|
Hey guys,
Just as the name suggests I'm looking for mapmakers who are willing to have their work as part of the map pool for a free to enter tournament. There will be no compensation for the submitted maps. There are also some stipulations that entries must follow to be considered. But first some background:
"Breadth of Gameplay" Reworked
Now that you've read that, lets continue.
I'm bad at making maps, and I feel that I could better utilize my time by utilizing the competent designers out there already. There will be at least one tournament, and if that goes well further tournaments may be held with maps from both the ladder and indie developers. If you would rather wait until we reach that point, please note that I will favour the mapmakers who helped move this project along in the early stages.
So how can you help?
There are two things I want to focus on right now. One of them is acquiring data from each of the ladder maps to achieve an accurate "standard" for the mineral collection rate per base of 16 workers. I have done Cloud Kingdom LE as an example. Listed below is the data I collected.
--------------------------------------------------------------------------------------------------------------
Main (bottom), Natural, + Third:
At 9:20 minutes -> 4389 minerals
At 16:30 minutes -> 18734 minerals
990 seconds - 560 seconds = 430 seconds of mining on three bases
18734 - 4389 = 14345 minerals off of three bases at 16 workers per base (two per mineral patch, no wandering)
therefor 14345/430 = 33.36 minerals per second
14345/(430*3) = 11.12 minerals per base per second
--------------------------------------------------------------------------------------------------------------
Third (top)
At 20:10 minutes -> 18304 minerals
At 24:00 minutes -> 20849 minerals
1440 - 1210 seconds = 230 seconds
20849 - 18304 minerals = 2545 minerals
therefor 2545/230 = 11.07 minerals per second
--------------------------------------------------------------------------------------------------------------
Natural (Top)
At 25:01 minutes -> 21349 minerals
At 31:30 minutes -> 25744 minerals
1890 - 1501 seconds = 389 seconds
25744 - 21349 minerals = 4395 minerals
therefor 4395/389 = 11.30 minerals per second
--------------------------------------------------------------------------------------------------------------
Main (top)
at 33:46 minutes -> 26429 minerals
at 37:40 minutes -> 28994 minerals
2260 - 2026 = 234 seconds
28994 - 26429 = 2565 minerals
therefor 2565/234 = 10.96 minerals per second
--------------------------------------------------------------------------------------------------------------
It's rather simple, unfortunately it is tedious. It is important that each mineral patch only has two workers mining it. There are to be no wandering workers. Please do not spend any resources while doing the measurement. As well, please save the replay file for easier data collection and review. See this link for the Cloud Kingdom LE example.
The second thing that needs to be done to move this project forward is to adjust the mineral fields so that each base of 16 workers mines at a lower efficiency. This will be done by placing:
- four mineral fields such that, when mined by two workers, the maximum resource collection rate occurs (aka Tier 1)
- two mineral fields such that, when mined by two workers, produce slightly less than the maximum resource collection rate (aka Tier 2)
- two mineral fields such that, when mined by two workers, produce much less than the maximum resource collection rate (aka Tier 3)
This second phase of the project will occur after we have collected sufficient data from ladder maps to get a clear understanding of what "standard" resource collection rates are. The purpose of the first part is such that we can design mineral patch layouts so that the shift from "standard" to "tiered" isn't so dramatic. Perhaps the tiered system will operate at 80% of a standard base, both utilizing 16 workers (2 per patch).
So, is anyone willing to help?
|
I have done this experiment before, in preparation to launch an effort much like you have. I never found the opportune moment to do so, but I wish you the best. This is far and away the easiest and best way to fundamentally improve SC2. One prays it catches on, but the general lack of effect that community mapping has on the competitive scene tempers hope.
In my mind, the 16 worker output of a tiered base isn't nearly as important as the falloff in worker efficiency, which is the real goal as per Lalush's old thread, and the main thrust of the scheme. (Incidentally, he prefers SC2BW's custom mining mechanics instead of forced inefficiency of mineral placement. I support mineral placement because it doesn't require a mod.) The target of 16 is no longer meaningful, because every player will have to make a decision about the added value of additional workers given the situation in game. If they want full saturation, they can do that with 24-26 workers on a tiered base that has 2 patches each at distances 3, 4, 5, and 6 squares away. A patch 6 squares away is just shy of saturated by 3 workers. So, you need two workers for the close patches (3 squares), and three for the rest, which makes 22, plus more for any bouncing and the smidgen of saturation you don't yet have.
In the current metagame which until recently relied on macro friendly maps, players commonly overproduced workers in preparation for the next expansion anyway, temporarily oversaturating. So I don't think you'd have to change your habits much, other than allocating APM to pairing workers on close patches and prioritizing mining of closer patches, even late into the game when you have a 4th and 5th or even 6th base before it's strictly necessary. Note that if you do the player a favor and arrange the mineral line a certain way, you can easily split off the most inefficient workers with one selection box and send them to a new base. Due to SC2's harvester AI, the workers on the close patches will stay paired up. On arrival at the new base, you'd have to spend time prioritizing close patches however.
As for the income rates, I don't think you have to make meticulous measurements on varying mineral layouts. Suffice to say, a close patch (3 squares) is saturated by two workers, and a far patch (6 squares) is saturated by three workers. A priori, a far patch with two workers is 66% of the income of a close patch with two workers, since it's 2/3 saturated. Furthermore, the intermediate distances are similarly linearly related, easily accounted by interpolating.
If a player added harvesters to their mineral line in precise order of efficiency, the graph would diverge from the present standard SC2 mineral line after 4 workers. A tiered mineral line would make every next pair of workers approximately -12% as efficient (additive) after the first 4.
As for sweeping claims about deathballs, I wouldn't hold your breath. But you can certainly count on more positional play when players spread out their bases. Not to mention it would open up exciting new possibilities for mapmaking with base placement and resource capacity.
Do note that the immediately reduced income curve would prolong the early and midgame, and probably reduce the efficacy of "all in" builds that hit a certain saturation benchmark while getting up their linchpin tech and completing their production infrastructure on schedule to bang out units once it all lines up, e.g. FFE 7-gate blink.
Another side effect to take into consideration is that mules are fractionally more powerful because workers are less powerful.
Hope this helps. 
|
Thanks for the in-depth post! I agree with many of your statements. Actually, all of them with the exception of the deathball. I truly think that the combination of more mining bases and a higher worker count would change the deathball concept permanently. Although it seems this post didn't achieve the popularity that I hope it would, I am going to continue taking in data , move forward on adjusting maps to this tiered system, and in the near future hold tournaments (with cash prizes) to entice players to test out the maps. I will agree with you that the competitive scene not moving in the direction of tiered mineral patches will likely be the largest obstacle to overcome.
|
I'm very excited to see if this gets anywhere! At the very least it should introduce the idea to a lot more people, which should pay off once we're all over the impending shininess phase after HotS release and people go back to bemoaning "the state of SC2". 
Deathball play is sloooowly ebbing naturally as players get better at getting value from unit control besides a-move/cast spell. I definitely think tiered minerals would push this along even more, especially due to multiple engagements at different mining locations and strategic points (route intersections/towers). It would definitely reward multitasking more, which could help reward players on a separate axis from deathball orchestration. I just can't quite see it as a "cure", although I'm sure it would help.
|
More data!
Newkirk District
Main (Left)
At 7:05 -> 2710 minerals
At 3:35 -> 335 minerals
425 - 216 = 209 seconds
2710 - 335 minerals = 2375
therefor 2375/209 = 11.36 minerals per second per base (16 workers, 2 per patch)
----------------------------------------------------------------------------------------
Third (Left)
24:40 -> 13570
14:31 -> 6640
1480 - 871 = 609 seconds
13570 - 6640 = 6930 minerals
therefor 6930/609 = 11.38
----------------------------------------------------------------------------------------
Natural (Left)
8:55 -> 3420
12:50 -> 6005
770 - 535 = 235 seconds
6005 - 3420 = 2585 minerals
therefor 2585/235 = 11
----------------------------------------------------------------------------------------
|
Interesting reading. I really think new types of bases need to happen on the large stage. If you have maps with a few different types of resource allocations, it becomes more about who the smartest, most innovative, most reactive player is, and less about who can execute their memorized build order better.
GSL with Icarus is sort of testing these waters, albeit on an extremely cautious scale. I wish they would be more ballsy. Kespa was smart (even if they are assholes in other areas), they knew they could have crazy maps and people would still watch. Heck, I think they got a lot MORE numbers because of the maps. But I'm kind of getting off-topic. GL
|
More data
Ohana LE
Main (Top)
4:00 -> 410
7:40 -> 2875
460 - 240 = 220 seconds
2875 - 410 = 2465 minerals
therefor 2465/220 = 11.2 minerals per second (2 workers per patch, no wander)
-------------------------------------------------------------------------------
Natural
9:34 -> 3605
13:05 -> 5955
785 - 574 = 211 seconds
5955 - 3605 = 2350 minerals
therefor 2350/211 = 11.13 minerals per second (same as above)
-------------------------------------------------------------------------------
Third
34:51 -> 19865
30:06 -> 16745
2091 - 1806 = 285 seconds
19865 - 16745 = 3120 minerals
therefor 3120/285 = 10.9 minerals per second (same as above)
|
|
|
|
|
|
EPT
