Personally I never split, but I always assumed the people who do either want to look cool (always a positive) or want to prevent a weird AI glitch putting them 1-2 seconds behind. Either way by the data it doesn't matter and that kind of upsets me.

woolly

United States56 Posts

June 02 2010 23:28 GMT

#182

I just tested as well, and can confirm that the F1 split yields 5 more minerals at the 40 sec mark. That 5 minerals isn't going to exponentially increase though, so I suppose it's more a question of preference.

roymarthyup

1442 Posts

June 02 2010 23:28 GMT

#183

the best split is moving your 6 workers to one half of your 8 mineral patchs targeting the third or fourth one from the end then take the worker furthest away from reaching that mineral and select it and move it to a the closest mineral to your main on the other half.

i do this every game. seems to get out that 8th scv way faster

pyr0ma5ta

United States458 Posts

June 02 2010 23:30 GMT

#184

On June 03 2010 08:28 roymarthyup wrote:
the best split is moving your 6 workers to one half of your 8 mineral patchs targeting the third or fourth one from the end then take the worker furthest away from reaching that mineral and select it and move it to a the closest mineral to your main on the other half.

i do this every game. seems to get out that 8th scv way faster

This cool story brought to you by Bro, Inc. Back up your claim. The OP has numbers. You have a super cool story.

Lexpar

1813 Posts

June 02 2010 23:34 GMT

#185

It's nerds like you that make this community great.

Kletus

Canada580 Posts

June 02 2010 23:36 GMT

#186

Alright! I have further testing with F1. Here is what I gathered:

I originally thought that the F1 selection AI wasn't random, I was wrong. However, 7 out of the 15 games I played ended up with the scvs being selected by this pattern: top, bottom, top, bottom, etc.
About 4 times I would get: top, bottom, bottom, bottom, etc. For the rest of the games the selection seemed random.

I ran through the tests(same ones as on page 9) comparing the F1 first pattern to the F1 "click down the minerals" and the difference in negligible; the scv that carries the minerals to raise the count to 145 is just about to hand it in before the timer hits 1:00.

SkCom

Canada229 Posts

June 03 2010 00:04 GMT

#187

if nothing else, the split will at least warm us up for a good match instead of just spamming mindlessly to keep your hands warm

Bosu

United States3247 Posts

June 03 2010 00:08 GMT

#188

On June 03 2010 09:04 SkCom wrote:
if nothing else, the split will at least warm us up for a good match instead of just spamming mindlessly to keep your hands warm

lol

If there really is no difference splitting to warm up would be pretty dumb since people screw it up all the time.

ClicheGuevara

United States11 Posts

June 03 2010 00:28 GMT

#189

auto split/mine and MBS are only expanding the player base. It isn't really going to make a difference in the higher levels of play

By "expanding the player base" you mean "fixing a shitty UI".

Redmark

Canada2129 Posts

June 03 2010 00:37 GMT

#190

Either way it's really an insignificant change.
If I practiced for half an hour I could do a practically-perfect split every times, or at least perfect enough for it not to matter. It doesn't add skill, it just adds a bit of time you have to practice something trivial and useless to get to playing the actual competitive game.

NonY

8751 Posts

June 03 2010 00:37 GMT

#191

On June 03 2010 09:28 ClicheGuevara wrote:

Show nested quote +

By "expanding the player base" you mean "fixing a shitty UI".

High level players didn't have a problem with the UI. I think you are mistaken here.

ejac

United States1195 Posts

June 03 2010 00:40 GMT

#192

If I do not split, I typically have to wait like, iono, .2 seconds after the first drone hatches before I build the second, and if I do split, I build it right when the drone hatches.

Does it really matter? no. I may be gaining a grand total of .1 minerals. I could probably play 100 billion games and you'd see no connection between me splitting and what not, but I do it anyway

CauthonLuck

United States93 Posts

June 03 2010 00:43 GMT

#193

It matters, a fair amount on some maps, an insignificant amount on others.

These tests that were run don't take into account important factors like the start time on your second worker nor the distribution of optimal mineral patches. Using a 3/3 or f1 split on a map that has it's optimal patches on the ends instead of the middle areas is conclusively better than sending all 6 to any one of the patches and letting the AI take over.

Kletus

Canada580 Posts

June 03 2010 00:44 GMT

#194

On June 03 2010 09:37 Liquid`NonY wrote:

Show nested quote +

High level players didn't have a problem with the UI. I think you are mistaken here.

Agreed, heck I didn't have a problem with the UI and I was C-. Then again I'm currently playing, "warcraft: orcs and humans" and having a blast. 4 unit selection only, have to use hotkeys to move units(right click moves your view, not moves units and I actually prefer this over todays rts games), units/buildings can't be hotkeyed to number keys.

Feels good, wish I could stream it but alas my upload rate sucks.

SnuggleKittens

Australia57 Posts

June 03 2010 01:11 GMT

#195

It could be to do with how workers get minerals. I dont really know how the system worked but in WC3 workers would walk slow or fast between resources and town hall. i guess this is how it kept the income consistent. Perhaps a similar mechanism is used in SC2 whereby once you tell workers to mine a mineral patch, the game knows what the income should be and then alters the mining speed or traveling speed of the workers.

qoiN

Sweden576 Posts

June 03 2010 01:13 GMT

#196

I use a 3-2-1 split, moves all, selects 3 to move to another closer, then move one in the last group before it has to auto-split, and I earn at least a few ms. I've tested it out and it does make a small difference. I think I had a ~5-10 minerals difference while trying out different splits for 1-2 hours, god was it boring though.

InfiniteIce

United States794 Posts

June 03 2010 01:19 GMT

#197

Yea, this is about what I expected. Kinda sad that it makes NO difference whatsoever. The intensive micro components that lead to small advantages (remember, very small advantages add up, over the time period of a game...) seem to be being removed. SC2 is becoming (whether intentional or not!) less micro-requiring than a lot of people hoped for..

This is just quibbling over 5-10 minerals, but to those players with the skill and speed to do nice splits, such a miniscule 5-10 mineral reward SHOULD be allocated, I think. It's never going to break or win a game, but micro should be rewarded with small advantages over time.

Just my opinion.

Also, thanks for the charts and tests! I was hoping somebody would get around to this.

michaelhasanalias

Korea (South)1231 Posts

June 03 2010 01:27 GMT

#198

I mean honestly, the funniest part of all this is that the position you put your cursor into plus adjustment reaction time more than eclipses any mineral gain.

This was the case in BW as well....

If you juts right click all 4 workers on the same patch, you could either jump a patch (if you're that retarded) or worst case, be back 1 gather.

So the difference there was genuinely up to 24 minerals, where as this you're not down gathers, but simply slowing your total gathers by milliseconds, which are really within the margin of error for the rest of your game.

If your micro/macro is up to where milliseconds of gather made a difference, you wouldn't be worrying about it anyway.

I mean, probably 40 people have commented on how misplaced the elitism over microing workers at the start is, but I think it's worth continuing to state until they stop crying about it.

I'm sorry you guys fail at basic mathematics and timescales in your brains, but it really doesn't make a difference beyond the margin of error of even a single mis-click over the entire game.

dogabutila

United States1437 Posts

June 03 2010 01:31 GMT

#199

Some of the videos I watch of pros have them sending all to 1 patch, but splitting off 1 drone manually to a corner one.

Playguuu

United States926 Posts

June 03 2010 01:39 GMT

#200

I didn't think splits mattered so much because the pathing and ai is pretty decent. I was surprised though with the timing of the first worker built. I guess I can relax at the beginning of the game now.