|
I've gathered quite a bit of empirical evidence about the optimal number of drones that should be placed on a mineral patch or gas mine and the numbers are surprising. Anything past 3 per mineral node nets you NOTHING. You will get upto 1200 minerals per minute max no matter how many drones you place on your minerals. Using the normal distance of 3 grid squares from the starting cc, it appears that 3 is also the optimal number for gas.
I see a lot of guides, especially on other sites, telling new and experienced players a like to never stop building drones. If you have some other use for them, or you are stocking up to move to an expo then there might be a reason for it, but don't think that more than 3x per mineral node is going to give you more minerals or gas, because it wont. 100 drones still brought in 1200 per minute.
I've also seen other guides tell people to make 1.5-2 per for optimal results. You might want to do this to shift resources elsewhere but I have found that it does scale evenly up to 3 per where it just craps out.
|
for zerg you need at least 5 on minerals and 3 on gas per hatchery to get the benefits.
|
3 has always been optimal for gas unless it was below the cc like luna or something..
|
On February 04 2010 06:13 wut_wut3 wrote:
for zerg you need at least 5 on minerals and 3 on gas per hatchery to get the benefits.
If it's a 8 patch mineral field, then that equates to 40 drones per hatchery. If you have three bases, that's 120 drones. FAR to many. As opposed to 24 per expansion, which comes out too approximately 72 drones for three bases. Still a bit excessive. For Zerg, it's harder to have a high drone count since you use a lot of your drones and larva for buildings and military. But the most important thing for a zerg player (especially in late game) is to have 3-5 running gas extractors with 3 drones each to support tech and units.
I really don't know where you got at least 5 per mineral patch.
|
I've actually been wondering about this for a long time and I would love to see some calculations... They must have been made at some point.
As for now I just look the base and see if it "seems" saturated, which can't be the optimal solution.
And yes, for 3 hex gas you need 3 - I do remember that one...most of the time 
|
I don't think he meant 5 per patch, but more like 5 per expo.
|
On February 04 2010 06:29 yarkO wrote:
I don't think he meant 5 per patch, but more like 5 per expo.
yes thats what i meant
|
You're probably right Xsebt. In the heat of the game I have never been able to successfully count the actual number of scv's I have on any patch. I usually stop a little early, I think. After a while though, you do get a pretty good idea of what feels right.
I did my tests on 8 node patches, which means 24 is the cut off point. Most people have half that by the time they are making any kind of combat units, which is about 3 minutes. I would say a lazy mans guild is to stop at around 6-7 minutes unless you have some other purpose for them like scouting, repairing units, making buildings out of them  or whatever ...
Of course you may want to build a few more than that to account for ones that are building and the 3 on gas.
|
On February 04 2010 05:45 sfdrew wrote:
I see a lot of guides, especially on other sites, telling new and experienced players a like to never stop building drones. If you have some other use for them, or you are stocking up to move to an expo then there might be a reason for it, but don't think that more than 3x per mineral node is going to give you more minerals or gas, because it wont. 100 drones still brought in 1200 per minute.
This is solid advice for early game, but you should not take the "never" part literally - it just means early to mid game (and even then, assuming you aren't cutting workers purposely for some sort of timing push). The advice is also meant for T and P, since Z sort of alternate between units and drones depending on the flow of the game / expansion status so the Z drone production can be "lumpy". The terminal "optimal number per patch" depends on the number of expansions you have: more expos = less workers/patch.
I would say first and foremost learn the basic build orders / standard expansion strats / worker transfers for your race. Also, the search function is your friend.
|
@citi.zen
I was merely trying to reveal the hard limit set by the game, which, despite your condescending "search function is your friend" comment isn't readily found here or anywhere for that matter.
New players probably will take the "never" part literally because they don't' know any better.
|
you might be able to get away with never stop building probes and scv but for zerg that could be dangerous
|
|
|
2.5 for prtoss and terran
1.5 for zerg
|
As a matter of fact I have given the search function a try, and I looked through all of liquipedia and I found nothing. I'm curious what you typed in the search to get that to appear? After an exhaustive second attempt to find that pdf on my own I finally had to boil my search down to one word "mineral" and then wade through a thousand mineral hack results to find the one that had that link half way down.
When people say things like "the search function is your friend", what they are really saying is, "you're a fucking idiot who shouldn't be posting because any noob can find this info by typing blowme in the search bar, look how l33t I am!"
I'm not wasting my life wading through page after page of back logged threads and search results for something that may not even exist. I check the first couple pages of threads and the most obvious searches along with liquipedia, after that I'm done, and I don't think that is unreasonable. Since there is no liquipedia entry on this subject I guess I'll make one and add a link to that document for good measure so this sort of thing doesn't happen again.
btw: I apologize for being so edgy. I am used to posting on message boards and blogs where things almost always degenerate into fighting.
|
|
|
I think the bigger problem here is the poor way in which forum data is orginized. Newer posts aren't automatically connected in anyway to older posts except through the literal duplication of words; duplication of words is a very hit or miss way of searching and doesn't pay proper repsects to the meanings and contexts in which words are used.
Look at all the different words that were used for this specific topic. Optimization; saturation; wander; scv; drone; probe; worker; production; etc... I might think to search for one of these words, but what about the others? The search function is very literal. The search algorithm doesn't know what words mean; it doesn't know if I am talking about cheese the food or the tactic. If I type drone it isn't going to think to look for scv and probe also.
Some gimmicks have been introduced to try and solve this, such as "tags". Web pages stuff themselves full of invisble words and meta-tags to try and get ranked higher on search engines; photo sharing sites and social networking sites allow you to "tag" things to aid searchers in finding it. Solutions such as "tags" only allow people to exploit a limited system, but they don't solve any problems.
Until the day when search engines know what you mean instead of what you typed and new information can be intelligently grouped with old information without a human hand sorting it, these types of problems will remain.
People who have been around for a long time get tired of seeing the same information appear over and over again, but there is some value to this duplication under the current situation. Because the search functionality is so literal, new threads that use different wording from the old ones increase the chance that somebody new will get a hit when searching for that content. Eventually every possible verbage will be used to describe the same thing and it will be almost a gaurantee that somebody will find a related thread when searching.
|
|
|
|
|
|
EPT
