|
Hello all, this is a quick informational guide to the community on how to properly mineral split. It is also a way for me (if I may) to continue to encourage TLers to watch my short, 10 minute show that I stream for TL readers (meaning not advertised or popularized on any other site)!
The premise of the mineral split in this context is to get the most minerals out of the split from the opening seconds of the game. In episode 4 I detail this in the beginning, so if you were to watch anything it would be the beginning of episode 4, where I go through the basics of my (Sidereal) particular mineral split vs a very common 3-3 split. It is a 4-1-1 split, where after the quick clicks each starting probe is on the closest mineral patches to the nexus, in the case of my race. By doing this, even though all players, especially in mirror matches, are identical up until usually the 12th probe, this split will put you ahead against other versions of mineral splitting.
I believe this split is superior. I, however, am a simple human and am capable of showing humility, so if it is actually not as good, please let me know! Also, if there are any comments (positive or negative, don't matta), please leave them on the TL thread for the stream! I am constantly reviewing it and trying to get better, so any feedback is def appreciated!
Link to stream:
http://www.livestream.com/tehemp
Link directly to episode 3, the beginning of which is the mineral split discussion:
http://www.livestream.com/tehemp/video?clipId=pla_8a9de969-ae7d-4e91-ab12-9519ada279d2
Link to stream thread:
http://www.teamliquid.net/forum/viewmessage.php?topic_id=185138
Happy hunting, jeee ouse.
|
|
|
The type of split does, I've seen that post that you linked and I have demonstrated that my split does in fact net higher gains, and it makes sense why it does as well. My split is different from the 3 done in the test, which wasn't very thorough to begin with. It doesn't even include where to add the 7th or 8th drone and why it matters where you add them.
I encourage you to try it out, I wouldn't post unless I was sure.
|
It's been a while since I read that thread, but I don't remember it ever discussing the whole "send workers to the closer patches" idea; rather, it just let the auto-split work its magic.
|
Did you say your name was Cytheria? Because if it is, I hope you still have that free name change floating around..
|
I do not believe in the ep 3 the opponent splitted his probes. It looks like her just sent to one patch and ai did the rest. Anyone else agree?
EDIT: Also tlo gets pairs on closest mineral patches as soon as possible I've noticed.
|
I think it's not so much about the initial split, but where your probes end up mining from after they are split. each mineral patch can support 2 probes without the probes having to skip around to other patches. in this aspect, you ideally want 2 probes on the nearest mineral patches and zero probes on the furthest mineral patches. doing so will obviously give you a higher mining rate
|
Just send all 6 to the closest patch then split 2 or 3 to a different section (depending on the patch set-up) in a way that all will mine without having to overlap each other and travel all the way to the other side. Very simple ^^
|
Its already been mathematically proven that the split's gains are so negligible (literally 5-10 minerals with no increase after 2 minutes) to make it completely worthless as the risk of misclick would cost you more.
Just send all 6 to the middle patch and 3 minutes in you'll have the same money no matter what
|
Hi
I watched the vod, and I'll have to disagree on the way you proved how your way was better. Basically, you pulled up the income tab, and then paused it the moment your income spiked the highest (right when workers sent in their minerals). As you can notice before, the icome fluctuates, going up and down, and you simply noted the highest point, then compared it to your opponents income, which is also in a state of fluctuation.
The best way to really see how effective your split is would be to not build any buildings and compare how many minerals you have compared to a 3-3 split at a certain time in the game.
Anyways, the split certainly looks nice, doesnt seem to bad, but i doubt it is a significant improvemnt to worker splitting (and i dont think 3-3 helps that much either)
|
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy.
|
On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy.
Way to read the post bro. Or even the topic title.
|
On February 02 2011 16:57 ch33psh33p wrote:Show nested quote +On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Way to read the post bro. Or even the topic title.
I did and the poster was proposing a way of faster worker splitting and/or how to start ones economy as fast as possible, (as stated in the post before mine (the one you quoted), he used a flawed method of obtaining data) not to mention that in addition to that its been proven worker splitting slows you down, thus making this whole topic redundant. Perhaps you should re-read the topic to make sure you retained everything, bro.
|
I started watching the video... there's a couple of things I noticed that make me immediately not trust this.
1: You're playing at different spawns of the map. We've seen many times that blizzard's maps aren't symmetrical and have positional imbalances (scrap station zvz, the base structure in steppes, etc). The difference could purely be a result of that. You need to test it either in seperate games on the same spawn, or also do a "control" test where you both 3-3 split and see if they're the same on the different spawns.
2: Now when I see that it's based on whether you got your probes to the close minerals or the far ones means it's just based on which mineral patches you click on, and also on the map.
3: Hard to tell but tbh from the video it looked like the bottom left guy didn't even 3-3 split, looks like he just sent them all to one patch and they spread out themselves (I say that as I don't see the left 3 all go to the same patch and then spread... they go straight to their split up patches)
|
5 more minerals aren't really going to benefit you to any slightest significance, so I don't even see the point in discussing this.
I also don't get how people can even think this would give an advantage in the mid or even late game. If you send your workers out half a second later than normal, or do some sort of failing split causing them to not mine effective for half a second, all this will do is set you behind half a second. How hard is it to realise that?
|
On February 02 2011 13:02 tehemperorer wrote:
I, however, am a simple human and am capable of showing humility
The first 60 seconds of your video proved that part wrong.
Also, it's been proven time and time again that it makes no difference whether you 3-3 split, 4-1-1 split or don't manually split at all. It's a nice warm-up and micro exercise and that's about it.
|
Talk about beating a dead horse.
|
Ive done auto/3-3-/2-2-2 it seems like the only difference from 3-3/2-2-2 and auto is the 3-3/2-2-2 allows all five workers to finish there first mineral drop at the exact same time witch allows for 50 minerals to build your first scv. Auto seems to always give me 45 minerals and there's a slight delay in between 45-50minerals where i can't build the scv. So i dunnno if this helps or if the mineral difference would even matter but it's nice to split for micro practice.
|
I usually just do the 3-3 drone split, then set my mining rally points each time I make a drone or two, moving it across the patches in a line... the AI seems to do a bunch of work for me...
The only thing I'm worrying about right now is which is the best way to get my second overlord out... 9OL, 11 extractor trick, DOUBLE extractor trick, 10OL?
|
On February 02 2011 17:36 Shockk wrote:Show nested quote +On February 02 2011 13:02 tehemperorer wrote:
I, however, am a simple human and am capable of showing humility The first 60 seconds of your video proved that part wrong. Also, it's been proven time and time again that it makes no difference whether you 3-3 split, 4-1-1 split or don't manually split at all. It's a nice warm-up and micro exercise and that's about it.
Technically, there is a difference. We just don't have infinite APM, so we can't abuse the difference >.<
Also, I still can't remember if they tested actual mineral distances (i.e. individually sending each worker to a closer patch or whatever); the only thing I remember is people saying auto-mine is just as effective as any other split. I'll have to look through the thread again, though.
Although I've seen plenty of people do REALLY stupid splits, forcing some workers to move much more than they should (i.e. sending the whole stack to a far patch, not leaving enough patches between two groups to have "flawless" mining, etc.). It's expected, though, since it's pretty much random.
However, that is not to say that splitting doesn't matter; if you're terrible at 3-3, then don't do it (and so on).
|
I've always found the classical posts 'disproving' the effect of worker splitting to be very flawed. The round trip time for a mining worker is about 7 game seconds. If you measure the amount of minerals at a particular time, you can miss a huge reduction of mining time because the mineral count is incremented so rarely. This is easiest to see with an example. Suppose you have 1 worker that you send a 0sec to mine a patch with RTT of 7secs. You measure the amount of minerals at 1min and get 60/7*5 = 40 minerals. You restart and send a worker at 4secs to mine the same patch. You measure the minerals at 1min and get (60-4)/7*5 = 40 minerals. You can't even detect a 4 sec reduction in mining time. The best way to do this is to measure the time it takes to reach a certain mineral threshold, preferable down to the millisec. I believe the advantage is about 1 sec to doing a real split. It is still small admittedly.
Here is the data I have on hand. I was checking these splits: 3/3, 2/2/2, F1, F1+Stacking. The F1+Stacking is a split that I devised myself. (I actually thought the split in this thread would be it based on the name). This F1+Stacking trick is using F1 to send 4 workers to the closest 4 patches, then wait a bit and send the next 2 workers to stack on 2 of these closest patches. Basically the idea was to get all workers on the closest patches as soon as possible. This data is the time it takes to reach 50 minerals (for the second worker).
3/3 Splitting 17.3
2/2/2 Splitting 17.8
F1 Splitting 16.9
F1+Stacking 16.7
*Note: I am bad at 2/2/2 splits.
As far as the split posted here, I don't think it is really going to be better than the standard splits. You don't actually really split off the workers until they are almost at the mineral patch, in which case autosplit would work almost as well. Perhaps surprisingly, it isn't that important to get to the closest patches immediately. After the workers return minerals, it is easy to adjust them to go to the best patches. Also, there is a much smaller distance to run if you switch patches at the nexus compared to switching patches in the mineral line.
|
On February 02 2011 17:42 teh_longinator wrote:
I usually just do the 3-3 drone split, then set my mining rally points each time I make a drone or two, moving it across the patches in a line... the AI seems to do a bunch of work for me...
The only thing I'm worrying about right now is which is the best way to get my second overlord out... 9OL, 11 extractor trick, DOUBLE extractor trick, 10OL?
Just shooting a little advice out there:
If I'm planning on 14/15 hatching I 10 OL.
If I'm sensing early pressure I go either 11 extractor or 9OL.
Going 11 extractor allows you to 11 overpool allowing for a quicker defense v. all ins and such. 9OL I transition into 14 pools and such.
|
On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Is there a link to a thread proving I should build a worker before I send my six workers to mine?
It should be better to send the 6 workers first, because you get 6 time the income of a single worker. (I am assuming that I am no perfect player, so I have to decide if I should delay the first worker to build or the six given workers to work.)
|
OP's concept is completely valid. No one is going to disagree that keeping your miners on the closer patches is optimal.
The way I split my probes is like this (and btw every time I split, my APM bursts to around 700, so keep that in mind for viability): I f1 split, and then individually select probes, before they reach the minerals, to send them to specific locations if necessary. For example, if I f1 split but accidentally send one to a far patch, I'll select him and have him go to the nearest close patch instead, before he ever reaches his destination.
PS: I agree that my original thread on splitting is scientifically flawed when trying to determine (taking the above example) fastest way to get 50 minerals. However, it does accurately represent the idea that "splitting basically doesn't matter", which is true.
|
it only matters for the first return but i have a theory as to why it balances out. suppose you get perfect split and then perfect placement for all 8 of your first workers. great. now every mineral patch has a worker on it and u now have the highest probability chance (at that moment in the game) for the next probe to spawn from your nexus to be rallied to a busy mineral patch, causing it to search for the nearest free one - and the likelihood of a free one being farther away increases corresponding to the efficiency of your spread. this is the reason why after 2 minutes its about the same, however a proper harvester spread will get you enough to get that first probe out with no lag time and with most efficient placement you will get your pylon as quick as possible which does matter if one is 10 gating or proxying but in most other builds its irrelevant. i'm not sure how the other races can make use of the early harvester efficiency but really the math is all within the explanation i gave you i just don't understand the numbers. i'll update my sig soon too, go by zugzwang now.
|
I always end up screwing it up, well most times I'm fine but when I start tilting I fuck up small stuff like this and it makes me rage lol. Nice contribution but!
|
To be honest, no matter how you split your worker, it is so insignificant in terms of gameplay.
It's all about warming your hand, using that skill throughout the game like marine split against banglings or high templars or siege tanks. What so good if you can do a 1-1-1-1-1-1 split at the start when your marine blob dies to banglings and lose the game?
|
On February 02 2011 13:02 tehemperorer wrote:Hello all, this is a quick informational guide to the community on how to properly mineral split. It is also a way for me (if I may) to continue to encourage TLers to watch my short, 10 minute show that I stream for TL readers (meaning not advertised or popularized on any other site)! The premise of the mineral split in this context is to get the most minerals out of the split from the opening seconds of the game. In episode 4 I detail this in the beginning, so if you were to watch anything it would be the beginning of episode 4, where I go through the basics of my (Sidereal) particular mineral split vs a very common 3-3 split. It is a 4-1-1 split, where after the quick clicks each starting probe is on the closest mineral patches to the nexus, in the case of my race. By doing this, even though all players, especially in mirror matches, are identical up until usually the 12th probe, this split will put you ahead against other versions of mineral splitting. I believe this split is superior. I, however, am a simple human and am capable of showing humility, so if it is actually not as good, please let me know! Also, if there are any comments (positive or negative, don't matta), please leave them on the TL thread for the stream! I am constantly reviewing it and trying to get better, so any feedback is def appreciated! Link to stream: http://www.livestream.com/tehempLink directly to episode 3, the beginning of which is the mineral split discussion: http://www.livestream.com/tehemp/video?clipId=pla_8a9de969-ae7d-4e91-ab12-9519ada279d2Link to stream thread: http://www.teamliquid.net/forum/viewmessage.php?topic_id=185138Happy hunting, jeee ouse.
Didnt finish watching the video, but there are several differences in both of your splits which give you the impression that yours is better.
1) Income fluctuates rapidly, the actual difference between splitting and not is down to like... milliseconds. If you actually care about a 30 mineral burst that is milliseconds different (and it is just a burst, since only the first 6 workers are effected, any after that are the same)
2) The reason your probes are coming out slightly faster is not because of income, but because he sent workers down then built his probe while you built your probe then sent your workers.
3) He didnt split, he did a select all -> send to minerals. The split you thought you saw was actually just him selecting all of his probes again.
|
Just asking a question from someone else's comment - I notice basically everyone builds a worker before sending the others to the mineral line. That guy says that's "proven" but I've always sent workers to patch BEFORE building because I really can't understand how delaying 6 workers for 1/2 second is better than delaying one.
Is it true? oO
|
I think people should use better terminology for splits.
You say 4-1-1, I think 6-1-1 (or could be a 6-2-1) is better notation though. 4-1-1 shows at a glance the different groups, but it doesn't elaborate on how they were attained.
Typical 3-3 is I think a 6-3? at least that's how I do my "3-3", if I make a mistake it usually becomes a 6-4 or 6-2.
I think it's pretty important to distinguish whether you leave some workers sitting at the start, or don't. Even if one says that all (good) splits would initially start with 6, and that it should be implied, it still doesn't explain things as logically or accurately.
|
I like the way this split feels, it's easier to do than the 3-3 split because half the time I end up doing a 2-4. It might be slightly better but as long as it's not worse I'll probably be doing this from now on.
edit: show is pretty cool actually
|
On February 02 2011 19:49 NeXiLe wrote:
Just asking a question from someone else's comment - I notice basically everyone builds a worker before sending the others to the mineral line. That guy says that's "proven" but I've always sent workers to patch BEFORE building because I really can't understand how delaying 6 workers for 1/2 second is better than delaying one.
Is it true? oO
I've wondered the same but I think most of the people are just not smart enough to realise it. We're on planet earth you know.
They could have a little argument though that sending+splitting workers takes more time than starting the production of an scv, so you basically have the shortest delay in between if you start the worker first, but this delay has to be 6 times bigger than the delay the splitting gives to make up for it.
|
On February 02 2011 19:49 NeXiLe wrote:
Just asking a question from someone else's comment - I notice basically everyone builds a worker before sending the others to the mineral line. That guy says that's "proven" but I've always sent workers to patch BEFORE building because I really can't understand how delaying 6 workers for 1/2 second is better than delaying one.
Is it true? oO
It shouldn't take more than milliseconds to "start". As the game loads I (left-handed) have my thumb on ctrl, my forefinger on q(build probe in grid layout), and my middle finger on F1. I put my mouse cursor in the exact center of my screen. As soon as the game starts i click and hit q, then ctrl+f1 milliseconds afterwards, then right click on the minerals. I usually do a 3/3 split while they are en-route, just because i can. I don't see how it really matters either way, but it literally takes less than a half second to make the worker first.
|
On February 02 2011 19:35 furymonkey wrote:
To be honest, no matter how you split your worker, it is so insignificant in terms of gameplay.
It's all about warming your hand, using that skill throughout the game like marine split against banglings or high templars or siege tanks. What so good if you can do a 1-1-1-1-1-1 split at the start when your marine blob dies to banglings and lose the game?
Lol. Bad example imo...
If you can do a 1-1-1-1-1-1 worker split, am pretty sure you wont have a prob doing a marine split against banes. 
|
I remember a video on TL from Berkeley University about binomial distribution and mining, the conclusion was that (in BW) you gain more minerals if you send your workers to the patches at the edges of mineral line. I don't know if it's the same in SC2.
|
Canada28 Posts
Anyone who thinks Splitting doesn't get them extra minerals simply doesn't have the raw apm to do it. To the people who do it they will notice their timings are slightly off. I can build SCVs and that first supply depot without getting blocked by splitting properly. If it's not done right I get blocked. There is an obvious advantage early game.
For example, say all your workers went to different patches on the fastest route as quick as possible right at the start with inhuman speed, you seriously don't think that would net you faster minerals? It's the same concept and how well you do it is directly effected by how fast your hands can move.
|
On February 02 2011 17:17 SichuanPanda wrote:Show nested quote +On February 02 2011 16:57 ch33psh33p wrote:On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Way to read the post bro. Or even the topic title. I did and the poster was proposing a way of faster worker splitting and/or how to start ones economy as fast as possible, (as stated in the post before mine (the one you quoted), he used a flawed method of obtaining data) not to mention that in addition to that its been proven worker splitting slows you down, thus making this whole topic redundant. Perhaps you should re-read the topic to make sure you retained everything, bro.
Everyone worker splits AFTER building the first probe, it isn't even a question here.
|
Canada28 Posts
On February 02 2011 18:15 [F_]aths wrote:Show nested quote +On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Is there a link to a thread proving I should build a worker before I send my six workers to mine? It should be better to send the 6 workers first, because you get 6 time the income of a single worker. (I am assuming that I am no perfect player, so I have to decide if I should delay the first worker to build or the six given workers to work.)
I send my 6 workers, build an SCV and split before they reach the patches. It always nets me at least 10 faster minerals that I would have non-split allowing me to get that depot without supply blocking myself.
|
During beta one guy make tons of experiment with this... and his conclusion was: theres no reason to split your worker in SC2.
Worst... you could misclick and now you could be behind...
Instead of trying to split the worker... just type "gl hf".
|
On February 02 2011 23:15 Figgy20000 wrote:Show nested quote +On February 02 2011 18:15 [F_]aths wrote:On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Is there a link to a thread proving I should build a worker before I send my six workers to mine? It should be better to send the 6 workers first, because you get 6 time the income of a single worker. (I am assuming that I am no perfect player, so I have to decide if I should delay the first worker to build or the six given workers to work.) I send my 6 workers, build an SCV and split before they reach the patches. It always nets me at least 10 faster minerals that I would have non-split allowing me to get that depot without supply blocking myself.
Always better to build, then send workers. No question whatsoever.
|
On February 02 2011 23:22 ch33psh33p wrote:
Always better to build, then send workers. No question whatsoever.
I understand this is common practice, and I even do it myself, but in theory it makes less sense. As a couple people stated above, having 6 start first should yield faster income. There's even usually a half second delay trying to create your second worker because you are waiting on the 50 minerals. This won't happen if you send workers first. Either way though, no method really makes enough of a difference.
Also, I remember people testing this in beta, but some people seem to think you can run the test 3 or 4 times and use those as results. In order for the test to be statistically relevant, you would need to run it many many times under the same exact conditions.
|
On February 03 2011 00:15 nickwtf wrote:Show nested quote +On February 02 2011 23:22 ch33psh33p wrote:
Always better to build, then send workers. No question whatsoever. I understand this is common practice, and I even do it myself, but in theory it makes less sense. As a couple people stated above, having 6 start first should yield faster income. There's even usually a half second delay trying to create your second worker because you are waiting on the 50 minerals. This won't happen if you send workers first. Either way though, no method really makes enough of a difference. Also, I remember people testing this in beta, but some people seem to think you can run the test 3 or 4 times and use those as results. In order for the test to be statistically relevant, you would need to run it many many times under the same exact conditions.
I build worker first because I think that the half a second that six workers will mine is less then every single worker made from the main having half a second more mining time. Seems logical to me @.@
My second worker is never late, though, so I guess just speed matters more then either option.
|
On February 03 2011 00:21 lyAsakura wrote:
I build worker first because I think that the half a second that six workers will mine is less then every single worker made from the main having half a second more mining time. Seems logical to me @.@
My second worker is never late, though, so I guess just speed matters more then either option.
That half second earlier worker does not compound over time. Its not like when your 18th worker is
coming out, you have (18 workers times half second) more minerals. You can only count that half second worker advantage once.
edit: disregard the previous paragraph as I just realized it was blatantly wrong
Like I said though, I still do the traditional 33 split with worker build first. Its easy as Terran because you just hold S as the map finishes loading and then you can just left click the CC and it starts.
|
the moment the probe you used to build the pylon stops mining, your income drops like 40 minerals at 5:44... and you lose that 'lead'.
have u not seen idra split before? i think his is the 'most optimal' (i doubt it really gives him any noticeable gain). but he does the 3/3 split then while half his probes are heading to a mineral patch, he splits those three to individual patches.
your op however, doesnt really give anything to give even a noticeable advantage -_-
|
On February 03 2011 00:15 nickwtf wrote:Show nested quote +On February 02 2011 23:22 ch33psh33p wrote:
Always better to build, then send workers. No question whatsoever. I understand this is common practice, and I even do it myself, but in theory it makes less sense.
No.
It is faster to build your worker before you split your workers, because you are build time limited, NOT mineral limited; ie. you have more than 50 minerals when your first probe is done building. This means you need to mitigate build time as much as possible, rather than mineral income.
|
Everoynes saying "5/10 minerals isnt a big difference" well when you place to players at even ranking, high masters (or mid like myself) and you need your buildings ina mirror MU to be SLIGHTLY faster, thus giving u just the slightest lead, it's a difference. Not big, but still there.
By having that extra 5/10 minerals, your sup depo starts faster, and then everyone evolves from there just a bit faster.
Or so I've found, anyway... chance of misclick is like -1000 if you can handle "boxclicksmallerboxclick" if not just quit lol.
|
On February 02 2011 23:22 ch33psh33p wrote:Show nested quote +On February 02 2011 23:15 Figgy20000 wrote:On February 02 2011 18:15 [F_]aths wrote:On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Is there a link to a thread proving I should build a worker before I send my six workers to mine? It should be better to send the 6 workers first, because you get 6 time the income of a single worker. (I am assuming that I am no perfect player, so I have to decide if I should delay the first worker to build or the six given workers to work.) I send my 6 workers, build an SCV and split before they reach the patches. It always nets me at least 10 faster minerals that I would have non-split allowing me to get that depot without supply blocking myself. Always better to build, then send workers. No question whatsoever.
Why no question whatsoever? You don't know how to defend your flawed "argument"?
|
On February 03 2011 01:37 Lythox wrote:Show nested quote +On February 02 2011 23:22 ch33psh33p wrote:On February 02 2011 23:15 Figgy20000 wrote:On February 02 2011 18:15 [F_]aths wrote:On February 02 2011 16:54 SichuanPanda wrote:
It was mathematically proven over 1000s of iterations in beta - build workers first, send all to closest patch is the fastest way to start ones economy. Is there a link to a thread proving I should build a worker before I send my six workers to mine? It should be better to send the 6 workers first, because you get 6 time the income of a single worker. (I am assuming that I am no perfect player, so I have to decide if I should delay the first worker to build or the six given workers to work.) I send my 6 workers, build an SCV and split before they reach the patches. It always nets me at least 10 faster minerals that I would have non-split allowing me to get that depot without supply blocking myself. Always better to build, then send workers. No question whatsoever. Why no question whatsoever? You don't know how to defend your flawed "argument"?
I just explained why there is no question what so ever. Unless I'm mistaken, it's literally impossible for him to be wrong.
On February 03 2011 01:19 MERLIN. wrote:
Everoynes saying "5/10 minerals isnt a big difference" well when you place to players at even ranking, high masters (or mid like myself) and you need your buildings ina mirror MU to be SLIGHTLY faster, thus giving u just the slightest lead, it's a difference. Not big, but still there.
By having that extra 5/10 minerals, your sup depo starts faster, and then everyone evolves from there just a bit faster.
Or so I've found, anyway... chance of misclick is like -1000 if you can handle "boxclicksmallerboxclick" if not just quit lol.
Early advantages like this don't expand over time. If you're .5 seconds ahead at 10 seconds into the game, and you both go neck in neck from there, then you'll be at .5 seconds ahead at 30 minutes into the game.
|
One thing that I've never seen discussed with mineral splitting is Zerg.
How do you properly get pairs on close minerals and stuff, I can understand how to pair up as terran/protoss where you can focus and determine which patch you can pair on, but as zerg it's hard when you have 2-3 drones in production and can't really focus on them all. So if there's some trick to it I'd love to know.
More importantly though, how much of an effect is it to send your drones away from your larva vs towards. The initial drones are always close to all of the mineral patches, but the additional drones are only close to patches near the bottom of the hatchery. Is it a benefit to send the initial drones to the furthest away mineral patches from the larva (the patches at the top) so the additional drones can take patches closer to where they spawn?
|
Say to build a car, I need 500 dollars, and 10 mechanics. Right now, I have 500 dollars, but only 9 mechanics. Should I get an extra 50 dollars, or hire one more mechanic?
If I have 550 dollars, but only 9 mechanics, what good will that do me? I still can't build the car because I don't have enough mechanics.
Thus, I need to hire a 10th mechanic, and not worry about the extra money.
The same goes for building probes: If you split correctly, you WILL have 50 minerals in time to build probe #8 BEFORE probe #7 is done being built, so you're waiting for probe #7 to complete before you can begin #8, meaning you're being slowed by build times, NOT income.
|
On February 03 2011 01:40 Logo wrote:
One thing that I've never seen discussed with mineral splitting is Zerg.
How do you properly get pairs on close minerals and stuff, I can understand how to pair up as terran/protoss where you can focus and determine which patch you can pair on, but as zerg it's hard when you have 2-3 drones in production and can't really focus on them all. So if there's some trick to it I'd love to know. Set the rally point of each egg if you need to.
More importantly though, how much of an effect is it to send your drones away from your larva vs towards. The initial drones are always close to all of the mineral patches, but the additional drones are only close to patches near the bottom of the hatchery. Is it a benefit to send the initial drones to the furthest away mineral patches from the larva (the patches at the top) so the additional drones can take patches closer to where they spawn?
No, because income between any two drones is equal. If you get 150% income from close patches and 100% from far patches, it makes sense to get that 150% asap, rather than waiting to let your next batch of drones get it. The increased income from the start might allow you to get out your next drone a tiny bit faster than if you had waited to let the new drone get the close patches.
This is kind of like maynarding. It makes sense to saturate immediately, rather than just changing rally points so that new workers go to the expansion. This way, you get the increased income immediately, rather than later. The sooner, the better.
|
On February 03 2011 01:46 Buddhist wrote:
Set the rally point of each egg if you need to.
Well that I understand, but I meant more of being able to accurately rally each egg to a patch that will create a pair vs the drone going to the patch then finding another patch.
No, because income between any two drones is equal. If you get 150% income from close patches and 100% from far patches, it makes sense to get that 150% asap, rather than waiting to let your next batch of drones get it. The increased income from the start might allow you to get out your next drone a tiny bit faster than if you had waited to let the new drone get the close patches.
This is kind of like maynarding. It makes sense to saturate immediately, rather than just changing rally points so that new workers go to the expansion. This way, you get the increased income immediately, rather than later. The sooner, the better.
Yeah I agree near patches > being far away from the eggs, but usually you start with more drones than there are close patches, so the two methods aren't exclusive.
|
Don't know if it's been said but TLO does 4-1-1 split, and to any naysayers who think for some terrible reason that having all your workers on the closest patches to start with yields no advantage... well don't bother posting because your view has been presented but this thread is about improving not naysaying
|
On February 03 2011 02:00 Logo wrote:Show nested quote +On February 03 2011 01:46 Buddhist wrote:
Set the rally point of each egg if you need to.
Well that I understand, but I meant more of being able to accurately rally each egg to a patch that will create a pair vs the drone going to the patch then finding another patch. Show nested quote +
No, because income between any two drones is equal. If you get 150% income from close patches and 100% from far patches, it makes sense to get that 150% asap, rather than waiting to let your next batch of drones get it. The increased income from the start might allow you to get out your next drone a tiny bit faster than if you had waited to let the new drone get the close patches.
This is kind of like maynarding. It makes sense to saturate immediately, rather than just changing rally points so that new workers go to the expansion. This way, you get the increased income immediately, rather than later. The sooner, the better.
Yeah I agree near patches > being far away from the eggs, but usually you start with more drones than there are close patches, so the two methods aren't exclusive.
It has to do with the time that the worker reaches the mineral. If there's already a worker on the patch, then it will automatically go to a different patch.
|
I dont get why people keep discussing this, the difference is so negligible no matter how you split. Your time and effort would be better spent doing other stuff -.-
|
Difference is very small but very noticeable if you split, how you split, etc...
I like to keep my timings in the early game very concise where I'm able to build a pylon and a probe and once my 10th probe is done, my 11th one starts instead of waiting for the pylon to finish.
I do the 6-3 split but I'll try the F1+Stacking as I've seen Jinro and TLO do it and should be optimal in the future
|
good insight, it does not really net people any huge advantage, but I thought it would be worth discussing as I have been doing it for a while now and wanted to know what peeps thought about it.
Also, I read that thread about splitting when I was in the beta and knew its contents before I posted, even though I never really thought it was explored to the extent that it could have been. I like being fine tuned in my game and I thought there were others who would be interested in getting this low to details that yes, in the grand scheme of the game, don't really matter. 
|
On February 03 2011 02:00 Logo wrote:Show nested quote +On February 03 2011 01:46 Buddhist wrote:
Set the rally point of each egg if you need to.
Well that I understand, but I meant more of being able to accurately rally each egg to a patch that will create a pair vs the drone going to the patch then finding another patch.
There's some timing that I honestly don't know and probably should.
Wouldn't be hard to figure out, lol. Like, maybe you send the new worker to a patch where another worker has just started mining, so by the time it gets there, the first one is returning.
Someone around here knows the timing.
|
Build->Split vs Split->Build has been argued before, and I posted something of a mathematical argument on it:
Paraphrased:
Let us define variables:
P = # of Workers
K = Minerals Per Second for 1 Worker
Thus, let Rm = PK, where Rm is the rate at which minerals are coming in. This is measured, obviously, in minerals per second.
Thus, let us have Mt = Rm*T = Total Minerals, where T is time elapsed in seconds.
Let us assume that a player can do "Build->Split" and "Split->Build" in precisely the same time (not actually true, in my experience, which changes this significantly).
Given the nature of the universe, Build and Split cannot be done simultaneously. Thus, no matter which option we choose, there will be a delay between the first action and the second action. Let us call this delay "d".
Thus, for a Build->Split, this results in an offset on the start of mining:
Mtb =Rm*(T-d)
If we build later, there is no delay:
Mts =Rm*T
Having established this obscenely oversimplified model:
Workers build in 17 seconds, iirc. Thus, with constant production, P increases every 17 seconds.
If we build first, then split, our 7th worker begins mining at T = 17. At this point, Mineral Totals are:
Mtb = RM(17-d) = 17RM - 17d
Mts = 17Rm
Thus send first leads in total minerals by an amount that scales with the size of "d", or the delay involved in sending the workers in a Build->Split.
However, if we Split->Build, d applies to the build time of the 7th probe, so the 7th probe will not begin mining until T = 17 + d
Thus there is an interval equal in time to d every 17 seconds where Build->Split has a mining rate advantage of K. Thus, every 17 seconds Mtb "catches up" by Kd.
It is obvious that Split->Build has an advantage in total minerals early on. Thus we need to know when Build->Split breaks even with Split->Build:
Split->Build has an initial advantage equal to 6dK, as it starts 6 workers mining d seconds before Build->Split does. Thus Build->Split needs 6 probes to be build to break even (although there is a very minor compounding benefit early on, so perhaps 5.9 or 5.8dK would be sufficient).
Thus, at 12 probes, Build->Split would begin to overtake Split->Build.
However, the supply mechanic creates a significant kink in this. For both T and P (I'm not confident that this model accurately handles Zerg worker mechanics), this occurs before Build->Split overtakes Split->Build, so Split->Build will drop the supply depot/pylon first. However, given that d is astonishingly small for most players, nearly any delay in the placement of the depot/pylon will likely wipe out the advantage.
Thus the crucial point is that one is able to constantly produce workers without mistakes up to 12 workers. This suggests that Build->Split is the better option for optimal play.
However, again, in the case of Protoss (I am not certain on Terran), a 9pylon is impossible to manage without cutting probes. I do not know how to account for this. This could conceivably make Split->Build the better option.
Moreover, this model assumes a continuous function for minerals returned per worker per unit of time, and this is not how it works. The stepwise nature of mining could also conceivably ruin conclusions drawn from this.
|
On February 03 2011 03:18 kzn wrote:
...
Thus there is an interval equal in time to d every 17 seconds where Build->Split has a mining rate advantage of K. Thus, every 17 seconds Mtb "catches up" by Kd.
It is obvious that Split->Build has an advantage in total minerals early on. Thus we need to know when Build->Split breaks even with Split->Build:
...
I'm sure you understand the math a lot better than I, so I'm not arguing against this, but can you explain this in a little more detail? I just don't understand why you basically get the time advantage of d compounded every time a new worker comes out. If you (build -> split) and I (split -> build), you are always only one worker ahead of me for a time duration of d. If you have 12 workers, you are saying you would count d six times (for the 6 you produced)?
I see that 6 workers will obviously produce more minerals in d time than one, but when you get your 6th worker, I still have 5 producing during d, so its still a 1 worker lead for a duration of d...
I don't pretend to be any kind of math expert, so if I'm missing something obvious, please just explain it. I'm just confused about the d advantage being compounded. Thanks
|
Honestly, the only time I put any effort into splitting is when the mineral patches are laid out in a 3-2-3 pattern like this:
![[image loading]](http://i.imgur.com/At0EN.png)
You build worker first and then quickly split them to the clusters of 3/3. All 6 workers will return twice before your first worker is done, unlike other splits where you usually have a quarter of a second delay before the second one builds. That's literally the only advantage that splitting will ever give you... Building another worker 1/4 second earlier.
It's almost so trivial as to be meaningless. It's basically the same as APM spam: if the first minute of the game wasn't so hopelessly boring then I wouldn't bother with it at all.
|
On February 03 2011 04:25 nickwtf wrote:Show nested quote +On February 03 2011 03:18 kzn wrote:
...
Thus there is an interval equal in time to d every 17 seconds where Build->Split has a mining rate advantage of K. Thus, every 17 seconds Mtb "catches up" by Kd.
It is obvious that Split->Build has an advantage in total minerals early on. Thus we need to know when Build->Split breaks even with Split->Build:
...
I'm sure you understand the math a lot better than I, so I'm not arguing against this, but can you explain this in a little more detail? I just don't understand why you basically get the time advantage of d compounded every time a new worker comes out. If you (build -> split) and I (split -> build), you are always only one worker ahead of me for a time duration of d. If you have 12 workers, you are saying you would count d six times (for the 6 you produced)? I see that 6 workers will obviously produce more minerals in d time than one, but when you get your 6th worker, I still have 5 producing during d, so its still a 1 worker lead for a duration of d... I don't pretend to be any kind of math expert, so if I'm missing something obvious, please just explain it. I'm just confused about the d advantage being compounded. Thanks
In strict terms this isn't a compounding advantage, but it works like this:
Build->Split:
T = 0, First Worker Starts
T = d, Initial Workers Mining
T = 17, First Worker Completes, 7 Workers Mining
T = 34, Second Worker Completes, 8 Workers Mining
and so forth
Split->Build:
T = 0, Initial Workers Mining
T = d, First Worker Starts
T = 17 + d, First Worker Completes, 7 Workers Mining
T = 17 + 17 + d, Second Worker Completes, 8 Workers Mining
And so forth.
Because you can only produce one worker at a time, if you start building later and everything is constantly producing, your final worker will finish later, by the same delay.
Basically, you start your worker d seconds later than mine, so I have a period of d seconds where I have 1 more worker than you every time one of my workers completes, until something stalls a worker completion. Every time this occurs, I get dK minerals as a result of that.
It doesn't compound, its not like there's 2d delay for the 8th worker, but worker production is bottlenecked because you can only produce one at a time.
[edit] Might have misunderstood you. The reason its 6dK by 12 workers is because I had 1dK for every probe 7->12
|
I didn't mention that this is done on specific maps actually, and thank you sevia for pointing that out. XNC, Scrap, Steppes, Metal, are maps that I know have the 3-2-3 mineral formation, which is where the 6-1-1 split is superior to other mineral splits. This is based on the irrefutable fact that all 6 probes are on the closest mineral patches, and that mineral patch distances are not all equal. Therefore, if you have 4 on the closest and 2 on the farther ones, your income will be logically lower than a player who has their probes on the closest mineral patches, it's simple math.
The 6-1-1 split works because of the natural spread the mining AI makes when 6 workers target the middle mineral patch. On every 3-2-3 mineral formation, if you target the middle mineral patch, and select the end probes to mine the outermost mineral patches, your 4 remaining probes will select the close mineral patches.
If you don't care about the extremely minor advantage this gives you, don't do it. However, not caring and trying to refute factual evidence that at the smallest degree that this split is better (however minor that advantage is) than other splits are different things.
|
On February 03 2011 06:10 tehemperorer wrote:
I didn't mention that this is done on specific maps actually, and thank you sevia for pointing that out. XNC, Scrap, Steppes, Metal, are maps that I know have the 3-2-3 mineral formation, which is where the 6-1-1 split is superior to other mineral splits. This is based on the irrefutable fact that all 6 probes are on the closest mineral patches, and that mineral patch distances are not all equal. Therefore, if you have 4 on the closest and 2 on the farther ones, your income will be logically lower than a player who has their probes on the closest mineral patches, it's simple math.
The 6-1-1 split works because of the natural spread the mining AI makes when 6 workers target the middle mineral patch. On every 3-2-3 mineral formation, if you target the middle mineral patch, and select the end probes to mine the outermost mineral patches, your 4 remaining probes will select the close mineral patches.
If you don't care about the extremely minor advantage this gives you, don't do it. However, not caring and trying to refute factual evidence that at the smallest degree that this split is better (however minor that advantage is) than other splits are different things.
It is insufficient to argue this purely on theory. We are simplifying things everywhere, including assumptions about probe AI, assumptions about the nature of ones mineral count as a function of time and workers, and all sorts of other stuff.
If you can do said splits and actually show an increase in income, fine. But "logically, it must be this way" is not actually the same as factual. I've tried a number of times to find a benefit from splitting in tests and I cannot.
|
On February 03 2011 05:31 kzn wrote:
Might have misunderstood you. The reason its 6dK by 12 workers is because I had 1dK for every probe 7->12
Alright it makes sense to me now. What each new worker is making in d time is additive, but d itself is not. I was thinking about this all wrong. all wrong...
Thanks
|
On February 03 2011 06:45 kzn wrote:Show nested quote +On February 03 2011 06:10 tehemperorer wrote:
I didn't mention that this is done on specific maps actually, and thank you sevia for pointing that out. XNC, Scrap, Steppes, Metal, are maps that I know have the 3-2-3 mineral formation, which is where the 6-1-1 split is superior to other mineral splits. This is based on the irrefutable fact that all 6 probes are on the closest mineral patches, and that mineral patch distances are not all equal. Therefore, if you have 4 on the closest and 2 on the farther ones, your income will be logically lower than a player who has their probes on the closest mineral patches, it's simple math.
The 6-1-1 split works because of the natural spread the mining AI makes when 6 workers target the middle mineral patch. On every 3-2-3 mineral formation, if you target the middle mineral patch, and select the end probes to mine the outermost mineral patches, your 4 remaining probes will select the close mineral patches.
If you don't care about the extremely minor advantage this gives you, don't do it. However, not caring and trying to refute factual evidence that at the smallest degree that this split is better (however minor that advantage is) than other splits are different things. It is insufficient to argue this purely on theory. We are simplifying things everywhere, including assumptions about probe AI, assumptions about the nature of ones mineral count as a function of time and workers, and all sorts of other stuff. If you can do said splits and actually show an increase in income, fine. But "logically, it must be this way" is not actually the same as factual. I've tried a number of times to find a benefit from splitting in tests and I cannot.
Fair enough, when I get a sec I will try my best to organize a controlled approach to testing my theory. I'm not sure at this point what it will entail, but I will think about it to make sure it is direct and without holes.
|
|
|
|
|
|
EPT
