|
Good Day Everyone,
While watching MikeD's stream yesterday, several of his regular viewers and myself had a discussion regarding the initial seconds of the game; is splitting your miners first better than building your initial worker first. I've done a little math and hope to hear some of your opinions as well.
On average, the player will split his workers to mineral patches so that you end up with 4 workers on close patches and 2 on far patches, so I'll assume this is true for the purpose of the following calculations. I'm also going to assume that it takes 2 in-game seconds to either split workers or build the first worker and hotkey your initial building. My last assumption is that the travel time of your initial workers will be the same no matter which path is chosen, so I will not include it in my calculations.
Build First With the above assumptions, choosing to build your worker before splitting will generate an immediate loss of roughly 8.6 minerals. (This is the amount of minerals that would be gained with the additional 2 in-game seconds with 6 workers split 4close/2far.)
Split First Choosing to split first will generate an immediate loss of 1.43 minerals and an anticipated loss of 1.43 minerals for each worker built immediately after the first. As soon as there is a cut to probes, this loss is nulified. This means; Protoss; 10 gate will lose roughly 5.7 minerals overall. (Loss over 68 in-game seconds) 12 gate will lose roughly 8.6 minerals overall. (Loss over 102 in-game seconds or 90.66 in-game seconds with 1 Chronoboost) 13 gate will lose roughly 14.3 minerals by the 16th probe. (Loss over 170 in-game seconds or 147.3 with two chronoboosts, 136 with 3 CBs) Note 1: Chronoboost will not effect the minerals lost, but if chrono causes an accidental cut, it may stop the loss chain. Terran; 15 Orbital will lose roughly 12.9 minerals overall. (Loss over 153 in-games seconds) Zerg; Larva mechanic nullifies this debate...everything I calculated says split first.
So the math seems to tell me to split first as Zerg, Build first as Terran, and that Toss should change based on their openings...10 or 12 gate = Split first, 13 gate = Build first.
I understand that we are talking about a very small amount of minerals...but I know that mirror matchups can be largely effected by such differences. Let me know your thoughts on the matter!
+ Show Spoiler +
|
Your math may be right, but your assumptions are wrong. For average player it takes more time to split your workers then to build a worker.
|
On April 30 2011 03:34 gejfsyd wrote: Your math may be right, but your assumptions are long. For average player it takes more time to split your workers then to build a worker.
The selection and first click to minerals is the more important part...splitting the second half afterwards is only beneficial to determine the average ratio of close/far patches. I do not think it would effect the numbers too much. (Change the split and it will equally effect both options anyways.)
|
Very interesting results for Zerg and Toss. I never would've thought that split or build first would depend on the race.
Even though it doesn't affect me cause I'm Terran and I already build first, I'm sure it will help out tons of Zergs and 13-gating Toss players.
|
i think its better to split first because:
_the 6 workers outmine the 1 worker ur building so you will mine more _with ctrl+f1 you can select workers so fast that you can ''gain'' 1 second.
building a worker first will hurt you in my opinion because those 6 workers will not be mining untill u put them on the minerals. and 6 workers outmine the 1 or 2 second delay on your 7th worker.
with the individual splitting you can make your builds even better and get faster hatches as zerg. splitting is the last thing you want to master though but its definately worth it. most players wont think about it but you can lose lots of minerals if your probes run between mineral patches without mining.
|
The math is flawed, there is no way it takes 2 seconds to build a worker, in fact i wouldnt be supprised if it was around 0.5 for myself. Personally i grab 6 workers, set them to the top far mineral patch, and then quickly build my worker as i pull off 3 and send them to the bottom.
|
It takes 2 seconds to start building the 7th worker? Just center your cursor during the loading screen and asap click and build, I don't see how that can ever take multiple seconds :O
On the actual topic though, I never really considered which would be better, building worker was always prioritized in bw so I just do that, and I play zerg anyway so even a small delay won't delay workers after that first one since I have larvae anyway.
|
lol everyone builds first and splits then. 100% sure that is the best way.
|
On April 30 2011 03:45 deanyo wrote: The math is flawed, there is no way it takes 2 seconds to build a worker, in fact i wouldnt be supprised if it was around 0.5 for myself. Personally i grab 6 workers, set them to the top far mineral patch, and then quickly build my worker as i pull off 3 and send them to the bottom.
In this case time will only change the base value, not the ratios or amount of time for the losses to equalize accross the two options. Keep in mind that it likely takes your brain close to 0.25 seconds to tell your hands to start...so adding that to your 0.5 seconds means that 0.75 seconds turns into 1.125 in-games seconds which is likely. (Also, I'm using averages here...)
|
I build first and always am able to start the second probe immediatly anyway, although I see what youre saying if someone were doing say a 10 gate opening or whatnot
Otherwise 100% do build first in my opinion
|
You can see how long it takes a player to build their first worker using the Build Order tab at a Score Screen. Most players take about 1 second, but the range is ordinarily between .48 seconds and a little over a second.
|
0.25 second reaction time? That's ridiculous. Please don't make a thread like this without actually looking into.
I build then split but I could see sending all, building, then splitting being the best
|
I think it slightly depends on the player. Since I have a keyboard that has macro keys, what I do is hold E to build my Probe (since I'm Protoss) and when the moment the game starts I click on the Nexus and it starts my worker. Then immediately after I hit the Nexus (and Probe starts building) my pinkie is easily able (and already positioned) to hit my G12 key which is CTRL + F1. So basically, I start building my Probe right away and already have my workers selected to send them on their way.
I'm still working on my getting the "perfect" split... But I seriously get them both started almost simultaneously that I think in my case it's better to build the probe first. In a general sense, I think it depends on how quickly you are able to do it. Before I had my keyboard I would build probe first and since my mouse was already in the vicinity of the probes, I'd box them and get them moving, and try and split them as they were getting off the start mining.
And as Blindgamer said, I've always found that I'm able to have no delay between building my second probe and the completion of the first one... So I'm personally of the opinion to build first (that and I remember seeing it a few times in some pro games that I was watching, so I just assumed it was the better way to go).
|
|
As a longtime Terran player, I build first then send workers to minerals. But really what matters most is that you get your commands in as soon as humanly possible and you select the best mineral(s) to drop your group of workers into. The thing to avoid is having one or two SCVs confused about what patch they should be mining.
My tactic is to spam S as the match is starting, and select my CC as quick as possible. Then as fast as I can, box my workers into one group and send them all to one mineral patch in the middle of the line, preferably the closest one to the CC. My "split" is to grab one of SCVs and send him to a patch away from the middle, but not all the way to the very end. (This is the SCV that would be left straggling if you sent all workers to one mineral without splitting). It took me a while to get good at selecting the single SCV without mis-clicking and screwing it all up, but the end result is I rarely rarely have stragglers left trying to find their patch to mine and everyone has their minerals back at the base in time for me to produce my next SCV with zero CC downtime.
That being said, the gain made from doing this is very small, and in a best-case scenario you will be maybe a half-second ahead of your opponent - hardly game-changing.
|
Well doesnt it make sense that zerg drone production is limited my minerals while terran and toss are limited by time?
So zerg players should splitt first.
Personally I send all workers to 1 patch->build probe->take 3 and splitt
|
I actually double-checked and graphed your math because I was unsure if the compounding would actually occur (instead of a linear loss) and turned up roughly the same results - interesting.
I can say for a fact that I've never lost a game by a razor thin margin like this, though, fortunately. I split first without splitting, then build. The lazy way.
|
I always start the game with the cursor over the hatchery, so it takes practically no time to build a worker. Basically I am moving the cursor to split whilst I click and press "S - D". Then I split, so basically I think this is the best method =)
|
On April 30 2011 03:51 Anomandaris wrote: lol everyone builds first and splits then. 100% sure that is the best way.
everyone also believed the sun rotated around the earth a few hundred years ago. you're just appealing to popular belief!
interesting thread though, i always assumed that splitting first or building first would make no difference. thanks for the contribution!
|
i knew people would start getting mad at this thread... everyone seems to have their own opinion. it seems not important but it actually is.
as for what i said about the 1-2 second delay on the probe, if u split first then build then your probe will be created at the 1-2 second of the game. depending on how you split. this means you will lose 5 minerals for the delayed probe max but the 6 workers will mine more than 5 minerals. thats why i think splitting first is better.
Edit: the reason building first could potentially be better is because you dont know where you will spawn. the mineral patches could be above the nexus or beneath the nexus. you need some time to react so if you build first you will be able to locate the mineral patches and click with maximum precision. if you are focussed when the game is loading, you can prepare to locate the mineral patches and click on them as fast as you would if you built a probe first.
Edit nr.2 : yea sure this wont matter too much but if you want to maximize the efficiency of your build by doing this kind of things its only gonna help you. isnt it better to build a nexus 5 seconds earlier? or constantly building stuff and have more?
|
|
|
|