|
The only reason I worker split is just to get my mind and my hands ready for the game. There's something about splitting them into 3 and 3, rather than boxing them, that makes you feel more 'prepared' in a sense?
It's purely mental for me. There's such a negligible difference to me.
Also @Mazer on previous page, I always tend to make my worker first, then split, because it's easier to click the hatchery (or whatever building you have) and press a button, than to do the slightly finer motions of selecting and splitting workers. To me, it's just FASTER to do it that way.
Suppose it's preferential though!
|
I think the problem I have with spitting and stacking is that if you mess it up badly enough it can actually cost you the game so I wouldn't suggest doing these on ladder until you have it down in custom games first. Also I'm not sure how much this actually helps your econ or if it is mostly just a psychological gain most of the time. However on some maps the mineral patch are set up horrible so spitting and stacking is almost required. Also I personally like to have a little calm before the storm at the start of a game to try and relax before the click fest begins so that is the main reason why I don't split or stack unless it is required by the map.
I think also knowing when to cut probe production and all that jazz could be discussed here. Because if you are 1 basing you don't need more workers once your fully saturated and a lot of the times I will watch a replay with someone who was clearly one base play that had like 10+ extra workers just bouncing in the mineral line.
|
Low masters Terran here,
I've been stacking my scv's like this for a few months now and it makes the standard terran opening (ie. 10 depot 12rax 13refinery) alot smoother for me. When I pull this off close to perfect I miss ZERO scv training time. More specifically my depot finishes before my 11th scv so no supply block occurs and I have enough minerals to throw down my rax and my refinery and make my scv's. When I'm too lazy to do this I usually am supply blocked for a few seconds and either have to delay my rax and refinery slightly or lose about 5 seconds of scv training time. Arguably this isn't much but it makes everything feel much nicer so I prefer it. Also I am not a spammer so this gives me something to do at the start of a game .
|
I do this every game (as posted in the OP)...but I notice that maybe 40% of the time, a probe will bounce off to a different patch after 3-5 mining trips paired up. ie, it looks as though I've got things paired up perfectly, but after several mining trips, one of the probes will just decide to wander. In the worst cases, the wandering probe will steal another 'occupied' patch, causing that worker to wander...within a few seconds, all of my precious mining pairs are gone. Anybody know why?
|
On April 19 2011 23:47 Chemist391 wrote: I do this every game (as posted in the OP)...but I notice that maybe 40% of the time, a probe will bounce off to a different patch after 3-5 mining trips paired up. ie, it looks as though I've got things paired up perfectly, but after several mining trips, one of the probes will just decide to wander. In the worst cases, the wandering probe will steal another 'occupied' patch, causing that worker to wander...within a few seconds, all of my precious mining pairs are gone. Anybody know why?
I think it's because it doesn't take exactly 1 trip's worth time to return the minerals so if there are patches open they still might break off (until you have at least 1 probe on all of your other patches). The probe thinks that a naked patch will be more efficient for this reason so it take more babysitting before all patches have a worker. I'm currently doing some testing on mineral stacking shaving off time on a 4gate and my results so far have been pretty crazy. stay tuned!
|
On April 13 2011 05:32 islandman wrote: Interesting.
Does splitting in SC2 actually affect the mining capability of the player? i.e would a player that doesn't split have the same amount of minerals at a certain point in time? The second you gain 5 minerals .5 seconds faster then the other guy you are at LEAST that much ahead until something changes so yeah you can stay ahead the whole game by that amount assuming no mistakes from him or you. so if he doesnt stack his probes and you do... lets say it nets you 20 extra minerals over a 3 minute period thats 20 extra minerls you will have ahead... which effects your build order timings making them faster.. etc... its small.. but in a game as competive as sc2 gets in master league EVERYTHING matters.. I have had people just leave a game after 1 battle where they didnt break even and they knew there was no easy way to get back into the game.
|
Actually splitting gives absolutely no advantage. Experiment after experiment has been done proving the lack of advantage gained from splitting.
|
On April 21 2011 07:42 Tallbus wrote: Actually splitting gives absolutely no advantage. Experiment after experiment has been done proving the lack of advantage gained from splitting. Where is you're link? where is your evidence...
|
Here are some benchmarks for the 4gate while I'm paying very close attention to mineral stacking (ie first probe out of the nexus gets stacked and the two that are mining slightly farther patches are corrected after a few trips to line up)
0:50 first pylon 0:56 10th probe (the one that get slightly cut after you put down pylon) 1:37 gateway (on 12, probe scouts) 2:44 core (17) 3:36 wg research starts 4:07, 4:09, 4:14 additional gateways added 5:22 warpgate finishes, warpins begin 5:27, 5:27, 5:27, 5:33 additional stalkers finish (you get a little capped on gas)
I will soon post benchmarks of a 4gate where there is no split and workers are rallied without any attention to mineral stacking (first 8 probes take unique patches, rally is not changed to assist stacking)
|
Becnhmarks for 4gate without ideal split (all 6 to one) nor attentional to mineral stacking
0:50 first pylon 0:58 10th probe 1:37 gateway, scout 2:49 core (17) 3:40 wg research starts 4:14,4:15,4:19 additional gateways added 5:27 warpgate finishes, warpins begin 5:32, 5:33, 5:34, 5:36 additional stalkers finish
So, you eek out aboutt 3-5 seconds on most of these timings. Some probes start to incidentally pair up after you have about 10 probes, so the advantage you get seems to really come from the first minute or so of the game. The benchmarks that are most significant are the times that wg starts (3:36 vs 3:40), and when the last gate is added (4:14 vs 4:19) as these timings are what give the 4wg its strength. I ran each test 4 or 5 times so this research is by no means flawless but I think it's pretty clear you are rewarded for babysitting your workers.
|
Been doing this myself. Although my APM is too low to worker stack past the 15 supply or so. However, worker splitting DOES work. I usually gauge my split on whether or not there is any delay from my 1st worker to my 2nd worker. A good split means you will have your 2nd worker already queued before the 1st one is finished being build. Whenever I look at replays, my workers are always a bit faster than my opponent.
Was watching Artosis' stream earlier and couldn't help but notice the nonexistent worker micro....
|
On April 21 2011 07:43 Meldrath wrote:Show nested quote +On April 21 2011 07:42 Tallbus wrote: Actually splitting gives absolutely no advantage. Experiment after experiment has been done proving the lack of advantage gained from splitting. Where is you're link? where is your evidence...
There was a thread on this, the general conclusion was that it doesn't matter too much.
http://www.teamliquid.net/forum/viewmessage.php?topic_id=129010
|
Well to test the efficiency of this you could just play a game standard with
a) no order, no split b) splitting, no order c) splitting and order
let say saturate a base to 16 workers let them mine for 1min. and then count the Minerals to have discrete numbers of the effect. so we can evaluate the effects to Income to a specific timing.(+XX Minerals)
Don't know if the economy graph is accurate but you could get that information easily by counting how much minerals have been mined by counting the minerals patches
|
so how much mineral boost u get compared to normal splitting or no splitting?
|
I do this after reading ths thread a couple of weeks ago.
Its not gamechanging, but its definitely something that keeps your APM high without spamming mindlessly. And I just love it when my 13th drone is able to be built right as the larva pops off ( normally its around 30 minerals, with efficient mining I usually get 45 or even 50 if Im on top of my timings).
|
Looking at the timings posted, on certain maps this could give you an insurmountable lead(mostly PvP only). On scrapstation and tal'darim, when perfectly executed if you can finish warpins when the other player is just starting to warp is huge. Just one or two free hits on warping units is all you need.
|
I think the OP also ommited a point to only pull probes off far patches to mine gas so that close patches are still stacked. This will lead to roughly a 100 extra minerals mined by the time your cybercore completes. This has the largest impact in the 3 stalker robo build where usually you will have to cut probes and chrono out 3 stalkers to reach 26/26 before building a pylon to unblock your supply. You will be able to prevent a supply block if you use my method mentioned above.
I dont believe that the OP's probe split is the most efficient too. I think the most efficient method is to send 4 probes to the close mineral patch nearest to the middle of the whole field and the other 2 probes to the 2 close mineral patches on the other end, then follow on the stack from there. I usually plop my pylon down at 0:48 using this method.
|
On April 19 2011 23:47 Chemist391 wrote: I do this every game (as posted in the OP)...but I notice that maybe 40% of the time, a probe will bounce off to a different patch after 3-5 mining trips paired up. ie, it looks as though I've got things paired up perfectly, but after several mining trips, one of the probes will just decide to wander. In the worst cases, the wandering probe will steal another 'occupied' patch, causing that worker to wander...within a few seconds, all of my precious mining pairs are gone. Anybody know why?
I've noticed that this happens only on some of the maps. So, it makes me think that most "close" mineral patches will maintain the stacked mining because the patch is at the optimal distance for the mining trip to be equal with mining time. However, some maps have patches that are too close. These patches have a distance where the mining trip is less than the mining time, and thus the worker bounces off. One map that comes to mind where this frequently happens is Metalopolis on the outermost mineral patches of each starting position.
|
here is a great zerg build if you want to benchmark for mineral line micro (wzp's 3 roach speedling all in): http://www.teamliquid.net/forum/viewmessage.php?topic_id=190292
You can cut around 5 seconds in being able to produce the crucial 3 roaches by splitting, stacking workers and using only far-patch workers in the gas transfers
|
for terran, if you do this perfectly, I've found that it actually smooths out the build times significantly in the early game. When I do it right, the completion of my supply depot and the start of the 12th scv is exact which gets rid of bumps in scv production in the early game. It's a minor thing that helps you optimize build orders and times so even though it won't have a huge difference in your gameplay. It feels good and it gets the game rolling smoothly! I'd recommend doing it
|
|
|
|