for the Terran to get behind 9-15 seconds by killing 1 SCV, but increases from there for as long as they constantly build SCVs.
In other words, if you're planning on attacking in the next 5 minutes go for the MULE, if not the SVC. Again, this math assumes the mule JUST came down and that the Terran player is steadily building SCVs. I would say 4/5 times it's better to just attack the SCV
Assuming your playing against a good player, that SCV will be replaced promptly. if the player is on his game, could even have that SCV out at the time one dies. So you'll only be setting him back the 50 mins to build a new one. If he's slow, then you cost him 50mins + the amount of minerals that would be lost while waiting for the new completion. The longer it takes your opponent to replace the miner, the more impact it will be against their economy.
Killing a MULE and preventing even 1 trip back will put stop 30 minerals from reaching the bank. if we want to use a formula to determine the amount of economical damage done by killing a single SCV vs a MULE we can use the following.
s = SVC m = MULE p = number of miners still intended to produce T = time elapsed (in seconds) between the death of the SCV and the birth of the new SCV t = the number of trips prevented by the MULE at 30 mins/trip D = damage done (number of minerals lost due to miner loss at any given time)
I read that the SCV collects at 1min/1.4 seconds (not sure if that is true or not, but we'll use that here). A 1min/1.4 seconds, that is 0.714 minerals/second. And I also read that the MULE moves 30 minerals per trip.
D(s) = p * (0.714 * T + 50) D(m) = 30 * t
So if you kill a MULE that will make X trips you are preventing 90 minerals from being banked.
Killing an SCV earlier in the game where your opponent intends to build say 5 more SCVs, will have that one replaced quickly, say the build time of one SCV (17 seconds).
D(s) = 5 * (0.714 * 17 + 50) = 310 minerals.
Assuming all 5 SCVs were queued up, that is 310 minerals lost over 85 seconds. If he already had as many miners as he felt fit, but planned on replacing the 1 SCV, but it took 30 seconds for him to get it out (including 17 second build time)
D(s) = 1 * (0.714 * 30 + 50) = 71 minerals.
So I'd say that early in the game you can do more lengthy damage by killing an SCV. But do more short term damage by killing the MULE.
Mid-late game, it will be a 50/50 choice, depending on how fresh the MULE is, and how much of an impact 1 SCV will have (assuming you target one on minerals and not gas).
On April 25 2010 21:24 ProoM wrote: full saturation is 3 scvs per mineral patch. 2 workers saturates 1 mining patch perfectly (they both mine at the same speed as they would mine separately). If u add more than 3 scvs per mineral patch, mineral income won't increase.
This is correct.
On April 27 2010 02:45 NiiPPLES wrote: Why the hell do people think 16 SCVs is saturated?
Diminishing returns do start around 16 SCVs but if you build 16 SCVs per base you're going to lose a lot of games very fast.
tl:dr people who aren't in the beta stop giving advice
Where did you get that information?
Number of Miners/Minerals collected per minute (based on mining at 1 location) 6/248 7/286 8/324 10/420 12/500 14/590 16/629 18/686 20/729 22/760 24/800 32/838 64/876
sorry for not stating why you should kill the mule. unless the mule is on its very last few returns it will always be better to kill it before anything else. the difference in hp doesnt matter. this is about killing a mule over the scv not the efficiency of killing either.
the big issue with this type of theorycrafting is how much mules can increase your resoure influx. sure killing scvs will hurt for longer assuming your opponent isnt over saturated and whatever other variables you can pull out.
but the truth is a mule racks in a lot of min, and those mins can be used to produce more cc's, depots, ocs, or scvs at a faster rate.
another way of looking at it is a mule is pretty close to a chrono boost on mineral intake. what if you could stop a building that had a chronoboost on it by attacking it? would you attack that one over the normal one? of course you would. it would be illogical to do otherwise, let alone bad play.
there is absolutely no circumstance where an scv should be valued higher then a mule.
keep in mind this is all theroretical obviously. in play if you reach the mineral line you need to kill as many resource collecting units as you can. and that usually means the one closest to your attacking units. if you can get a mule, awesome, but just kill the workers. any workers.
Your conclusion would be true if the only factor in a match-up was to kill an scv or a mule. However, as we all know, the objective is to win the game. By what means? Economic and strategic advantage. To understand economic/macro advantage you must learn the time-value of money. A dollar today is worth more than a dollar tomorrow. While the SCV will cost the Terran player its mining ability and 50 mineral cost, the MULE will cost the Terran player 270 Minerals in 1 minute as opposed to 407 minerals over the course of 15 minutes. Therefore, you are both right and wrong. Killing the SCV is better than killing the MULE if time didn't matter, however, not if you are trying to win the war.
Think about it this way. The MULE can bring in 270 minerals over 1 minute to build 5 scvs over the course of their build-time. This is certainly less than the time it would take to build 5 scvs if you only had 1 scv mining. Don't believe me? Map editor. 1 command centre starts with 1 mule the other starts with 1 scv. Both start with 0 minerals. Now start producing SCV as soon as you have 50 minerals and then go from there to see who gets more scvs faster. The gains of a MULE are even higher while in-game and not in map-editor.
On April 25 2010 02:32 duckhunt wrote: yea but the thing is if you kill a mule 4 mins into the game with your reaper, then he will lose 270 minerals over the next 3 minutes, instead of 1min/second for the next 8 minutes
The math seems wrong idk. Wouldnt the amount of minerals lost by killing a scv be 50+ (scv build time)(scv mins per second(which depends on saturation)) and maybeeeee add the amount of minerals you lose by not mining optimally for the 17 seconds you are behind in being saturated which is essentially negligible. Which idk according to your numbers is 67 minerals lost maybe another 17 that you lost by not mining optimally. And if the mule mines 270 mins in its lifetime you lose 135 on average by killing that.
OP math is so bad the post should have been deleted. lol at "This equals to 21*17 + 50" ... no , not even close, put down the nitrus. It's at max 67, and less so depending on saturation. Probably on average something like 65. A mule just about breaks even at that in 2 trips, and owns it and then some after 3. Chances are a mule will have more than 2 trips left in him well over 50% of the time. Side factors:
Killing an scv frees up farm space, an advantage which killing the mule does not give the terran In RTS (as in many things) a bird in the hand is worth two in the bush , meaning stopping him from having 200 minerals in the next 40 seconds is far more important than stopping him from having 300 minerals in the next 5 minutes... it gives you opportunity to gain advantage and leverage in that time that is worth far more than that petty 100 or so mins that you even then only arguably get.
Saturation on most bases is 22 for 8 mineral patches and they mine slightly less than 40 minerals a minute(using replay time) at saturation. Also, it makes a ton of difference whether the MULE has just been cast or it's duration is almost over.
On April 27 2010 02:39 niteReloaded wrote: Go for the mule. It makes a bigger shock, as it disrupts the mineral/gas ratio and makes it harder to optimize the Build Order.
Your theory is good if computers play the game.
yeah, but how exactly does it make "a bigger shock". OP brought a fairly convincing analysis to the contrary, so if you're making a claim against it, i think you should provide evidence to support it. it seems like the burden of proof is on you now.
Do you know what shock means?
I agree with him that for the long term killing the SCV is better. But for the short term killing the mule is ALOT better. It takes 270 minerals in a very short period of time, providing a type of shock to the opening build order.
On April 25 2010 04:42 SuperJongMan wrote: Just kill the mule. Slowing him down 270 soon means his factory or whatever will be late etc etc. Its a big chain event of slowdown and 270 short term is strong early game. LOL! Cmon ppl... CMON!!!
I'd have to agree with this, at least for earlier on in the game. It's not just about the minerals lost. If you make him lose minerals earlier, rather than more minerals later, it's going to create a lot bigger of a chain reaction later in the game in terms of when he gets his tech/expansions/army production going, and it'll most likely put you a lot more ahead than if you were to just do more damage to his minerals in the long term. I think it does mean that killing SCVs over MULEs later in the game is a lot more worth it though still.
On April 27 2010 02:39 niteReloaded wrote: Go for the mule. It makes a bigger shock, as it disrupts the mineral/gas ratio and makes it harder to optimize the Build Order.
Your theory is good if computers play the game.
yeah, but how exactly does it make "a bigger shock". OP brought a fairly convincing analysis to the contrary, so if you're making a claim against it, i think you should provide evidence to support it. it seems like the burden of proof is on you now.
Do you know what shock means?
I agree with him that for the long term killing the SCV is better. But for the short term killing the mule is ALOT better. It takes 270 minerals in a very short period of time, providing a type of shock to the opening build order.
Why is killing the scv better long term that doesnt make any sense. Your essentially setting the player back the 17 seconds it takes to make a new scv. It doesnt get compounded over the rest of the game or however long it takes to reach saturation. Unless if your strategy focuses around on doing some absurd 17 second timing window push killing 1 mule is always going to be better then killing the scv unless if you have some hacks telling you how much life is left on the mule. Also its not like 50% life is some magical number that blizzard devised to make killing a scv more optimal then killing a mule. Killing a scv and killing a mule are about equal in terms of damage when the mule has 24.4% life assuming you lose 67 minerals by killing a scv. Not used to thinking so im still unsure of whether you lose additional minerals for mining suboptimally for 17 more seconds so some numbers might be off a lil but I cant fathom it being more then 17 if any at all.
The mineral cost of losing an SCV is 60 minerals / minute until the SCV is replaced, at which point you lose a fixed 50 minerals. Note that this latter loss only occurs at the point that you otherwise wouldn't have built SCVs, not immediately.
If the opponent wouldn't be building SCVs anyway, killing an SCV costs them 50 minerals + 17s of lost mining time. Say 70 minerals.
If the opponent never rebuilds the SCV, they lose 60minerals/min until the end of the game. It becomes better to killing a full mule after 4.5minutes.
Note: building 16 SCVs takes 4.5 minutes as well, so it is unlikely that the SCV would be replaced much after that.
But the mule may not be new!. True, but the base is likely not at 0% saturation, either. Any way you slice it, the mule is almost certainly going to provide a greater short-term shock to the opponent's economy, and probably going to create a greater long-term shock as well.
Reason to bump is that this is still relevant. The only value I can add to this post is to mention a few things that I feel were not considered:
Maybe you should be killing in base orbitals instead of going for scvs or mules?
Also no one mentioned that the cost to upgrade the command center to create mules is 150 minerals.
16 scvs have about hit points 800 hit points an orbital command has 1500? But the the 16 scvs cost more then an orbital command.
In the late game a terran would have no scvs or more or less no scvs and only mules, attacking the orbital would force both mules and scvs to repair and stop mining.... There is also a cost associated to repairing.
If its generally better to go for an scv I would assume it would have to be early on based off the factors I brought up above.
Maybe the answer is something like this:
If terran opened double orbital then mules....
If early mid game scvs
If late game mules since killing scvs would actually help the terran.
Nobody would spend 50 energy in the early game to save an SCV that was about to die. A mule is way more valuable in the early game as it is necessary to hit expand timings, and general build timings. Obviously a single mule is less valuable over time than an scv, but the time in a game that this dilemma occurs is one where mineral income is so vital because of exponential economic growth.
On July 22 2013 14:51 furyofSkanks wrote: Reason to bump is that this is still relevant. The only value I can add to this post is to mention a few things that I feel were not considered:
Maybe you should be killing in base orbitals instead of going for scvs or mules?
Also no one mentioned that the cost to upgrade the command center to create mules is 150 minerals.
16 scvs have about hit points 800 hit points an orbital command has 1500? But the the 16 scvs cost more then an orbital command.
In the late game a terran would have no scvs or more or less no scvs and only mules, attacking the orbital would force both mules and scvs to repair and stop mining.... There is also a cost associated to repairing.
If its generally better to go for an scv I would assume it would have to be early on based off the factors I brought up above.
Maybe the answer is something like this:
If terran opened double orbital then mules....
If early mid game scvs
If late game mules since killing scvs would actually help the terran.
No not really. Accounting for the cost of an orbital upgrade is irrelevant because terrans get it regardless of what's happening. Also talking about killing in base orbitals is also mostly irrelevant, as if you have enough time to drop their base and kill an orbital command, then the game is probably over then and there, it's no longer harassment to damage their economy it's a game-ending move. Mining time and minerals lost to repair is not as damaging as killing a bunch of their workers, as if you have enough to try to focus down the CC, then you have enough to kill all the workers which is far more damaging than trying and failing to kill an orbital or planetary, especially in the late game where terrans will most likely have lots of orbitals so they just lift one up and land it at the base you just killed, and go right back to mining.
The only relevant thing to this discussion i can say from what I think is: 1. early game where every mineral really matters --> kill mules to fuck up their BO and initial building production 2. Mid game where you want to do a timing/allin --> kill mules to hurt their production in the lead up to this attack, or if you don't want to risk killing a mule that is almost expired, kill gas SCVs. Although honestly I would always kill mules in this situation. 3. Mid game where you want to take it to the long game --> kill SCVs to hurt the long run income and possibly gas income 4. Late game where terrans moves to relying mainly on mules --> i would kill SCVs here because at this time mules are so easily replaced, and only SCVs can mine gas/repair/build things
Long story short: Kill mules for short term economic datamage if they have more than 33% life, kill mule for long term damage if it has more than 55% life.
For calculation purposes, the max minerals gained will be used per minute for both the SCV and the MULE.
Mineral loss per building time An scv gets 45 minerals p/m. 45/60 = 0,75 minerals p/s. 0,75 * 17 = 12,75 minerals lost during building time.
Total damage for SCV's scv cost = 50. 62,75 minerals lost per scv, if saturation is 16 or less With 3 scv’s per mineral patch the loss is (12/60)*17 = 3,4. 3,4 + 50 = 53,4 minerals
Additional damage The big thing with killing an scv, is that you can state that all scv’s that havent been build yet lose 17 seconds of mining time as well. Lets say that the 16th scv is killed, so after recreating this, all 8 future scv’s lose 3,4 minerals mining time, so 3,4 * 8 = 27,2 minerals per minute. If an scv is killed before 16 scv’s, every extra scv costs an additional 12,75 of mining time. So if the 12’th scv is killed, on top of the calculations done before an additional 12,75 * 4 + 3,4 * 8 = 78,2 minerals are lost if a constant scv production is upheld. However, these minerals are lost over a much longer timespan.
So for short period damage: up and to 16 scv’s: ...62,75 After 16 csv’s: ......... 53,4 Long period damage: 12th scv dead: .........140,95 16th scv dead: ..........80,6
MULE Damage A mule gets minerals in batches of 30. It gets 180 minerals per ingame minute, or 270 minerals during 90 seconds. 270/30= 9. 100/9 = 11,11. So about each 11,11% of a mules time, 30 minerals are mined.
This means, that if a mule has about 50% of its time left, 50/11,11 = 4,5 = 4. 4*30 = 120 minerals are mined.
So what is the breaking point that it becomes more efficiënt to kill a worker than a mule? 55,55% 150 minerals 44,44% 120 minerals 33,33% 90 minerals 22,22% 60 minerals 11,11% 30 minerals
So to conclude Short term damage Kill mule if 33% time left or more if low saturation Kill mule if 22% time left or more if high saturation
Long term damage Kill mule if 55% time left or more if low saturation Kill mule if 33% time left or more if high saturation
i mostly hold an easy answer. IF you see the mule fall, kill the mule. IF you don't see the mule fall but you see it mining already, go for scv! (<- because for all you know, the mule only has 5sec lifetime left, so the damage you have done to it you could have killed an scv)
I always target mules. It's better to go for short term advantages in this game, and then leverage those short term advantages to get even further ahead. Yes, killing an scv does more economic damage than killing a mule if that's all you consider. But if you kill the mule, you could be creating a window of opportunity for yourself where your opponent's defenses are less developed than they could be, allowing future harassment to kill an additional five scvs (for example).
Also, if you have a single unit like a reaper or a mothership core targeting down SCVs, a good player can often just run the SCV taking damage away and end up saving it. If you force them to run the mule instead, they'll lose more income during that brief window where a single unit is running.
Yes, but now you know that according to the math, if a mule has less than 33% time, an scv will give more short term damage if the scv count is less or equal to 16.
You need to analyse the time value of money aspect to sc2. I would have no idea where to start generating an appropriate discount rate as to define it accurately you would have to know all of the possible 'paths' to winning the game (each mineral is contributing to the outcome) and then from this determine how many 'future' minerals are the equivilent to a mineral right now.
I suppose if you really wanted you could base assumptions around a set of standard builds and standard timings and map statistical probabilities of the win/loss outcomes. By doing this for each matchup against terran you would be able to generate an applicable discount rate for various simplistic outcomes. These generated discount rates for different build openings should then be compared with what lets call a zero net benefit rate (the rate at which it makes no difference whether you kill the scv or the mule). Through the comparison of the statistical discount rate (dependent on the build) and the zero net benefit rate, you could determine whether to kill the scv or the mule.
Note: It is entirely probable that upon observation of rates for a few statistical builds and the zero net benefit rate it could very quickly make obvious that killing one over the other will always be far superior - which one that is though i'm not sure (other variables such at the number of active workers and the number of command centers could also have a large impact upon the results)
With all that being said, i've always thought it best to kill scv's unless you've seen a freshly dropped mule - the average mule you come across will have half its life gone, and killing it really only costs a terran 135minerals.
How do you take into account when the kill comes? Obviously 270-300 minerals in a 90 second window at the beginning of the game is more important than other points in the game. You also have to take into consideration that an SCV can actually build stuff and attack so its usefulness goes beyond how many minerals it will eventually mine.
The thing is that the math will basically say that the longer the game goes on, the better it is to kill scvs early compared to MULEs, but remember that there are more tjings to consider, such as "what will he do with the minerals?"
The math is the same if you kill 3 scvs when he has 16 workers or 6, but in %, the difference is huge. So killing a mule will not according to the income be better over a long game, but remember that he will invest these minerals in either army tech or income, which means that killing a mule will: 1. Delay an expansion 2. Delay/weaken an incoming push 3. Delay/forcing him to skip tech
The first one is obvious, the second one will force more units and less income from you, and the third will do the same.
Thats why people love sniping the mules, because on paper, scvs are better, but mules will give you early economy, which will exponentially increase its value depending on investment
Really interesting thread. I was thinking about this myself a little while ago. It seems that the correct choice varies according to the situation. I think a lot of times though that killing SCVs would be better because when you think about it... if you 'kill' a mule you never really kill it, just put it out of commission for a while, depending on how fresh it is. To really KILL a mule, you'd have to kill an orbital... which is unlikely to happen.
If you kill a mule, it goes away for a little while, but comes back when the energy regens. When you kill an SCV, it's gone until the terran stops making SCVs. Usually terrans will continue constant SCV production until 70 or so and won't stop until that point unless doing some sort of timing attack or something. Say you kill one SCV when the terran has 25 SCVs. He now has 24 SCVs and is building another. Let's say it comes out 17 seconds later. After 17 seconds, he has 25 SCVs again, but in reality he's still missing a worker, because had you not killed that SCV, he would have 26 SCVs at this point, so really, an SCV is still dead and not collecting. This continues until the terran stops producing workers, which can be a LONG time from then.
EDIT: also, killing an SCV forces the terran to build a new one, costing 50 minerals, where killing a mule forces nothing. He will simply drop another when he has the energy. No extra cost to him.
Theres also the chance that his mule would have run out very shortly after you kill it. Also mules have more health than an scv, a banshee needs 3 shots for example.
Since i play protoss, to maximize oracle harrassment, i would go after the scvs instead of the mule unless i just saw it get called down and even then i would go for the scvs.
Explanation:
Oracle shows up in terrans mineral line, there like oh shit! got to run everything away. They take the mule and the scvs as well. In this instance you are already doing economic dmg since the mule is pulled off which is why i would go after scvs.
By the way to everyone who wants to know a good way to see if the MULE is fresh or not, look at the energy left on the OC, if it is close to zero than it is still a new MULE, but if it has more than 30 than probably you should kill SCVs instead.
I feel like even though you may not be planning on killing within the next 5 minutes, if you are killing one of the first mules to get called down, it will mess up his b/o timings a lot more than killing an SCV. It may also allow opportunities to do more damage with early pressure at a time when his defense would normally be solid, since his first couple units will be a lil late.
Perhaps we could include variable like mule frequenzy (number of orbitals available), because the energy gained on orbital is constant and assuming energy is only used on mules we could define some timings related to build orders where killing mules is particularly strong?
On July 24 2013 23:04 woreyour wrote: Does anyone feel like the terran's early mules are crucial than SCVs?
I mean the early mules could really mess up the build and timing while the SCVs are not really that big.
(we assuming the you are just going to kill one SCV or Mule since killing a lot would not be the thing we are discussing here.)
Then again, if you're killing his SCV you are probably messing up his build already.
And you forget to factor in that if you kill an SCV - if he expands early enough not to transfer 22 scv for full saturation - he will be missing that one worker every time he expands.
On July 24 2013 06:12 nickyboy909 wrote: Really interesting thread. I was thinking about this myself a little while ago. It seems that the correct choice varies according to the situation. I think a lot of times though that killing SCVs would be better because when you think about it... if you 'kill' a mule you never really kill it, just put it out of commission for a while, depending on how fresh it is. To really KILL a mule, you'd have to kill an orbital... which is unlikely to happen.
If you kill a mule, it goes away for a little while, but comes back when the energy regens. When you kill an SCV, it's gone until the terran stops making SCVs. Usually terrans will continue constant SCV production until 70 or so and won't stop until that point unless doing some sort of timing attack or something. Say you kill one SCV when the terran has 25 SCVs. He now has 24 SCVs and is building another. Let's say it comes out 17 seconds later. After 17 seconds, he has 25 SCVs again, but in reality he's still missing a worker, because had you not killed that SCV, he would have 26 SCVs at this point, so really, an SCV is still dead and not collecting. This continues until the terran stops producing workers, which can be a LONG time from then.
EDIT: also, killing an SCV forces the terran to build a new one, costing 50 minerals, where killing a mule forces nothing. He will simply drop another when he has the energy. No extra cost to him.
Opportunity cost. Since energy is no unlimited resource, he does have a cost.
Well if you are comitting to an all-in for instance it would be smarter to kill the mule as the SCV won't be able to pay for itself yet and you also have to consider that you can't actually see how much it has mined yet. So unless you see him call down a mule OR you have some kind of insane inner clock to know exactly how long it has been since the last 50 energy on the OC you really won't know if you've killed 270 minerals worth.
If you are preparing an all-in and happen to have a chance to kill either a fresh mule (that you've just seen plop down) and an SCV. The mule is the better choice, but I don't really think even Code S players take that into consideration.
I prefer killing MULEs, especially in the early game, since one MULE mines as fast as 4 SCV's, you effectively remove 4 SCV's from the Terran's economy for the remaining duration of the MULE.
As opposed to killing an SCV which technically hurts more, but it's over a longer period of time so it's not as profoundly felt, where on the other hand killing the MULE could mean that the Terran has to choose between building more marines or building more SCV's or building more production, which is psychologically taxing as well as if he misses an SCV or two due to this, then it hurts even more.
i still dont understand from the op wich one is better . talking in % like 60% of times is better to kill a scv 40% is better to kill a scv ... can some1 explain to me wich one is better overall ?
On August 01 2013 00:57 xsnac wrote: i still dont understand from the op wich one is better . talking in % like 60% of times is better to kill a scv 40% is better to kill a scv ... can some1 explain to me wich one is better overall ?
Mules at 50% time mines as much as an SCV (close enough)
Mules have 60hp, SCVs have 45
Kill new mules, but if the mule has only half its timer left, kill SCVs.
On average you will meet mules at half health, play it safe and kill SCVs unless you see the mule landing with your two eyes. (or one, I guess, if that's all you have.)
On August 01 2013 00:57 xsnac wrote: i still dont understand from the op wich one is better . talking in % like 60% of times is better to kill a scv 40% is better to kill a scv ... can some1 explain to me wich one is better overall ?
Mules at 50% time mines as much as an SCV (close enough)
Mules have 60hp, SCVs have 45
Kill new mules, but if the mule has only half its timer left, kill SCVs.
On average you will meet mules at half health, play it safe and kill SCVs unless you see the mule landing with your two eyes. (or one, I guess, if that's all you have.)
On August 01 2013 00:57 xsnac wrote: i still dont understand from the op wich one is better . talking in % like 60% of times is better to kill a scv 40% is better to kill a scv ... can some1 explain to me wich one is better overall ?
Mules at 50% time mines as much as an SCV (close enough)
Mules have 60hp, SCVs have 45
Kill new mules, but if the mule has only half its timer left, kill SCVs.
On average you will meet mules at half health, play it safe and kill SCVs unless you see the mule landing with your two eyes. (or one, I guess, if that's all you have.)
thanks . op should put this on the first page .
so ppl kill scv unless you see mule landing .
Theoretically yes, I didn't actually test/ran the numbers myself; just basing it off what the OP says.
EPT
![[image loading]](http://i.imgur.com/ZhcHMV8.jpg)
![[image loading]](http://i.imgur.com/NmE96A2.jpg)
