|
On February 10 2011 02:29 Duka08 wrote:
I love this post, and the thought behind it, and especially love your second section about efficiency and the way you spell out plain and simple what all the numbers mean. Reminds me of a great discussion in a research paper after a lengthy and understandable mass of data analysis. Hope this sparks some great discussion.
I'd actually love to hear your Zerg bit, since you seem to have a great grasp on economy and macromanagement in SC2. I'd also be interested to hear what you have to say about the 300 supply cap in more detail, and especially what you think Blizzard could change otherwise (you alluded to worker AI) that you kind of shrugged off as "they'll never do," which is probably very true.
Edit: Gagnar's post above me actually reminded me of the creativity of map designers and how possible "strange" expansions might work out. Perhaps maps will be push out with fewer main base mineral patches, reducing 1-base efficiency and encouraging early expos. Or gas only expos (perhaps with only very few min. patches) somewhat nearby that would encourage tech. I don't have enough BW experience to draw correlations, but it seems to certainly be a possibility, just not in the form of "mineral only" expos because that would, as OP demonstrated, really not achieve much.
And for clarification, these "fewer mineral patch" main base map(s) could obviously be choices in a loser-picks style event by macro-style players, and consequently vetoes for stronger timing/all-in players. I'm not suggesting that it become the norm, I think it would be more interesting to have some with and some without. Map diversity is great, and something BW seemed to have a lot of even with my lacking experience.
yeah it would be awesome to hear that zerg bit, edit it in please :D
given how excellent the rest of your post is, the zerg bit wont spoil it
also the supply caps, theres something about 200 that id rather keep if possible, but enough to make 4 mining bases possible would be cool
|
lalush, fantastic write up, thank you for clarifying and confirming the suspicions of many. kudos to you sir.
|
Your analysis was incredible. Very interesting.
However, I don't agree that this effects Blizzard and balancing. They have failed miserably to balance small maps. They just seem incapable of making the difficult decisions necessary to make close positions maps balance.
You need very effective early game scouting tools for all races and much more balanced unit efficiency and economy gain. Mules are a prime example of an ability that simply doesn't balance in close positions.
Blizzard however seems much better at balancing late-game. A lot of their changes (void rays, battlecruiser, Thors etc....) help make late-game play fair.
I just think it's easier for them to look at unit stats and balance tier3 big battles. They can run their simulations with 200 supply armys and see what happens.
However this doesn't take into account the explosive macro capabilities of Zerg. Hopefully this won't turn out to be a problem but we'll have to wait and see.
|
Many have suggested a 300 food cap, but really, 230 would be good enough. It's enough to saturate a 4th, with as large an army as before. You'll be hard pressed to take a 5th before your main is mined out, on any map, lol.
I was actually pondering the other day that, perhaps for HotS, the cap could be raised to 230, and perhaps in LotV, it would be raised again to 260.
|
I see 3 solutions for this:
1) Changing the worker AI
2) Rising the supply cap, and
3) Moving the mineral patches closer to the Hatch/Nexus/CC and therefore removing 1 or 2 mineral patches
If you apply the 3rd one, all the races would saturate their bases much faster and thus there will be a need to expand. If the patches were closer to the base you could secure the same amount of income but you will be able to saturate more than 3 bases with ~75 workers.
|
On February 10 2011 08:23 Plutonium wrote:
What would you guys say about a hive tech upgrade at the roach warren that makes roaches take up one supply instead of two?
I could imagine Artosis's voice "SO MANY ROACHES" instead of "SO MANY BANELINGS"
I still believe then if you have Hive tech you may as well spend that gas in actual Hive Tech units since more roaches aren't going to be the solution to the same death balls either way. massive collosi count? massive siege tank count? more roaches simply wouldn't be the solution to that problem but I'm of course not going to say no for more options as a Zergie.
On topic. I know that the OP has stated that Zerg are so messed up to model you still at least have to try (science/maths minded here)....
I'm just wondering if someone is mad enough to also try to do this with multiple lines for Zerg based on one hatch two hatch with typical openings and queen inject times assuming they make only 2/4 zerglings for scouting purposes and the rest of the larvae for droning. This would be of course for the first 5 mins only as from then Zerg will probably be wanting to make 2-3 spines/ roaches/ lings to fight the 1st timing push. If the maps are getting bigger then cheese will be less likely (so no early zealot from main and less chance for bunker rushing) and so 14 pool 15 hatch (or 11 overpool 18 hatch) would actually be a good idea to model on assuming double queen on pool spawn with one inject one creep tumouring
I know that most of the time a zerg player will be harassed before the 5 minute mark either with a pseudo push with early units (a bunch of marines and scvs and then falling back forcing zerglings) or the double stalker push whatever early harass that come out of the gates. The point is to still have a comparison even if there are parameters set around the lines for Zerg or I could just be talking crap and everyone's gonna have a go at me.
|
Honestly, I doubt something like that could be done with zerg....ever.
Zerg has the possibility to produce multiple drones at the same time even off one hatch. I can't imagine how you could model this to be comparable with toss/terran.
|
very impressive read, thanks for work you did, and gz on 1k post 
I think Blizzard is aware of this, in fact i think gold mineral patches are in game exactly due to this "issue" because 2 base vs 3 base where 1 is gold there is great difference, 2 base play is a result of such macro mechanics system, because the difference between taking each expansion is not big enough.So i think reduce mineral patches in each map and each base are crucial to reduce 1 base / 2 base play.But then again, zerg ability to produce multiple drons at time would be of too big succes, so changing one thing in system will mean to change a lot more that one can think of.
|
In this light, it seems slightly foolish of the community to expect the metagame of SC2 to eventually evolve into something resembling that of Broodwar. You will likely never see players opt for as early of a third base as used to be the norm in Broodwar. Rather players will tend to be bottlenecked on 2 bases for longer (especially on Blizzard-sized maps).
I can say with experience with T from one recent game on Megaton (Pure Mech) that getting 2 > 4 bases (4th being a gold) during the first 5-15 mins of a game in a gradually expanding manner is what's most effective/optimal. Basically after you get your orbital, worker saturation and mule finished mining is when you gain back the minerals you invested into the Orbital (550 mins) and start going surplus.
And you should expand every time you are able to afford another expo if your still keeping up with full unit/depot production, your macro gradually escalates and explodes at a point, 3 fully saturated bases or a gold! when you go up like 100~ supply.
I think having 60-70 scvs (with mules) divided between 3 bases is the most income you should at most ever get while still maintaining an army, going for more means you have less supply for combat units.
It also shows that even if you do no damage to a Terran (Banshee, DT etc) but force him to waste scans instead of dropping a mule then that's like 240-270 mins lost for the Terran meaning that it's more beneficial to just place turrets for detection then relying on scans alone.
Doing like in BW and going for two early game expansions at the same time (like MVP did in GTSL today) just slows you down a lot as you won't have enough workers or income to keep producing units/scvs/depots all at the same time.
And seriously an orbital (550) is expensive compared to a hatchery (350), you would need to drop 2 mules to earn that back after 90 secs and scv mining time would give the bonus surplus.
|
Man, I LOVE U ! i just had best 2 hours in school, thanks !
Mule boost is crazy but you can also kill mules so i think thats the point.
|
Canada5565 Posts
I bookmarked this and will be coming back to it many times, I'm sure. Thank you for collecting all of this data, explaining it in an organized fashion, and presenting an interesting argument. I might edit this post to add a more useful comment but I'm still looking through the OP. But damn, this is good.
|
Protoss, BW, with 54 workers equally distributed on 6 bases: 18120 minerals over 5 minutes.
Terran, BW, with 54 workers confined to 3 bases: 13200 minerals over 5 minutes.
Zerg, SC2, with 54 workers equally distributed on 4, 5 or 6 bases: ~15384 minerals over 5 minutes.
Protoss, SC2, with 54 workers confined to 3 bases: 14586 minerals over 5 minutes.
Do you have the statistics for 70 workers? Thanks.
Great post btw, it is interesting stuff. Please don't make it a balance discussion though because there are many other factors in play rather than just "who can mine the most minerals", such as cost efficiency, mobility, micro..
|
Good read.
There are "some" factors I think were skimmed over that more or less leads me to disagree with the 3 base cap conclusion or rather that Blizzard did think of these factors alot more than people will give them credit.
For one, the reasonable point where your first base starts fizzing out.
There's a reason there's a heavy point of conflict around the point you get your third base. As you pointed out, it's the point where you start getting closer and closer to that saturation point while sticking to about 1/3rd supply on workers, but also the point where your first base threatens to give out. Terrans who mule, while they do get that initial burst which puts them ahead, they also threaten to burn out the first base at a quicker pace.
To me, the larger maps give me a little pause, since an easily defendable 4th or 5th base, while it may give room for players to breathe, would also mean that players can obtain that ideal 3 base income quicker, and like to mention, there is little advantage for a Zerg to get 5-6 bases as long as your opponent keeps in mind that he shouldn't confront the Zerg unless he can either wipe him out, destroy a base with little to no losses, hit his drones, or bring him below 3 bases.
That said, I do think some of the smaller maps could be up for review, however with Vote down features on maps, including many tournaments, and pro players choosing to invest time practicing maps they might end up on... I'll just say that it's hard to judge a book (talkin to you Steppes of War) that somehow can produce 2hr draws. However I do think Metal and Lost Temple should be reviewed, spawn locations shouldn't dramatically impact how a MU plays out.
300 Supply cap... I have one serious issue and one minor gripe.
Minor gripe? As a spectator, and even a player, 40-45 minute games can be entertaining... If there were signs it was a close game. If it becomes the norm because neither player feels it is to their advantage to hit the other player before they reach 300 cap... It can become pretty boring.
Serious issue? Tech. There was already a thread about how given the resources, one can provoke crashes in the other players computer. (Nydus worms most notably) More supply, means more units, which means bigger hogs on both your computers resources and your internet connection. 300 supply might as well be a nail in the coffin of anything above 1v1, and force many players to expand a ton of money to play the same game they love. Which is why it always surprises me that Day[9] advertises this solution when he will talk about how he often plays 2v2s and various other team games with his roommates from time to time.
To those that TL:dR this post, just read the last part on how I think 300 supply is not a great idea.
|
On February 10 2011 22:32 NIIINO wrote:
Man, I LOVE U ! i just had best 2 hours in school, thanks !
Mule boost is crazy but you can also kill mules so i think thats the point.
The MULE boost is not crazy. The graph of Minerals mined with respect to number of workers is extremely misleading. Given constant worker production It is not physically possible for a Terran to have equal numbers of workers as the other races and have an orbital command for MULES. An orbital command and thus a single constant MULE costs 150 minerals and 34 seconds, this is at minimum 2 workers less at any given point in time. Protoss can produce workers 22% faster with constant chronoboosting and should at all times be ahead on workers, with just 2 chronoboosts to the nexus a Protoss will be at 20 workers while the Terran is at 16 and a MULE.
The MULE only starts to have a significant impact on income after saturation comes into play, which the later graph of minerals with respect to time shows the rate of income levels out after a roughly 30 second spike for both races. This is the only thing you could consider "imbalanced" about the MULE, that it gives no diminishing returns on mining rate, thus allowing a Terran to maintain similar economic growth to the other races while expanding slower. Which the graphs in the latter part of the OP somewhat show.
The total amount of minerals mined is actually much, much closer that the graphs seem to say and is a very important factor to consider in this discussion, also the fact that Terran has a more mineral based military. Protoss and Zerg units require more gas with respect to minerals, so a mineral only comparison with taking into account the uses for those minerals is flawed and paints a misleading picture. What would a Protoss do with all those minerals? The can only make Zealots, gas is a much more valuable resource for both Zerg and Protoss compared to Terran's largely mineral focus.
With respect to the OP I think it is very insightful and provides some very good information. However it doesn't actually address anything as it only provides specific information that you can apply to a part of the equation, it doesn't complete both sides or take into account the other variables. A good source of information, but not in itself a conclusion or explanation to any problems.
|
This is really good stuff. I always wondered about the 3 base thing, not in a such a grand coherent way - only the in way of 'if I take more bases, then my army food drops to 120.' Never applied it to zerg but it makes sense.
|
|
|
Great write up. Very interesting and informative.
This really just proves how lucky(?) blizzard got in creating bw, which although not perfect is the most balanced game I have ever played. For example when I get my ass handed to me in bw it can only be my fault and not the games.
|
On February 10 2011 23:51 Kazang wrote:Show nested quote +On February 10 2011 22:32 NIIINO wrote:
Man, I LOVE U ! i just had best 2 hours in school, thanks !
Mule boost is crazy but you can also kill mules so i think thats the point. The MULE boost is not crazy. The graph of Minerals mined with respect to number of workers is extremely misleading. Given constant worker production It is not physically possible for a Terran to have equal numbers of workers as the other races and have an orbital command for MULES. An orbital command and thus a single constant MULE costs 150 minerals and 34 seconds, this is at minimum 2 workers less at any given point in time. Protoss can produce workers 22% faster with constant chronoboosting and should at all times be ahead on workers, with just 2 chronoboosts to the nexus a Protoss will be at 20 workers while the Terran is at 16 and a MULE. The MULE only starts to have a significant impact on income after saturation comes into play, which the later graph of minerals with respect to time shows the rate of income levels out after a roughly 30 second spike for both races. This is the only thing you could consider "imbalanced" about the MULE, that it gives no diminishing returns on mining rate, thus allowing a Terran to maintain similar economic growth to the other races while expanding slower. Which the graphs in the latter part of the OP somewhat show. The total amount of minerals mined is actually much, much closer that the graphs seem to say and is a very important factor to consider in this discussion, also the fact that Terran has a more mineral based military. Protoss and Zerg units require more gas with respect to minerals, so a mineral only comparison with taking into account the uses for those minerals is flawed and paints a misleading picture. What would a Protoss do with all those minerals? The can only make Zealots, gas is a much more valuable resource for both Zerg and Protoss compared to Terran's largely mineral focus. With respect to the OP I think it is very insightful and provides some very good information. However it doesn't actually address anything as it only provides specific information that you can apply to a part of the equation, it doesn't complete both sides or take into account the other variables. A good source of information, but not in itself a conclusion or explanation to any problems.
alright im going to question that 22%, got the math?
yes there are misleading things in there, but a mule is 6 workers (close patch) worth of income so its like those two scvs turn out three times better afterwards for an extra 50 minerals, given that this is around 15 supply then a toss hits 19 supply-ish when the mules come out
logic: it takes two workers to build orbital so thats elsewise 17 supply, and thats 11 more than the six you start with so at ~22% thats 19.4, where the terran has the equivalent of 21 if the mule was close
given that scvs take time to build structures and probes dont though, this only begins to matter later when the toss saturates, and if anything, untill then a terran is behind because of that
zerg get a worker every 15s instead of 17s, but they have to cut one for every building and other unit so im wondering how that compares
and that goes down to 6s after the queen pops
i believe those build order tools could be useful
the more mineral based thing only applies to low tech units, bio is a lot more solid on terran
around the same time some decent infantry pushes arrive, the mc mass sentry push and kyrix baneling aggression could arrive, both much more gas intense
if it was a sky terran or terran mech, its gas intense just like the others
|
Great read, thanks.
Does your analysis imply that it's better to temporarily stop making workers if you're not going for a fast third then? Or maybe should the builds be optimized to incorporate a faster third in them?
For example in my standard PvZ build I usually have around 60 workers on 2 bases on 12 minutes which means I'm way oversaturated. I don't usually get third before 13 minutes mark, where I'd have about 72 workers (the optimal number ?). Is it better to cut workers production around 54 until you can safely get third or should the build be changed to get third at 54 workers? Also, assuming no successful harassment occurred, should you completely stop workers around 72?
|
Rofl @ 300 food cap. It would be waaaay imbalanced, but towards zerg!
I mean, with a production as big as zerg they would be at 300 supply waaaaay before p/t could even think of being there. Which would be imba.
|
|
|
|
|
|
EPT
