|
United States7166 Posts
a problem several starcraft 1 players will have will be breaking their habits and intuition based from their sc1 knowledge. a lot of SC1 knowledge is ofc very helpful in sc2, but some of it will hurt
the main one I can think of is:
in SC2, 2 workers per mineral field gives you almost complete saturation. According to some replies in this thread, it's about 75% of full saturation 3 workers per patch gives you 100% saturation. *of course you want to have extra workers to transfer when you make expansions, the point is that the worker ai has improved greatly and 2 workers per patch in SC2 is way better than 2 workers per patch in Brood War.
---
EDIT:
Let me give an example of how players can sometimes think a unit is worse than it actually is:
New zerg players encounter fast 2 or 3 gate zealot rushes, tried using zerglings but they get demolished. They incorrectly come to the conclusion that zerglings suck vs zealots, then they discover roaches and how hard they counter zealots. In reality, in all my tests, speed zerglings beat zealots in terms of equal cost every time, the problem with going zerglings vs a fast 2 or 3 gate rush is that they can outproduce you on zerglings, zerglings are too cheap to spend all your minerals on with just 1 hatch, even with Spawn Larvae. Proton Charge of course helps the rush quite a bit. In my experience, protoss has no hard counter to zerglings before Colossus, which decimate them like nothing else of course. but by then you should have mutalisks which completely wreck colossuses. The point here is that players need to carefully analyze and think things through before quickly judging "oh this sucks vs this or I can't do this because of that."
__
If anyone else has any thoughts of habits or tendencies people have that aren't helpful in SC2, please post them here
a few zerg-specific ones I can think of that people probably figured out quickly is:
Don't spam hatcheries, get used to using spawn larvae from queen instead.
melee units like zerglings have improved significantly because of how closely tight they can fit together, being able to surround units much better. don't underestimate them.
try to spread creep as much as you can, (creep tumors are excellent for this, as are overlords) the speed boost is very significant and they help give map control/vision.
|
Even if you reach saturation, you should still make workers for maynarding to expansions.
However, I've ran into people making 40+ workers for one base play and 50+ for two.
|
Its still semi-useful to get extra workers if only to be able to transfer a lot of them to your next expo and have it up and running quickly.
edit: beaten!
|
well u have to build more workers if u plan to expand and want expansion to work to its fullest right away...
edit: beaten x2 :D
|
On February 21 2010 07:41 silencefc wrote:
Even if you reach saturation, you should still make workers for maynarding to expansions.
However, I've ran into people making 40+ workers for one base play and 50+ for two. I think Zelniq is zerg, in which case you really don't need an advance worker count for maynarding because you can saturate in moments with spawn larvae. It probably deserves an asterisk though 
|
Yeah, about that creep thing, I hardly ever see Zergs use it. =/ It doesn't even cost anything does it? That's definitely a good tip.
|
As many people have been saying, it seems that wc3 players will be better off than sc players in sc2.
Notable human players.
|
People gets the pool way early vs protoss, just because you have 200 minerals doesnt mean that you need the pool that early, you can make it 3 drones later and thats like +33 minerals or so which allow you to expand faster...
|
On February 21 2010 07:52 peidongyang wrote:
As many people have been saying, it seems that wc3 players will be better off than sc players in sc2.
Notable human players.
soooo wrong haha
|
I already eco-cut from time to time, but 2 scvs / mineral patch is not enough unless you don't want to expand..
|
I still try to use f buttons from time to time. Its a shame they were taken out.
|
What testers have to figure out is whether or not making more than 2x workers per patch still gives you enough return to justify spending that money (and what number of workers this occurs at) so transferring is efficient.
|
You should only stop and not maynard if you're zerg or terran, as terran you have to have 18-19 per base's mineral patches so that your special worker guy(I always forget his name since I don't play terran) can actually be of some help to you.
|
I did a quick test using probes. Mined for three minutes at various probe counts (clicked all probes on middle mineral patch and started timer the second the mining animation on that patch began - hit S on all when the timer hit zero). Here's the results:
+ Show Spoiler +16 probes: 998 -> 3768 = 2770 minerals gathered = 15.388888888 per second
20 probes: 10513 -> 13588 = 3075 minerals gathered = 17.083333333 per second
24 probes: 7143 -> 10483 = 3340 minerals gathered = 18.555555555 per second
51 probes: 3768 -> 7103 = 3335 minerals gathered = 18.527777777 per second
So yes, worker AI is amazingly better. Over 16 and the effects of adding additional probes drops fairly quickly. When you hit 24 (3 probes per patch) it stops COMPLETELY. There was even slightly more mined with my 24 probe run than my 51 probe run, I would be convinced that doing this test over and over would show they mine exactly the same. So think about it this way: Every probe you build and throw on minerals after 24 is doing absolutely nothing. If you are adding them because you plan on expanding, you must remember that they are just wasted minerals until the moment that expansion comes up. If you could get away with it, it'd even be better to long distance mine than add a 25th probe onto your main.
|
On February 21 2010 08:35 3FFA wrote:
You should only stop and not maynard if you're zerg or terran, as terran you have to have 18-19 per base's mineral patches so that your special worker guy(I always forget his name since I don't play terran) can actually be of some help to you.
The MULE can mine a mineral patch that is already being mined by an SCV.
|
Making a very sturdy wall, only to have blinked units and colossi walk right over the cliff and pwn your tanks.
Worker saturation is about 5-10 units lower for minerals, but higher for the double gas (apparently gas demand is higher in SC2).
Trying to fast expand on minimal units is a bad idea without scouting first. Unless 1-base play is truly the strongest (which I doubt for some matchups), most of your opponents will be doing all-in builds or are hyper-aggressive.
|
I did the test for protoss. Did it in like 10minutes with a friend me building probes he as observer checking my income / harvesters.
Results:
12Probes: 457-533
16Probes: 629-686
20Probes: 725-781
22Probes: 743-819
+24Probes: 762-838+
26Probes: 762-838
30Probes: 781-857
34Probes: 781-857
Posted this 3 days ago but got ignored '-.-
|
On February 21 2010 09:41 Ziph wrote:
I did the test for protoss. Did it in like 10minutes with a friend me building probes he as observer checking my income / harvesters.
Results:
12Probes: 457-533
16Probes: 629-686
20Probes: 725-781
22Probes: 743-819
+24Probes: 762-838+
26Probes: 762-838
30Probes: 781-857
34Probes: 781-857
Posted this 3 days ago but got ignored '-.-
This needs more attention (and information - how many patches? what do the numbers mean? did you stop at 34?)
|
With 8 patches at each base (this seems to be standard), you can have 16 workers, meaning you have 2 workers every patch.
If you add a third worker for every patch, you will take away that tiny, miniscule time-gap between when your mining worker gets off the patch and your incoming worker grabs the patch. If you have a third worker, it will stay at the patch and immediately grab it once the mining worker leaves.
That's why resourcing doesn't go any higher after 24 workers at a base. Basically, you can have up to 3 workers per patch without wasting any minerals.
In the end, you still only make like 200 extra minerals in the long run - and I mean the long run. You're better off going with 16 workers, while also stocking up additional workers if you're meaning to expand.
|
On February 21 2010 07:39 Zelniq wrote:
a problem several starcraft 1 players will have will be breaking their habits and intuition based from their sc1 knowledge. a lot of SC1 knowledge is ofc very helpful in sc2, but some of it will hurt
the main one I can think of is:
in SC2, you should never go past 2 workers per mineral field. that's the maximum amount of gathering you can get. *of course you want to have extra workers to transfer to new expansions, but not for mining out a base* This is because of the improved A.I. lack of wandering around looking for mineral patches. I figured this is not something a lot of players will realize on their own.
---
If anyone else has any thoughts of habits or tendencies people have that aren't helpful in SC2, please post them here
a few zerg-specific ones I can think of that people probably figured out quickly is:
Don't spam hatcheries, get used to using spawn larvae from queen instead.
melee units like zerglings have improved significantly because of how closely tight they can fit together, being able to surround units much better. don't underestimate them.
try to spread creep as much as you can, (creep tumors are excellent for this, as are overlords) the speed boost is very significant and they help give map control/vision.
In addition to this, I've also noticed the following very important difference:
Zerglings can chase retreating units very effectively and get in potshots on them. So retreating from Zerglings is gonna give them free hits. This is much unlike SC1 where retreating units caused Zerglings to wig out pretty badly, making them effectively do no damage except to stragglers.
|
For zerg
On February 20 2010 18:22 Virtue wrote:
After 1 min this is what i got in minerals
16 drones 925
17 drones 945
18 drones 990 -- 995
19 drones 1040 -- -1050
22 drones 1080
32 drones 1060
19-20 Drones is full saturation on 8 minerals in a field for a zerg player
How did i do this well i put 32 drones on the min field waited 2 mins so they can shuffle all though
then i picked numbers at random till i saw a drop to show that it's no longer full saturation.
Basically i took units away waited 1 min then counted again to try to get the most accuracy but it's only to the second.
Still 19-20 drones seems to be the right numbers i know i could probably figure out if it's 19 or 20 but i already wasted like an hour doing this
3 on Gas is optimal ~160 a min on gas
4 doesn't bring anymore and 2 doesn't bring enough
I only really ran it once and didn't like triple check it but around 19 units on 8 min patches is fully saturated for zerg.
|
On February 21 2010 09:45 d3_crescentia wrote:Show nested quote +On February 21 2010 09:41 Ziph wrote:
I did the test for protoss. Did it in like 10minutes with a friend me building probes he as observer checking my income / harvesters.
Results:
12Probes: 457-533
16Probes: 629-686
20Probes: 725-781
22Probes: 743-819
+24Probes: 762-838+
26Probes: 762-838
30Probes: 781-857
34Probes: 781-857
Posted this 3 days ago but got ignored '-.- This needs more attention (and information - how many patches? what do the numbers mean? did you stop at 34?)
its 8patches as thats the standard in sc2 for now. When watching the income screen as observer you can see that its constantly changing even with the same amount of workers. So the first number is the lowest mineral/minute it gives and the second number is the max minerals/minute it gives. Yes i stopped at 34. It didn't increase at all from that point. and only increased with 20minerals/minute from 24workers to 34 workers.
|
THe maximum is 3 workers per patch, thus most bases are standard 8 patches, saturation is 24 workers. Any more produces 0 increase in production.
For protoss, this amounted to an average of 800 minerals per minute per base, presuming 8 patches.
QED: Do not produce more than 24 workers at a base (30 including your gas miners) unless you intend to expand.
Currently doing gas tests.
220 gas per base, 450 overall. Maximum saturation is 3 probes, more is worthless.
2 Probes you get: 146 per base, 292 overall.
Unsurprisingly, with 2 workers as opposed to 3, you gas 33% less.
This was on Steppes of War, where every base has the same gas setup (two opposite sides of base). Did not teset both same side.
|
Since you can select so many units you can easily tell where you are at with worker count.
3 full rows in the selection box = 24 workers.
|
16 probes ought to be enough for anybody
|
On February 21 2010 10:44 floor exercise wrote:
Since you can select so many units you can easily tell where you are at with worker count.
3 full rows in the selection box = 24 workers.
usually plus two, which are currently mining gas. or are the workers in the geyser selected as well?
|
On February 21 2010 09:41 Ziph wrote:
I did the test for protoss. Did it in like 10minutes with a friend me building probes he as observer checking my income / harvesters.
Results:
12Probes: 457-533
16Probes: 629-686
20Probes: 725-781
22Probes: 743-819
+24Probes: 762-838+
26Probes: 762-838
30Probes: 781-857
34Probes: 781-857
Posted this 3 days ago but got ignored '-.-
Nice work. Wonder if anyone has tested this with drones and scv's. To see if anything is different.
|
On February 21 2010 07:52 peidongyang wrote:
As many people have been saying, it seems that wc3 players will be better off than sc players in sc2.
Notable human players.
Can you elaborate a bit? I thought that WC3 players specialize in micro, yet most of the TL staff have said that the game is heavily macro favored.
|
Currently, it appears that SC2 focuses heavily on macro, not micro. You can see that in most streams.
|
"spamming workers" is not a question of maximizing ONLY that base! It's having the extras to have the next base running ASAP.
|
United States7166 Posts
yes but I was mostly referring to people who keep spamming workers w/o even expanding soon after
im just saying, use your judgment and be watchful of your worker count.
ALSO i updated the main post with another thing I thought up that's been bugging me
|
Aotearoa39261 Posts
I agree with your edit Zelniq
Speedlings are difficult to stop as protoss, if you factor in trying to prepare for Roaches as well... Im starting to think gate forge is a necessary build;;
|
On February 21 2010 10:02 Zeke50100 wrote:
With 8 patches at each base (this seems to be standard), you can have 16 workers, meaning you have 2 workers every patch.
If you add a third worker for every patch, you will take away that tiny, miniscule time-gap between when your mining worker gets off the patch and your incoming worker grabs the patch. If you have a third worker, it will stay at the patch and immediately grab it once the mining worker leaves.
That's why resourcing doesn't go any higher after 24 workers at a base. Basically, you can have up to 3 workers per patch without wasting any minerals.
In the end, you still only make like 200 extra minerals in the long run - and I mean the long run. You're better off going with 16 workers, while also stocking up additional workers if you're meaning to expand.
this post was very helpful!
So the basic strategy is 16 mineral workers per base + up to 24 when you're maynarding - and over 24 is never needed/wasted investing?
|
Iloveoov proved you can never have too many workers. From every game I've watched I've seen a distinct lack of worker production.
|
On February 21 2010 10:40 ghosthunter wrote:
THe maximum is 3 workers per patch, thus most bases are standard 8 patches, saturation is 24 workers. Any more produces 0 increase in production.
For protoss, this amounted to an average of 800 minerals per minute per base, presuming 8 patches.
QED: Do not produce more than 24 workers at a base (30 including your gas miners) unless you intend to expand.
Currently doing gas tests.
220 gas per base, 450 overall. Maximum saturation is 3 probes, more is worthless.
2 Probes you get: 146 per base, 292 overall.
Unsurprisingly, with 2 workers as opposed to 3, you gas 33% less.
This was on Steppes of War, where every base has the same gas setup (two opposite sides of base). Did not teset both same side.
Yeah, because you never get harassed or raided....Workers have a lot of uses besides purely gatherers.
|
On February 21 2010 19:58 Rothbardian wrote:
Iloveoov proved you can never have too many workers. From every game I've watched I've seen a distinct lack of worker production.
gj completely missing the point of this thread
|
On February 21 2010 20:04 Urth wrote:Show nested quote +On February 21 2010 19:58 Rothbardian wrote:
Iloveoov proved you can never have too many workers. From every game I've watched I've seen a distinct lack of worker production. gj completely missing the point of this thread
The point of the thread was to see how many workers completely saturate a mineral field. What I am saying is that is irrelevant.
|
does this means at a certain point u would gain more from long distance mining with your extra workers then u would if u just keep them mining at your base(even if it wasnt alot more its still more then keeping them at your base)
|
Your point is valid to a degree, but I would argue that throwing more on your patches really is a complete waste. If you need probes around your base for building, you may as well have them idling in good spots over putting them on the minerals. If you suspect a successful harass, by all means pump up production, but why throw them into the fray of the harass when you can keep them separate? Maximum saturation at 24 means there is literally no benefit to putting extras on the mineral fields compared to idling them anywhere else. If you plan on expanding, and are at 24 probes, you may as well rally them anywhere else in your base.
edit: On February 21 2010 20:20 drivec wrote:
does this means at a certain point u would gain more from long distance mining with your extra workers then u would if u just keep them mining at your base(even if it wasnt alot more its still more then keeping them at your base)
Yes. In fact I made that point in my post on the last page I believe.
|
On February 21 2010 20:06 Rothbardian wrote:Show nested quote +On February 21 2010 20:04 Urth wrote:On February 21 2010 19:58 Rothbardian wrote:
Iloveoov proved you can never have too many workers. From every game I've watched I've seen a distinct lack of worker production. gj completely missing the point of this thread The point of the thread was to see how many workers completely saturate a mineral field. What I am saying is that is irrelevant.
Mining works very differently in SC2, in SC1 you never really had 100% of the mining potential possible even if you pumped workers all the time, you certainly did not hit full mining potential at around 20 workers like it seems to be in SC2.
|
On February 21 2010 20:20 drivec wrote:
does this means at a certain point u would gain more from long distance mining with your extra workers then u would if u just keep them mining at your base(even if it wasnt alot more its still more then keeping them at your base)
Slightly off topic, sry:
Bisu vs Stork. I think it was Incruit OSL Quater finals on Chupung Ryeong.
|
As for mining. I just did a test. On 8 mineral patches total, my income when viewed in the replay fluctuated from 600-700 with 16 workers. I introduced an additional 4 workers, bringing the total to 20 workers and the income jumped to 700-800 range. After introducing an ADDITIONAL 4 workers, bringing the total to 24, the mineral range stayed the same. So it seems the cap is approx 20 workers on 8 mineral patches; or approximately 2.5:1 worker to mineral patch ratio.
I would be interested to see if other people get the same results. The best method to test is just to play an easy computer, and let workers mine for 3-5 minutes at each worker count level (16/20/24).
|
Someone should test the following: (give me a key and I'll do it myself)
24 workers in main
compared to
16 workers in main, 8 long distance mining from natural
|
It's really no different than SC1... the best test to do is to time it for clearing the mins at 8, 16, 24, and 32 workers because the longer time interval will make for less error. If you do this you'll see that 24 is superior to 16 even in SC2.
|
All in ling rush is really really powerful if you add another hatchery and start spamming zerglings out of those two with your spawn larva - it can only be compared with producing mass zerglings out of 3-4 maybe even 5 hatcheries in BW.
|
TY for this thread and contributors. I've been wondering about both gas and minerals (max workers) and I think this thread pretty much answered them both. In some early games I had some friendly players tell me I had way too many probes. He was definitely right 
|
On February 21 2010 12:10 nukular wrote:Show nested quote +On February 21 2010 10:44 floor exercise wrote:
Since you can select so many units you can easily tell where you are at with worker count.
3 full rows in the selection box = 24 workers. usually plus two, which are currently mining gas. or are the workers in the geyser selected as well?
It's +4, since there's 2 geysers per base, the 2 workers inside the refineries are not selected.
|
Here are the results of the some comprehensive tests I did for all three races using combinations of workers. I tested workers @ 8,16,24 (did 32 for protoss, had a negative effect so I didn't bother using it again). Shown below is Minerals Mined in 1 minute, Minerals Mined per Second, Minerals Mined per Worker per Second, Time to Recouperate the cost of the worker (from the time it finishes) and 'Time til Dry' is the time it would take to drain the mineral source from full to zero at that rate -- I think this is an often overlooked factor.
+ Show Spoiler +
Probes Minerals in 1min Minerals per second Minerals per Probe per Second Recoup Time (s) Time til Dry (min)
8 480 8 1 50 25
16 765 12.75 0.796875 62.74509804 15.68627451
24 1155 19.25 0.802083333 62.33766234 10.38961039
32 1140 19 0.59375 N/A
SCVs Minerals in 1min Minerals per second Minerals per SCV per Second Recoup Time (s) Time til Dry (min)
8 480 8 1 50 25
16 855 14.25 0.890625 56.14035088 14.03508772
24 1125 18.75 0.78125 64 10.66666667
Drones Minerals in 1min Minerals per second Minerals per Drone per Second Recoup Time (s) Time til Dry (min)
8 470 7.833333333 0.979166667 51.06382979 25.53191489
16 940 15.66666667 0.979166667 51.06382979 12.76595745
24 1120 18.66666667 0.777777778 64.28571429 10.71428571
Avg
Workers Minerals in 1min Minerals per second Minerals per Drone per Second Recoup Time (s) Time til Dry (min)
8 476.66\ 7.944 0.99 50.35 25.17
16 853.33 14.22 0.88 56.649 14.16
24 1133.33 18.88 0.78 63.54 10.59
Simple conclusions -- all races mine the same (and all workers cost the same gold and time to build).
Not-so-simple conclusions -- 20 workers is probably the best for your starting base as it provides a good balance between speed mining vs slow & steady, but also provides backup workers in case of loss or need to scout/expo/whatever.
|
Sorry, first time posting on TL.net
The chart isn't super easy to read (did it in excel), but the conclusion is this, averaged over three trials.
Workers || Minerals in 1min || Minerals per second || Minerals per Drone per Second || Recoup Time (s) || Time til Dry (min)
8----476.66-----7.94-----0.99----50.35------25.12
16--853.33-----14.22----0.89----56.65------14.16
24--1133.33----18.89----0.79----63.54------10.59
|
On February 25 2010 05:57 Gescom wrote:
Sorry, first time posting on TL.net
The chart isn't super easy to read (did it in excel), but the conclusion is this, averaged over three trials.
Workers || Minerals in 1min || Minerals per second || Minerals per Drone per Second || Recoup Time (s) || Time til Dry (min)
8----476.66-----7.94-----0.99----50.35------25.12
16--853.33-----14.22----0.89----56.65------14.16
24--1133.33----18.89----0.79----63.54------10.59
you should have gone higher than 24
|
3 workers on gas is still the norm tho right
|
3 on gas gives full saturation. 2 gives about 80% of the capacity, which can be useful at times.
|
Going higher than 24 does absolutely NOTHING on an 8 patch base. 3 per patch is 100% saturation. The only variation you'll see is natural error (You can't perfectly time anything).
I would suspect that 4 workers on gas is full, since 3 workers does sometimes leave a small time window in which you aren't mining full time (due to positioning).
There are some maps that have a 2nd gas that is farther way, so you'll definitely need 4 for those.
|
On February 25 2010 06:59 commanderchobo wrote:
3 workers on gas is still the norm tho right
Having more then three causes the excessive workers to have to wait in order to start gathering. So yes, it is still three.
|
That's not 100% saturation, though. 100% saturation is getting ALL possible mining. With 3 workers mining gas, there's still a small window of time the gas is not being mined from; with 4, the worker already waiting there will fill in that time slot.
Remember, saturation doesn't necessarily mean efficiency. It's probably most efficient to get 2 workers per mineral and 3 per gas. However, you still aren't getting the most out of the minerals/gas.
|
I tested 32, open the spoiler. Had 0 benefit.
|
On February 25 2010 07:31 Zeke50100 wrote:
That's not 100% saturation, though. 100% saturation is getting ALL possible mining. With 3 workers mining gas, there's still a small window of time the gas is not being mined from; with 4, the worker already waiting there will fill in that time slot.
Remember, saturation doesn't necessarily mean efficiency. It's probably most efficient to get 2 workers per mineral and 3 per gas. However, you still aren't getting the most out of the minerals/gas.
No, actually. If you watch a geyser with 3 probes on it, they do "bump" outside the assimilator and need to wait to get in. Thus, if you add another one... they'll both wind up waiting, and you don't gain anything.
The first day or two of beta I did some testing on this by ridiculously over-saturating minerals and gas, and then seeing how well you did at lower levels. Short version of the conclusions:
- A fully saturated gas geyser produces just under 2 gas a second - my best estimate is 116 gas per minute.
- 3 probes on a gas gives this amount.
- 2 probes on a gas gives about 75% as much - it may depend on the exact configuration, but the geyser I tested returned 88 gas per minute with two workers on it.
- A fully saturated mineral patch produces about 105 minerals per minute.
- 2 workers can mine at full capacity on each patch; doing so usually produces "about" 80-85 minerals per minute. However, there are "good patches" and "bad patches" which produced as much as 95 or as little as 75 minerals per minute with two miners
So, as a rule of thumb: 3 probes is full saturation on either gas or minerals; 2 probes is about 75-80% capacity. What I didn't test is how well mining saturation scales in the 2 to 3 probe per patch range - I think that's going to work out to be somewhat subject to one's exact mineral configuration.
Also interesting about this is the fact that a mineral patch is 1500 minerals, while a gas geyser is 2500 gas; however, gas only mines about 10% faster. Hence, it takes almost 50% longer to mine out a gas geyser than it does a mineral patch, meaning that as the game goes on, you wind up with a fair number of bases that are "gas-only" for at least some period of time.
|
Actually, the other reason for the early worker-spam is so you aren't creating workers *later* as you expand (especially in 1v1 or 2v2 on decently-sized maps with multiple fields/geysers). Because all the workers are already there, you can concentrate on structure building (and upgrades) as opposed to ramping up the economy. (The idle worker actually costs more later than earlier; also, excess workers can be used for repairs, in addition to tech-up/tier-up.)
With Zergling opposition, the tactical retreat can be used (like in original SC or BW), by drawing them towards Bunkered Marines/Firebats (especially if you also have Missile Turrets) or Photon Cannon, especially when mated up with Zealots and/or Dragoons. In short, *invite* the chase, then have them basically walk into a trap.
The default stance for Zerglings (same applies to Zealots and Marines) *is* to chase down opposing units (and a smart opponent will take advantage); instead, change the default stance (or a specific group stance) to Sentry/Guard. That also enables you to have an actual reserve (in case your attempt at a rush gets stonewalled).
|
@Steve496: It might be different for some maps, because I definitely remember seeing at least two maps where workers will be slightly behind.
|
On February 21 2010 08:18 Monokeros wrote:
I still try to use f buttons from time to time. Its a shame they were taken out.
+ infinity. F buttons are my favorite part of starcraft, bar nothing.
|
On February 25 2010 09:59 love1another wrote:Show nested quote +On February 21 2010 08:18 Monokeros wrote:
I still try to use f buttons from time to time. Its a shame they were taken out. + infinity. F buttons are my favorite part of starcraft, bar nothing.
Yeah, F buttons are really important for macro and defending against harass. I have my F2 and F3 buttons on my up/down scroll wheel. I hope Blizzard doesn't wait until the final release to add the F buttons.
|
On February 25 2010 05:57 Gescom wrote:Sorry, first time posting on TL.net The chart isn't super easy to read (did it in excel), but the conclusion is this, averaged over three trials.
Workers • Minerals 1 min • Minerals per second • Minerals per Drone per Second • Recoup Time (s) • Time til Dry (min)
w M/min M/s M/w*s Recoup(s) Time til dry(min)
8---476.66-----7.94-----0.99----50.35------25.12
16--853.33-----14.22----0.89----56.65------14.16
24--1133.33----18.89----0.79----63.54------10.59
reposting it for you, should be easier to read now
|
For zergs, did you try with 20 ? 
Because semantics wrote :
After 1 min this is what i got in minerals
16 drones 925
17 drones 945
18 drones 990 -- 995
19 drones 1040 -- -1050
22 drones 1080
32 drones 1060
19-20 Drones is full saturation on 8 minerals in a field for a zerg player
|
Some bad habits:
Sending ur overlord to the enemy's base in zvz - surprise, surprise, queen shoots air.
Spamming a hotkey - causes ur unit to go into attack move (i've lost my scouting units countless times that way)
|
We still need to keep pumping drones/scv/probe to transfer to expansion to really have the economical advantage of having 2 bases asap.
On the zergling, part, they are still very good against the zealots early game because of surround, but we must take note that if the player goes for a 3 gate rush with zealot/sentry, lings are definately not enough and roaches will be needed to fight because u cant get the surround that easily
|
This is a great thread. Any more information from you sc1 players on what skills from sc1 you should moderate? I'm a wc3 player learning about the game.
|
Its the balance of how many drones to make at a time that is hard to judge. I might get a spawn larva popping and have 4 larva ready to use early game and just don't know if I should spam a lot of lings to be safer or drones for economy.
If on 1 base it feels distinctly safer but that time where you put down an early/fast expansion you get so vulnerable to a variety of harassments. Some spine crawlers is really nice, but I guess learning to scout a lot more effectively is the key to knowing how hard you can pump drones.
|
All these numbers about mining eficiency are really cute and i'm tankfull to you guys who have crunched them.
BUT, i think is should be considered theorycraft. In most games, almost everygame i would say, you should make workers all the time, for so many reasons. The biggest reason is also on those numbers, and it is "time til dry". You will need to expand if the games moves on and you want that second exp saturated asap. Ok, maybe you had 35 workers on your base, that's not eficient, but when you expand you will have 5 extra workers that can be moved to your natural, and that is a good thing!
But also you will be harrased, you may need to use workers to defend some atacks. I play terran, Terrans do move workers out of mineral lines and use them on battlefield, they repair and build bunkers and turrets. Zerg do turn workers into buildings.
Just consider time til dry of a 100% saturated base is less than the average game (educated guess).
Am i wrong? I never felt like stoping making SVC is a good idea at all.
|
As Zerg, (assuming your expo is at your natural) instead of overproducing drones with the intention of Maynarding, would it not be better to build 7 (3 spawned larva 4 from vomit) drones from your hatch and set their spawn point to the expansions minerals with the intention of them arriving at (or moments before) the expo hatch completes & setting your main hatch's default drone spawn point to the expo minerals.
That way you are better able to match your income to spending without wasting income that could be spent on army/tech on surplus drones that will be needed "in the future"
Does this make sense, is it practical?
|
i constantly rewaypoint my nexi, and check my main to manually send probes to mins often.
its a very hard habit to drop :[
|
Manually distributing harvesters between patches is still beneficial, auto-mine supplements this it does not replace it. Zerg units move faster on creep making them faster then their counterparts and requiring lower numbers for saturation. Coupled with the fact that you use the same building as production as expansion (hatchery) then it isn't as much about optimal saturation as it is income to production ratio. As a zerg player saturating beyond 16 mineral and 6 gas per base (22 drones) is detrimental because you are either not keeping your production capacity large enough to effectively utilize your economy or you are significantly reducing the effectiveness of production expansion by not taking advantage of the added benefit of larger economy that would come if the hatchery is placed as an expansion as opposed to another building in the main . If however there are external factors like you have no map control for some weird reason then there might be some extenuating circumstances but for the large part as a zerg saturate to 16 and expand rinse and repeat. To keep this focused on topic I will refrain from going into more depth on how to work army into your droning.
|
I'm pretty sure drones don't benefit from creep movespeed though, do they?
But yeah, I usually stick to 16 drones on minerals unless I'm planning to expand soon.
|
T.O.P.
Hong Kong4685 Posts
I usually stick to a max of 16 workers per base. In SC2, you can't expand unless you have units and you can only get so much units if you don't make so many probes. I rather get a faster expand than more probes in my main base.
|
Drones do not benefit from creep, for obvious reasons.
|
I didn't do the testing but the theory is sound without the numbers for expanding > saturating beyond 2 per mineral and 3 per geyser and the numbers are as follows:
Probes Minerals in 1min Minerals per second Minerals per Probe per Second Recoup Time (s) Time til Dry (min)
8 480 8 1 50 25
16 765 12.75 0.796875 62.74509804 15.68627451
24 1155 19.25 0.802083333 62.33766234 10.38961039
32 1140 19 0.59375 N/A
SCVs Minerals in 1min Minerals per second Minerals per SCV per Second Recoup Time (s) Time til Dry (min)
8 480 8 1 50 25
16 855 14.25 0.890625 56.14035088 14.03508772
24 1125 18.75 0.78125 64 10.66666667
Drones Minerals in 1min Minerals per second Minerals per Drone per Second Recoup Time (s) Time til Dry (min)
8 470 7.833333333 0.979166667 51.06382979 25.53191489
16 940 15.66666667 0.979166667 51.06382979 12.76595745
24 1120 18.66666667 0.777777778 64.28571429 10.71428571
indicating drones do gain advantage by being on creep and attain optimal saturation earlier.
|
I ran a test like this a month ago with up to 200 probes.
Diminished returns start in after you get 2 workers for each node. You do get more minerals per minute with 3 workers, adding a 3rd worker does not increase collection by 50% over 2 workers. If 1 worker per node is 100% production, 2 workers gives us 200%, and the 3 gives us 255%. With 4 workers I got to around 270%. I pushed it all the way up to 25 workers per node and did not gain much more than that 270%.
I think that very small gain after 3 workers is from the distance of the back nodes.
To test this I made 200 probes and sent them in 4 at a time. I watched the income chart in the replay after it was over. It did go up as I added more and more probes to the patches, but the gain was very very small at 4+ workers per node.
|
On April 20 2010 20:49 CanT.ThinK.oF wrote:
Some bad habits:
Sending ur overlord to the enemy's base in zvz - surprise, surprise, queen shoots air.
Spamming a hotkey - causes ur unit to go into attack move (i've lost my scouting units countless times that way)
not sending a overlord to enemies base is rather silly.. you want to know at which time hes getting his pool / gas
|
On April 21 2010 01:20 Frenzied_Tank wrote:Show nested quote +On April 20 2010 20:49 CanT.ThinK.oF wrote:
Some bad habits:
Sending ur overlord to the enemy's base in zvz - surprise, surprise, queen shoots air.
Spamming a hotkey - causes ur unit to go into attack move (i've lost my scouting units countless times that way) not sending a overlord to enemies base is rather silly.. you want to know at which time hes getting his pool / gas
Yes. Just don't leave it there or go in range of the queen.
|
would sending extra workers in with troops to a battle, [i.e. Jinro vs. LucifroNNNN match] be a viable reason to continue worker production even after saturation?
|
This is true, but how does it relate to 'sc1 bad habits'? SC1 was basically the exact same- maxing out at 3 workers/patch, and there was significant diminishing returns as you reached 3..
|
On April 21 2010 01:40 Gokain wrote:
would sending extra workers in with troops to a battle, [i.e. Jinro vs. LucifroNNNN match] be a viable reason to continue worker production even after saturation?
for terran yes (repair, make bunker/turret), for toss/zerg, no. Unless you are zerg late-game, you have extended your creep-route all the way to your opponent's base, you have extra minerals, and you want to supplement your attack with some spine crawlers.
|
Did some quick calculations using the spreadsheet data and assuming the data is reliable...
It's not important what the per scv mineral rate is but rather the EXPECTED value of creating another SCV beyond 16.
We expect the expected value to behave like a piecewise step function, where the expected value is constant to 16, then falls to a lower value from 16-24, as there are 8 mineral patches to be filled to 100% capacity.
The expected value of an scv before 17 total scv is 57 minerals/min/scv/patch. The expected value of an scv from 17-24 total is 27.5 minerals/min/scv/patch. so SCV after 16 is worth 48% of an SCV before 16.
This is assuming the effect of scvs moving is of little effect. Does someone have multiple trial point to point data? I would like to calculate the expected value of each scv at each point in time as I think that is important, not the average mining rate of the whole group.
|
Couple things... I ran numbers for blistering sands using Zerg and Terran. The expo in the middle top and middle bottom on blistering sands are not the same for Z and T. Using 24 workers each the zerg mined the 12 out about 3 minutes earlier than the T mined the 6. I don't know if it's because the drones AI is better or because drones have a slight advantage or the mineral patches themselves are superior, but if given the choice I take the top middle expo.
The golds seemed identical with only a slight advantage to the zerg.
The other thing is that I win 90% of my games against 1800+ plats based solely on my macro. I make more workers than anyone i've played against on NA and I think that yea having too many on one base is bad, but I like to have my workers ready to move to my expo when it comes up.
edit: oh yea.. one other thing is that you can't rely on the in game "mineral counter". The best way to find out your mineral/minute is to have a set number of workers mine all of them for a set amount of time (8 minutes) or something and then divide the total gain over total time. The longer time period the better results you'll get. The mineral range that you get from the little mineral counter isn't going to give you reliable results for determining the value of each addditional scv.
|
starcraft911 You are failing to factor mule when you discuss the relative value of gold minerals for terran vs zerg. Terran can utilize the yellow minerals as a launch platform into an insane endgame. They can also create a planetary fortress to hold the minerals more easily. I disagree that zerg have some kind of edge or are even on gold minerals.
|
|
|
|
|
|
EPT![[image loading]](http://blogs.seattleweekly.com/dailyweekly/Bill_Gates_718639.jpg)
