EPTA Treatise on the Economy of SCII - Page 15
| Forum Index > Legacy of the Void |
I have received requests on how to try the model out: Search "Double Harvesting (TeamLiquid)" by ZeromuS as an Extension Mod in HotS Custom Games to try it out. Email your replays of your games on DH to: LegacyEconomyTest@gmail.com might have partnership with a replay website soon as well  In Game Group: Double Harvest | ||
Teoita
Italy12246 Posts
| ||
|
Hider
Denmark9376 Posts
I do not think that 15% more mining over 35minutes of mapcontrol vs passive style is that significant to be honest. I don't think I ever implied it was noticeable higher. But I implied that its not true that immobile and mobile always mine the same and I think people are wrong if they assume that in an average BW come the income assymetry is +50%. Rather, zerg would often have fewer workers on more bases. Honestly wouldn't surprise me if zerg vs mech or zerg vs toss or toss vs terran mech income assymetry (in BW) ranged between 10-30%. Point being here is that income assymetry is irrelevant in itself if its still signficiantly below cost efficiency. Just go and watch your average Avilo game. He stays on 3 bases for 30 minutes. Has half the income but 3 times the cost efficiency ---> incredibly lame. Now lets think about what happens with LOTV economy --> More bases for mech and more reasons to expect immobile race not averaging 3 bases --> More armytrading lategame. So I don't see why the late game armytrading is the big argument for the BW economy. It's not like Firecake lost 30 drones to hellion/banshee or anything...seriously that series really really really isn't indicative of balance or game design, at all, in any way This misses the point. Its not about a specific game. Rather this was an example from today of a typical trend I observed over a longer time period: The slower the mech player take bases --> The higher the income assymetry is + the lamer gameplay is. Do you disagree with my observations here? When TvZ mech was most enjoyable was around late 2013 when korean terrans would go more viking heavy and take bases faster. As the mech meta stabilized around heavy Raven play, it became signifciantly more turtlish as it increased the cost efficiency of the mech player while also reducing mobility --> Resulting in slower base taking + much harder for the zerg to armytrade. Hence one should consider if proper redesigns of units + more bases isn't enough to get the desired incentives in the late game. Which is why I cannot understand why everyone ignores the mid-game as no other economy can replicate what BW does here. The ability to stay on fewer bases against a mobile opponent being on many more faces creates unique advantages. | ||
|
Big J
Austria16289 Posts
On April 13 2015 08:15 Hider wrote: I don't think I ever implied it was noticeable higher. But I implied that its not true that immobile and mobile always mine the same and I think people are wrong if they assume that in an average BW come the income assymetry is close to 50%. Rather, zerg would often have fewer workers on more bases. Honestly wouldn't surprise me if zerg vs mech or zerg vs toss or toss vs terran mech income assymetry (in BW) ranged between 10-30%. But did it feature the same passiveness? There is a huge strategical dynamic behind you being able to theoretically getting a 50% income advantage if your opponent is just passive. The immobile player is forced to attack before getting into his compositional comfort zone just through the possibility. The dynamics balance on a smaller income advantage, but a more aggressive immobile player. Meanwhile in SC2, the income advantage of at most 20% is a given. There is no strategical dynamic behind it that would give the players reasons to attack, or not to attack. Point being here is that income assymetry is irrelevant in itself if its still signficiantly below cost efficiency. Just go and watch your average Avilo game. He stays on 3 bases for 30 minutes. Has half the income but 3 times the cost efficiency ---> incredibly lame. Now lets think about what happens with LOTV economy --> More bases for mech and more reasons to expect immobile race not averaging 3 bases --> More armytrading lategame. So I don't see why the late game armytrading is the big argument for the BW economy. I don't see how a real defensive, immobile style would even be playable in LotV economy. You just run out of money and that's it. The blizzard solution is that former immobile styles become mobile with cyclones and siege pick ups and BC teleports. They just kill the style with their solution. Obviously the problem is not just economy. There needs to be some form of balance. Before the Swarm Host patch that balance was relatively simple: Zerg has a somewhat equally costefficient army against both Protoss and Mech. There is no other good solution for that problem if we don't give players the possibility to overcome inefficiency with income advantages. | ||
|
Hider
Denmark9376 Posts
But did it feature the same passiveness? There is a huge strategical dynamic behind you being able to theoretically getting a 50% income advantage if your opponent is just passive. No, and that's my point here. The main reason it didn't was only partially due to income assymetry and more related to unit design, hence why I previously wondered how turtly TvZ mech game would have been if teran replaced Goliaths with Vikings and Ravens while removing Dark Swarm from zerg. BW simply had a lot more anti-turtling counters while Sc2 interactions are more about preventing anything from occuring in the first place. I think we can look at the incentive for army trading as a simple formula: X/Y. Armytrading occurs if the ratio is above 1. X = Income assymetry Y = Cost efficiency assymetry. E.g. let's say that in current Sc2 X is 1.15 and Y is 1.5 --> the ratio is calculated as 1.15/1.5 --> below 1 --> no armytrading is rewarded. If those numbers are correct, isn't it a bit missing point if all the analysis focus on ways to increase the X value to 1.25 while ignoring the Y value? Remember that the Y value is reduced when your forced to spread your self thinner as it makes it easier for the enemy to find the weak links. And on top of that, remember that X most likely will be increased going from HOTS to LOTV economy as the immobile race cannt always be on 3 active bases. I don't see how a real defensive, immobile style would even be playable in LotV economy Here is a suggestion: - 2 supply Siege Tanks - Upgrade that requires Fusion Core = +15 damage to Siege Tanks (PDD needs to go completely 20 sec still dumb) Effect: Tanks get cost efficiently enough (raises Y value from what probably in a post Raven world is close to 1.15 to 1.3 in the late game. if X is around 1.3 as well you should get some really solid gameplay.) The point being here isn't about the exact numbers, but rather how you should balance/design units around the income assymetry in the economy. An economy can not be analyzed as good or bad in a vacuum. Rather you have to look at the opportunites to fix the flaws through changes to units. | ||
|
TheWinks
United States572 Posts
On April 13 2015 08:15 Hider wrote: This misses the point. Its not about a specific game. Rather this was an example from today of a typical trend I observed over a longer time period: It's been 3 days. | ||
|
Big J
Austria16289 Posts
On April 13 2015 08:41 Hider wrote: No, and that's my point here. The main reason it didn't was only partially due to income assymetry and more related to unit design, hence why I previously wondered how turtly TvZ mech game would have been if teran replaced Goliaths with Vikings and Ravens while removing Dark Swarm from zerg. BW simply had a lot more anti-turtling counters while Sc2 interactions are more about preventing anything from occuring in the first place. I'm not really sure you are right with that. I've not only once read a comment how Terran would just win if they could split the map or get a critical mass of Vessels. But it's quite hard for me to argue that BW stuff, so I would like to just leave it to more knowledgeable people how the dynamics were exactly. Regardless of that, I like the strategical implications of a scaling economy. I can't see the game being played with the same standardization if this would go through. The advantages of having a 5th or 6th base if you have the mapcontrol would be way more subtle and in itself lead to more diverse strategies just by players taking that extra risk or not believing it doesn't pay off. Sorry, for the late edit/reply. Didn't realize you had also edited On April 13 2015 08:41 Hider wrote: I think we can look at the incentive for army trading as a simple formula: X/Y. Armytrading occurs if the ratio is above 1. X = Income assymetry Y = Cost efficiency assymetry. E.g. let's say that in current Sc2 X is 1.15 and Y is 1.5 --> the ratio is calculated as 1.15/1.5 --> below 1 --> no armytrading is rewarded. If those numbers are correct, isn't it a bit missing point if all the analysis focus on ways to increase the X value to 1.25 while ignoring the Y value? Remember that the Y value is reduced when your forced to spread your self thinner as it makes it easier for the enemy to find the weak links. And on top of that, remember that X most likely will be increased going from HOTS to LOTV economy as the immobile race cannt always be on 3 active bases. The whole point with a defensive style is that you are hard to break if you don't want to be broken. If such styles simply result in the opponent only having more stuff and you needing less stuff, I don't see the value of having that style to begin with. - that's not quite true, but it's just "I'm using other units" then. The point is not to just have a given amount of income advantage vs a given amount of costefficiency. The point is that players should be able to manipulate how X and Y look. It's about having a strategic choice, where I either go for 50% more income and turn myself into the defensive player, or try to bluff you and make you run into an open field battle that you shouldn't be taking at that time. But the current and the LotV economy inevitably give us X implicitly, so Y also has to have a certain value to put the ratio to 1 (=balance). - 2 supply Siege Tanks - Upgrade that requires Fusion Core = +15 damage to Siege Tanks (PDD needs to go completely 20 sec still dumb) Effect: Tanks get cost efficiently enough (raises Y value from what probably in a post Raven world is close to 1.15 to 1.3 in the late game. if X is around 1.3 as well you should get some really solid gameplay.) The point being here isn't about the exact numbers, but rather how you should balance/design units around the income assymetry in the economy. An economy can not be analyzed as good or bad in a vacuum. Rather you have to look at the opportunites to fix the flaws through changes to units. The main problem of Mech at the moment isn't even costefficiency. Mech builds a more expensive army to begin with, than zerg does. The problem is that zerg theoretically can't even match the cost of a fully developed mech army, which is why zerg needs to trade for free against it to just prevent that scenario to begin with. Costefficiency becomes moot when supply limits and unit costs simply dictate that one race has better tools than the other in a "perfect" scenario. Putting the tank to two supply AND buffing it so that you need even less of them to do whatever the tanks role is in the army just makes this even worse and increases the incentitive for a Terran not to attack, because his 200supply army will be even stronger. | ||
|
Hider
Denmark9376 Posts
I've not only once read a comment how Terran would just win if they could split the map or get a critical mass of Vessels. I think that's more of a theoretical scenario than something that is realistic at pro level (too gas intensive I assume). Most TvZ mech games - in the late game - are very focussed on overlord drops with Ultralisks, Lings and Defilers with the terran mech player being spread all over the map. Regardless of that, I like the strategical implications of a scaling economy. I can't see the game being played with the same standardization if this would go through. The advantages of having a 5th or 6th base if you have the mapcontrol would be way more subtle and in itself lead to more diverse strategies just by players taking that extra risk or not believing it doesn't pay off. I don't disagree here, though my biggest issue is mainly the lack of comeback-potential. I really dislike how LOTV games are right now. I rather prefer less action in the late game with a higher likelyhood of a great lategame. | ||
ZeromuS
Canada13389 Posts
On April 13 2015 09:00 Hider wrote: I think that's more of a theoretical scenario than something that is realistic at pro level (too gas intensive I assume). Most TvZ mech games - in the late game - are very focussed on overlord drops with Ultralisks, Lings and Defilers with the terran mech player being spread all over the map. I don't disagree here, though my biggest issue is mainly the lack of comeback-potential. I really dislike how LOTV games are right now. I rather prefer less action in the late game with a higher likelyhood of a great lategame. But when it comes to unit design and balance that is completely outside the scope of this article. The difference the DH model makes is also much more strongly felt in the mid game than late game since the natural and third bases make the biggest income difference to the flow of the mid game (compared to hots) we cant really begin to wonder what the game will look like far beyond that point. What I can say is that the player with more expansions has a lot more income on even worker counts, and that even in a turtle situation the expanding player due simply to a higher income per minute than in HotS, that isn't capped in nearly the same way as HotS, will probably not starve himself out on attacks vs a turtle player. This is because from my perspective, and what I've seen in HotS the aggressor rarely trades efficiently with the defender and if the income is capped, net losses will always favour the player who loses less assuming both income rates are the same. Its a very simple fact that I think is more often than not borne out in games we see in HotS. I posit that by unlocking the cap (here is where theory comes in) the net losses for the defender and attacker will even out so that one player doesnt eat their own bank through futile attacks. Of course there is such a thing as being TOO inefficient (and firecakes game today was an example) and this specifically needs to be solved over time through unit balance and exploring new strategies (or old ones). Broodlords are still good and ultralisks are getting major buffs in LotV and so are roaches (with burrow move) and ravagers and any the new viper spell vs air, and lurkers etc etc. | ||
|
Paljas
Germany6926 Posts
not quite sure if i am making sense or if i misunderstood your proposal. | ||
|
ZeromuS
Canada13389 Posts
On April 13 2015 09:09 Paljas wrote: the double harvest method seems somewhat unnatural, in the sense that a worker going back after mining but not returning the minerals (thats how the model works, right? only giving the minerals back every second trip. i cant play right now to test it) could be problematic in both early (scouting/worker micro) and out-mined late game scenarios. would it be possible to alter the model in a way that workers still return 5 minerals after every trip but still "stay" on the mineral patch for two trips (this could maybe lead to some sort of early game worker micro to boost effeciency)? not quite sure if i am making sense or if i misunderstood your proposal. If you look at the OP, there is a short gfycat of how it looks and works. They just harvest twice, they dont go back to the nexus in between | ||
|
dyDrawer
Canada438 Posts
And we can then tweak the number of near patches vs. far patches in map design. It can either be standardized on all maps, or it can even be left upon the map designer to decide as a potential factor in map balancing. Like, if we have 4 near patches and 4 far patches, it would reduce the optimal worker number at a mineral line from the current 16 to 12. If this is not good enough, then we can have 6 near patches and 2 far patches, which brings the number down to 10. Forcing a worker pull wouldn't be so devastating, and it doesn't really cause a severe income rise as the mining rate stays the same. I know next to nothing about the Map Editor so I don't know if we can test this (the worker AI's probably something hard coded tho). But if anybody knows how to achieve this I think it might be worth the try. | ||
|
ApBuLLet
United States604 Posts
The way they are currently going in LOTV seems too drastic and really doesn't address the problem that is at the root of the issue. The idea that a player can be punished by the game not for NOT doing something seems like bad design to me. Rewarding players for expanding rather than punishing players for not expanding leaves it up to the expanding player to punish the player that does not expand, instead of the game implicitly doing it for them by taking their mineral patches away. It may seem subtle but to me that is a very distinct difference. I think this makes the biggest difference for low level players who might not expand very quickly. Not expanding won't necessarily doom them, even if their opponent does choose to expand, as it is still up to the expanding player to properly use their economic lead to punish their opponent and win the game. | ||
|
Umpteen
United Kingdom1570 Posts
Although you can tweak the amounts so that the gold minerals run out at a similar time to the blues on a saturated base, the very act of spreading your workers across unsaturated bases causes them to be mined unevenly. With DH, if you split 16 workers across two bases you get increased efficiency until the base is exhausted. But with the mixed gold/blue bases, you're going to mine out of the gold patches first because that's where you send your workers. For Zergs taking a quick third, this is made worse because they'll start in on their third gold minerals before saturating the first and second bases. So they'll get a heavily front-loaded burst of benefit, and then the income will dry up to a trickle. I personally can't see that ending well. Zergs will have a gigantic mineral advantage to begin with, after which they'll be starving to death unless they can take a fourth and fifth in short order. That's way too bumpy a ride. Eh, well, at least it wasn't my idea XD | ||
|
knyttym
United States5797 Posts
On April 13 2015 08:41 Hider wrote: + Show Spoiler + But did it feature the same passiveness? There is a huge strategical dynamic behind you being able to theoretically getting a 50% income advantage if your opponent is just passive. No, and that's my point here. The main reason it didn't was only partially due to income assymetry and more related to unit design, hence why I previously wondered how turtly TvZ mech game would have been if teran replaced Goliaths with Vikings and Ravens while removing Dark Swarm from zerg. BW simply had a lot more anti-turtling counters while Sc2 interactions are more about preventing anything from occuring in the first place. I think we can look at the incentive for army trading as a simple formula: X/Y. Armytrading occurs if the ratio is above 1. X = Income assymetry Y = Cost efficiency assymetry. E.g. let's say that in current Sc2 X is 1.15 and Y is 1.5 --> the ratio is calculated as 1.15/1.5 --> below 1 --> no armytrading is rewarded. If those numbers are correct, isn't it a bit missing point if all the analysis focus on ways to increase the X value to 1.25 while ignoring the Y value? Remember that the Y value is reduced when your forced to spread your self thinner as it makes it easier for the enemy to find the weak links. And on top of that, remember that X most likely will be increased going from HOTS to LOTV economy as the immobile race cannt always be on 3 active bases. I don't see how a real defensive, immobile style would even be playable in LotV economy Here is a suggestion: - 2 supply Siege Tanks - Upgrade that requires Fusion Core = +15 damage to Siege Tanks (PDD needs to go completely 20 sec still dumb) Effect: Tanks get cost efficiently enough (raises Y value from what probably in a post Raven world is close to 1.15 to 1.3 in the late game. if X is around 1.3 as well you should get some really solid gameplay.) The point being here isn't about the exact numbers, but rather how you should balance/design units around the income assymetry in the economy. An economy can not be analyzed as good or bad in a vacuum. Rather you have to look at the opportunites to fix the flaws through changes to units. I think I understand what you are getting at. So what you would want is pairs of X and Y values that come at different points in the game. So let's use your example of the siege tank and think about a TvZ. At the beginning stages of the game when both players are on 3 base you have a pair (X1,Y1). Once the game progresses, terran stays on 3 base due to immobility and zerg moves on to 4-5. To account for this we have a new pair (X2,Y2) that results from the siege tank upgrade (changes Y) and the zergs increased mining (changes X). I think I have everything correct up to this point. So my issue is then the immediate power spike once you transition from Y1 to Y2. The transition from X1 to X2 is more gradual as patches are depleting and zerg is moving drones to new bases. What happens then is terran has its highest chance to strike when he obtains Y2 while zerg the game is transitioning from X1 to X2. All upgrades have this power spike but the proposal of an upgrade that allows (X1/Y1) = (X2/Y2) has to to be incredibly powerful. Why does this differ from a suggestion to first focus on X and then later Y? Let's look at (X3, Y3) -> (X4,Y4) Y3 -> Y4 is a gradual process predicated on the accumulation of tanks. As X3 -> X4, Y3-> Y4. Similarly we can have (X5,Y5) pairs and so on and so forth. The introduction of an upgrade to solve this process makes it so we can only have Y1 and Y2. So now back to your suggestion, maybe if the tank has some kind of extremely specific scaling upgrade system we could obtain a similar gradient style approach. So +1 vehicle weapons gives siege tanks an extra 5 damage but +2 gives it an extra 15 or something like that. This would be extremely difficult still because the transition in LOTV from X1 -> X2 is relatively quick (it's still enough for the tank upgrade power spike, Y2, to be a problem) while upgrading something like vehicle weapons is spread over close to 10 minutes. Secondly, all immobile forces would have to adhere to this awkward scaling upgrade system. | ||
|
AySz88
United States83 Posts
On April 13 2015 09:45 dyDrawer wrote: I'm wondering if it is possible to tweak the mining stats of workers, so that only the farther mineral patches can be paired and the nearer ones cannot? And we can then tweak the number of near patches vs. far patches in map design. It can either be standardized on all maps, or it can be left upon the map designer to decide as a potential factor in map balancing. If I may plagiarize myself, I made a similar suggestion on reddit - just move some patches closer to the town hall, and some farther away: that accomplishes the basic desire to have the first few workers more efficient and the last few less efficient. For example, maybe move 2 mineral patches closer (right next to the base?), 4 patches at roughly current distance, and 2 at 3-worker distance. (...) The worker AI might need a little adjusting (if you don't want everyone to manually micro workers to patches!). Maybe the CC/Nexus/Hatch should "own" the AI for all the mineral-harvesting workers at a base, and direct the workers to maximize efficiency. I noticed the suggestions regarding gold minerals are along the same lines - the core idea seems to be to make the first few mineral patches more valuable when you expand, so that there's always a reward for expanding even past 3 bases. I'll be happy if that's the takeaway by Blizzard from all this discussion. | ||
|
BronzeKnee
United States5217 Posts
On April 12 2015 05:58 ZeromuS wrote: I really hope people take the time to read this entire article. In it I break down the HotS economy, the LotV economy, and I provide what is truly more "BW-like" an economy that the TL strat team would love to see at least get a chance in LotV Beta. Its a long beta. Give other economic models a chance. Player influenced expansion based gameplay is we believe, a far better approach than time influenced expansion based gameplay. Thanks for reading this huge thing, we spent a LOT of time on. I read the whole thing, it is a good read and the conclusion here is obvious: some type of worker efficiency by double harvesting or double mining is better than the half patch approach. My only critique is that the article itself was overly simplistic with a lot of obvious statements and had many redundant statements and ideas. I suppose if you wanted to argue this to someone without any Starcraft knowledge, the ideas would need it to be simple and obvious (though the redundancy is unnecessary in all cases), but for the intended audience it seemed overboard. I do want to say that I don't think it is a bad thing that pulling workers off the line is punished more with double mining as compared to current HOTS mining because if you're pulling workers then one of the following is happening: 1) You're Terran and doing all-in and at that point you don't care about your economy. 2) You didn't scout worth a damn, and therefore you're being punished for not scouting. And scouting has really fallen off in HOTS compared to WOL. 3) You're being harassed. I really think the overall skill level of Starcraft goes up when there is more strategic diversity. And more strategic diversity means an increasing need to scout because there are more strategies to prepare for. And scouting and making reads is an important part of the game. Workers aren't combat units, don't use them as such. I also really want to play one base again more often because it's awesome, so that's why I also like double mining. Double harvesting is probably the way to go. But anyway, this article is very well done and logically sound, probably the best I've read on TL. The conclusion is inescapable. I just hope Blizzard listens. Actually I'm praying they do. | ||
|
OtherWorld
France17333 Posts
| ||
|
EsportsJohn
United States4883 Posts
On April 13 2015 13:42 BronzeKnee wrote: I do want to say that I don't think it is a bad thing that pulling workers off the line is punished more with double mining as compared to current HOTS mining because if you're pulling workers then one of the following is happening: 1) You're Terran and doing all-in and at that point you don't care about your economy. 2) You didn't scout worth a damn, and therefore you're being punished for not scouting. And scouting has really fallen off in HOTS compared to WOL. 3) You're being harassed. Actually, this is more of an issue specifically with hyper early proxies like proxy 2-gate in PvP and 6pools in ZvZ where you HAVE to defend with workers or at least dance them some. In these cases, you have a much slimmer chance of defending all-ins which are already difficult to hold because you're likely losing a lot more mining. Some of these are literally not scoutable on 4-player maps in time, which is why they're already powerful. The double harvest model helps a lot with making these all-ins a little less harsh in comparison. EDIT: also, 10-gate into 3-gate and proxy gate/core PvP | ||
|
BronzeKnee
United States5217 Posts
On April 13 2015 14:36 SC2John wrote: Actually, this is more of an issue specifically with hyper early proxies like proxy 2-gate in PvP and 6pools in ZvZ where you HAVE to defend with workers or at least dance them some. In these cases, you have a much slimmer chance of defending all-ins which are already difficult to hold because you're likely losing a lot more mining. Some of these are literally not scoutable on 4-player maps in time, which is why they're already powerful. The double harvest model helps a lot with making these all-ins a little less harsh in comparison. EDIT: also, 10-gate into 3-gate and proxy gate/core PvP 4 player maps for a 2 player game never really made sense to me anyway. Either way, double harvest is certainly going to be the easier (and for our sake and Blizzard's that is better) model to implement. | ||
|
docvoc
United States5491 Posts
| ||
| ||
