-----

6/14/2010

So today I did a three tests using a Nexus and three workers. The method was simple, and is shown in the picture below.



The time was measured in HT counts, and shown in the form of gas. I recorded each of the time it took from the minerals until the minerals were actually gained. The location size and method used makes no matter on the type of worker. Mechanically they are the exact same size, use the exact same range, and move exactly the same. The only thing different is the graphics.

Having said that, I began recording the HT counts from gathering 1500 minerals until it ran out. It wasn't long until I discovered a pattern for each worker:

SCV: 64, 64, 64, 64, 65, 64, 61, 60, 60, 60 ...
Drone: 64, 64, 64, 64, 65, 64, 61, 60, 60, 60 ...
Probe: 64, 64, 64, 64, 65, 61, 60, 60, 63, 61 ...

SCV and Drone pattern is exactly the same, however, the Probe's was different. Very peculiar. But if we evaluate the average it all equals out to the same 313/5 for each worker, that's 62.6 HT counts/trip. This suggests that at the angle 0 radians all workers mine at exactly the same speed in a controlled environment.

In the data there are random numbers that appear at the start of the mining process, but typically the worker will fall into the pattern rather quickly. The variance of numbers were from 60 - 66, with 64 occurring the most and 66 occurring the least. If I take the average of it all I get these values:

SCV: 61.91
Drone: 62.27
Probe: 61.25

All in all, very close, and the difference is most likely due to errors having added up the numbers by hand. I will make a higher volume test that will be fully automated (it will keep track of the numbers for me) and will last longer and use 2x fastest speed.

CONCLUSION: There is no difference in mining speeds by different workers using a Nexus as a base.

Post has been edited 1 time(s), last time on May 11 2012, 12:18 pm by Sacrieur.

I did some quick testing with the attached map about mining speeds for a perfect 4-worker split on python and this is what I found:

Numbers shown are the number of deaths it took the computer (which always does a perfect 4-worker split to the same mineral patches) to accumulate 64 minerals from mining. The fact sample sizes are not the same are results of SC's "random" race selection. I was also too lazy to get more data.

12 o'clock spawn:


2 o'clock spawn:


6 o'clock spawn:



8 o'clock spawn:



GG if you end up as zerg on 6 o'clock.

Attachments:

Just a suggestion: you should not only test different centers, but move the centers so that the RIGHTMOST position of each center has the same X value. (Since different centers are different sizes, you will need to move to match wherever the Nexus' right side is). Doing this can narrow down the factor that causes a Command Center to return slower than a Hatchery. Similarily, we should move the Nexus so that its right side matches the coordinates of a Hatchery and Command Center.

We can conclude that all 3 workers are exactly the same, but you're free to test each one as you see fit.

Yes, I'll do that next probably.


I've also confirmed all three workers mine at the same speeds. After testing 1500 minerals with the automated counter I got these results:

Probe: 62.6011
SCV: 62.5851
Drone: 62.6277

These all of course round off to 62.6, meaning that any difference in them is probably caused by the beginning input, which is random. I've attached the map I used. It is fully automated, so all you have to do is click to start mining and then let it sit there while it does the counting for you. All players are their default colors. The numbers are kept on the custom scoreboard.

red = 60
blue = 61
teal = 62
purple = 63
orange = 64
brown = 65
white = 66

Attachments:

Post has been edited 3 time(s), last time on Jun 15 2010, 2:26 am by Sacrieur.

This does make testing of different mineral positions easier if all workers do mine at the same speed (even though you pointed out the mining weapon of SCV is range 0 with cooldown of 15 while everything else was 22). However, practically you will almost never mine with an SCV and return to a nexus or a drone and return to a hatchery, so it's important to test mining in ways that will actively influence how it works in melee.

I'll probably help with this, but this is Finals Week and there is one non-trivial final remaining, so I won't have all the spare time in the world...

@Aristocrat: I don't think the mining weapon applies.
@Sacrieur: When I made tests, I did Leaderboard Greed and just placed 7 Computer centers and workers, and compared their minerals after a certain time.

You can probably test in single player by having a countdown timer and using the pause trigger after the time runs out.