|
LotV economy: non-linearity in time
Abstract
With the release of Legacy of the Void closed beta, Blizzard decided to change one of the fundamental mechanics of the Star Craft 2 game: The economy. This has sparked a lot of debate around the topic, and the team liquid strategy team wrote an excellent article analyzing this. They also came up with an alternative model, which they argue is better than the current LotV economy model. Blizzard has since recognized the ideas and given a formal reply. I would like to contribute to the discussion by providing some insight into the arguments of both parties, as well as propose a (third) alternative model of how to change the economy for the better. It requires no change in game mechanics or AI, but will fix a lot of the problems raised by the community.
Disclaimer
I am in no way affiliated with Team Liquid or Blizzard. I am however a long time fan of both. My views are my own and you may take them or disregard them as you please.
Problem Statement
There are in essence two problems which I hear voiced over and over. They are the following
a) HotS: There is a base cap of three saturated mining bases, of which there is little to no economic gain in expanding to more.
b) LotV: “Expand or else” - there is a too hard penalty of not managing to acquire new expansions
On one hand you want to encourage expanding, but on the other hand you want to keep the strategic diversity and make it possible to to turtle on few bases. To quote a problem raised in other threads. If you have one player with 32 workers and 2 mining bases versus a person with 32 workers and 4 mining bases, you want the 4-base player to have some economic advantage, but not a devastating game-ending advantage. The numbers are of course worker-dependent, but it seems to be a consensus that for “reasonable” worker counts the four base player should be at an advantage, but not have as much as twice the income.
The LotV model
The change in Legacy of the Void is that bases have two types of mineral patches. 4 patches containing 900 minerals, and 4 patches with 1500 minerals. What this means is that bases start operating at 50% after a certain amount of time have passed, encouraging expanding, while still maintaining some mining capacity in the original base.
The DH10 model
The proposed change from the TL strategy team is as simple as it is brilliant. Make workers mine twice, carrying a total mineral load of 10 minerals before returning the load. What is neat about this is that it disrupts the otherwise perfect AI of worker-pairing, making workers co-operate less. They mine more effectively on their own (less walking, more mining), but co-operate worse resulting in a non-linear income vs workers curve. What this accomplishes is that additional workers beyond the first one-per-mineral-patch is put to better use on a fresh expansion (mine at 100%), rather than in your main base (mine at less than 100%). Expanding is encouraged and provides meaningful advantages, while still maintaining the few base play-style viable.
The income function
To start the reasoning here, I would like to introduce the income function. Traditionally, this is displayed as “collection rate (income per minute) as a function of time and is readily available at each end game summary and it looks like this:
![[image loading]](http://i.imgur.com/miDcE36.png)
Lately however we have looked at it from another point of view. Instead of having elapsed time on the x-axis, we have looked at worker count, and it then looks like this
![[image loading]](http://i.imgur.com/NUcoXxR.png)
To understand my reasoning here, I would like to the discuss the income function Inc(x,t) as a function of both parameters: the number of workers x, and the elapsed time t. When discussing base saturation, this seem like an odd thing to do at first, since usually the whole base mine out at the same time. Mineral patches disappear, all within a very short time window. You would then drop from full income (given by the number of workers you have) to zero income, for that particular base. This however assumes a “normal” game. If you, for the sake of argument, did not have full saturation you would find that any mineral patch with 2 harvesters disappear quicker than those with a single harvester.
![[image loading]](http://i.imgur.com/JTuJ6CE.jpg)
Having 12 harvester for a base means that half the mineral patches are depleted before the rest and you are left with 4 mining patches.
![[image loading]](http://i.imgur.com/gHD4COY.png)
An almost correct visualization of the income function. It is however slightly more descriptive than the correct version below
![[image loading]](http://i.imgur.com/ncCTSNL.png)
This has all been pretty irrelevant in the past due to several reasons. First, the small window of time between when the three-worker mining patches disappear (00:14) and two-worker mining patches disappear (00:17) is too short (and prone to data noise) to pay attention to. Secondly because it is not common to leave less than 16 workers at a mining base for extended periods of time (17+ minutes). However, the reason we bring it up now is because of recent developments from blizzard. With the current LotV economy model, this picture changes dramatically.
![[image loading]](http://i.imgur.com/hU7wS9L.png)
The thing screaming for attention in this figure is the transition happening at around 10 minutes, where (at least one of) the 4 small mineral patches (900 minerals) disappear. It happens for all bases with more than 12 workers, but is most devastating for bases with 16 or more workers, where the income is suddenly sliced in half.
![[image loading]](http://i.imgur.com/JwFUnmX.png)
LotV economy with different worker count. Plotting these in a more familiar format of income rate vs time, the 10 minute drop where mineral patches disappear is clearly visible.
There is a lot of interesting things you can say about the income function when looked at from this point, but for now it is sufficient to look at the discontinuity happening at the 10 minute mark. It is my belief that the reason for the “expand or else” rhetoric is because many players feel that they are on a clock. And that clock is set to exactly 10 minutes (7 minutes at the start of the beta) and if you have yet to expand by then, then you are severely punished. While you had an equivalent timer in HotS, this was set to 14 minutes and players were more accustomed to this.
The degrading patch model (DPM)
What I propose here is a different model than the DH10 model, but builds on the ideas introduced in LotV. Instead of introducing an alternative income/worker curve, I propose to refine the income/time curve. It is believed that the effects of slicing your income in half is to severe, so we smooth this out by slowly decreasing it from full mining capacity to zero in multiple steps. By realizing that income drops happen due to mineral patches getting mined out, we simply let the mineral patches have an even more discrepency in size. For instance, consider the following mineral distributions:
![[image loading]](http://i.imgur.com/lovLrBj.png)
Income function with mineral patches containing 200,400,600,800,1000,1200,1400 and 1600 minerals
![[image loading]](http://i.imgur.com/2TvRZNu.png)
Income function with mineral patches containing 900,1000,1100,1200,1300,1400,1500 and 1600 minerals
![[image loading]](http://i.imgur.com/OAiHG0S.png)
Income function with mineral patches containing 600,800,1000,1200,1400,1600,1800 and 2000 minerals
This will accomplish several things. First of all, the current mining model and worker-pairing ensures that the initial income with 8 workers or 16 workers or 24 workers is identical to the current HotS economy. However, you will realize that you quickly mine out the smallest patches, thus reducing the available number of mineral patches available to mine from. After a given number of game seconds the first mineral patch will be mined out and the effective worker count at that base reduces from 14, down from 16, and the max worker count is 21, down from 24. This continues and the longer that time goes by, the less and less effective your worker force will be. The difference here is that this is a gradual change, and not a sudden one. While the numbers can be tweaked, we note that 600-800-1000-1200-1400-1600-1800-2000 minerals seem to be a decent mineral distribution for multiple reasons. First it has a total mineral deposit of more than the LotV model, and less than the HotS model, and secondly the smallest mineral is 600 which provides a good starting time gap before minerals start depleting. The numbers can however be tweaked and the primary message of this post is to let all patches have different sizes.
Multiple mining bases
Let us investigate what happens in a late-game scenario. Say that one has either 3 or 5 bases available and between 40 and 70 workers. The income function for the current LotV economy and DPM economy would then be
![[image loading]](http://i.imgur.com/XdKZNy8.png)
3 Bases with LotV ecomony
![[image loading]](http://i.imgur.com/vTDtcij.png)
3 Bases with DPM economy
To simplify matters, let us fix the worker count to 60, and see how resource collection rate over multiple bases develop
![[image loading]](http://i.imgur.com/fk2SIjk.png)
Comparing 2 base play with 4 base play with the same amount of workers across different economy models
![[image loading]](http://i.imgur.com/xqHmXMN.png)
Comparing 3 base play with 5 base play with the same amount of workers across different economy models
As is seen here, having 5 bases in the HotS economy only provides a fractional advantage when compared with 3 bases. You have the same income til the 15 minute mark after which the 3-base player runs out of minerals. With the DPM model the starting advantage is small, but as time goes by the advantage keeps growing penalizing the turteling player in a gradual way. Make no mistake. The current LotV model makes things better, but the DPM model takes the ideas one step further.
New micro opportunities
Both the current LotV model and the DPM opens up for optimal control of resource management. Just like 4 marines with 10 health is better than one marine with 40 health, the same is true for minerals. You want to distribute your workers to harvest minerals (and pair workers on minerals) with more resources. Having multiple patches with small amount of resources means that your resource collection rate remains high. This allows for micro during the first seconds of a game to get workers on the right mineral patches (Brood War nostalgia anyone?) and it rewards terran players for making their mules hit the right resources.
Worker scouting can break pairing on the good mineral patches and finally run away with 5 minerals from the already smallest patch of your opponent
TL; DR
Let bases have patches with the following mineral distribution: 600-800-1000-1200-1400-1600-1800-2000. This changes the sudden drop in income (current LotV model) with a smooth drop in income. It rewards players for expanding, while still keeping 2-3 base play viable. The primary difference from the DH10 model is that this reduces effectivity from time as opposed to reduce effectivite from worker count.
Even if you discard the DPM model, there is still a lot to take from this post with respect to the current LotV economy and the effects of suddenly dissapearing mineral patches.
|
Very good post, thanks for your work. Clearly the “domino collapse” is superior for the stability of the game, and map makers could use this... if they had licence from Blizzard in that regard. It would be an ingenious solution indeed to mitigate the effects of the bad LotV economy (whose crash is sadly intended). GG!
|
I very much prefer this idea to the LotV economy!
|
United States4883 Posts
Interesting idea. This is definitely a new and unique solution to what we've been talking about, and it follows a lot of Blizzard guidelines while still attempting to accomplish the same goals as the DH model. I also really like that a) this is very information-based and has lots of data to back up what you're trying to say and b) it's laid out in a way that's very easy to understand.
I do have one concern though, and I would appreciate any correction on any point where I'm wrong. Basically, because this model is function of time rather than worker count, I feel that it encourages one expansion at a time rather than any kind of mass expansion. That is, you don't need to outexpand your opponent by much; for example, being on six bases does not actually grant any advantage over three bases in the short run, assuming each player is still mining optimally (or close to it). That said, I believe it may force a quicker fourth, but once each player has 32 nodes - X (mined out nodes), they will still be capped at a similar value to HotS, and while you can't turtle on 3 bases forever, you can still turtle relatively well and hit brutal timings with a hyper economy that can't be touched. This is a function of the time to mine out or significantly lose income to be too far into the game.
If you were to counteract that by decreasing the amount of time to mine out -- to mine out nodes (in general) sooner -- you run the risk of making each base mine out too soon, in which case it might be even more brutal than LotV, and in which model newer expansions really don't pay off very well in high economy games and pay off too well in low economy games.
Let me know what you think of this concern and if I'm wrong in any areas. I don't disagree with the model, I just think that choosing a time-sensitive model is difficult for finding the exact balance, especially because then it is up to the game designers to fix how the game should be played rather than giving players free reign of what to do.
|
This was my idea also. What a polished, legit write-up! Well done and thanks!
The only thing I would suggest differently is that I would have the first few patches run out sooner. That way there's an earlier incentive to take your 3rd-5th bases.
Perhaps
300-600-900-1200-1500-1800-2100-2400
(fyi: total minerals is 400 more)
rather than
600-800-1000-1200-1400-1600-1800-2000
What do you think of that configuration?
edit: of course, this is compatible with DH as well.
|
I like your clear explanation of this idea that has been tossed around a lot recently. My preference is still pretty strongly for a more fundamental change, such as DH. This is a great backup plan though. If Blizzard remains arrogant about necessary economic changes, mapmakers can use your DPM system to bring about a bit of a jagged curve in terms of mining out, in order to mitigate the issues caused by the half-patch model. + Show Spoiler +I'm predicting the half-patch model to become the two-thirds-patch model within the next few months, for a few reasons.1) 900/1500 minerals is still a bit too punishing at mining out around 10 minutes, 2) 1000/1500 is a more pleasing number spread, and Blizzard loves them some easy readability, 3) more of a 1.5, but it's the next logical step for them to retreat to, after they see that their 60% patch model isn't working as they imagined either.
|
Aotearoa39261 Posts
You know shit is serious when matlab gets pulled out.
|
I skipped trought some parts so I may be wrong in this but, isn't the lack of a constant income also bad?
I mean, if you are on 3 fully effective bases, and you know you will have Y amount of income for some time, you will build X amount of production facilities that can spend your income nicely, but with this, since your income starts to slow down too soon (even thought it is slowly) you will never be able to have compatible number of production for your income.
That is why I think having a constant income -> dive -> constant income could still be better.
Blizzard wants people to expand faster, and I think they got it.
|
On May 15 2015 10:33 rpgalon wrote:
I skipped trought some parts so I may be wrong in this but, isn't the lack of a constant income also bad?
I mean, if you are on 3 fully effective bases, and you know you will have Y amount of income for some time, you will build X amount of production facilities that can spend your income nicely, but with this, since your income starts to slow down too soon (even thought it is slowly) you will never be able to have compatible number of production for your income.
That is why I think having a constant income -> dive -> constant income could still be better.
Blizzard wants people to expand faster, and I think they got it.
You'll still be able to figure out how many production buildings you'll have because player will time out how much will be coming in by the time they attack.
Therefore if you're doing a two base timing, and you have 5 mineral nodes left at your main at the time of your attack, you can plan for that.
On May 15 2015 08:37 SC2John wrote:
...because then it is up to the game designers to fix how the game should be played rather than giving players free reign of what to do.
This has been the struggle with SC2 from day 1.
I don't understand why, but Blizzard erroneously believes that if people expand faster more action will result, and that will fix the three base turtle. That is only true if expanding isn't safe.
And isn't expanding rapidly part of the problem? People spend all their money in economy going up to three bases, that is why there is no action early and the game is so boring. If we make them go up to more bases rapidly, then even more time will be spent developing economy, unless expanding isn't safe. And if expanding isn't safe, won't we just be back to square one here, WOL Beta?
What really is the difference between Protoss not being able to take a 3rd due to Stephano's Roach max (which required big map changes) and Protoss not being able to take a Y base due to X LOTV strategy? If Y base is necessary, like a third was in WOL/HOTS then we'll end up with the same problems.
And if Protoss can take Y base safely, then Zerg isn't going to bother attacking, and we'll have no action.
People can expand quite rapidly in HOTS and go up to three bases quite often without any major losses because it is easy to expand safely. The reason for that is because Blizzard and map makers gave people the tools they needed to expand to two/three bases without much fear.
Why not look at dialing back some of those tools so that expanding can be done, but it takes skill and resource commitment to defend? I don't think Whitera was talking about pressing F and clicking on the Nexus when he talked about defensing his expands and skill in general, it was a bit harder than that back in the day and that made it fun. No one should have any inherent right to an expansion, expansions should all be earned. And not just the 4th expansion, every expansion.
More than just the economy needs to change to avoid these issues, and a higher level of thinking than just "if people expand more the game will be better" is necessary to fix SC2. We already tried that.
Sorry for the rant.
|
Great post, this seems like a better alternative than the half/half option from blizzard.
from what I understood, at full worker saturation, first mineral node mines out at 2-3 minutes, so more bases is always a bit better.
However I fear that this system only encourages constant expanding and not holding numerous bases. For example, someone with 3 bases initially and expanding at the same time as his opponent on 6 bases initially would have the same income, even though he is 3 bases behind the whole game.
|
I've been saying this for a while now too. I am really glad that someone made an elaborate post such as this about it. Great job! This is a great solution precisely because it requires only minimal effort to have a desired effect, and doesn't require quite as many potential balance changes to make it work properly.
There is one additional thing I think might add even more to the game flow: The addition of Gold patch + high yield geyser bases that are harder to get (say everything beyond the 3rd or 4th) that apply the same idea of gradual decrease over time but also have the added benefit of requiring even less workers to saturate. Then there is a stronger incentive to go for more expansions faster because the income boosts are bigger for less workers beyond the 3rd or 4th base. If someone turtles you get rewarded more for taking extra bases than you normally would with non-gold/high yield geyser bases. It at least makes expanding vs turtling a better natural counter earlier in the game than it was before.
Very nice write up all in all. Keep up the good work!
Original post I made on this:
+ Show Spoiler +On April 21 2015 15:33 Masayume wrote:
While I really do like your idea, it feels like it moves away from the LotV design that Blizzard is toying with right now. I think it might be better to try something that fits into their idea better, while still having a more rewarding situation for expanding vs not expanding, while not killing off the non-expanding player right off the bat.
I propose that instead of 4x1500 and 4x750 patches, we take different values. Example:
4x2250 nodes
2x 1500 nodes
2x 750 nodes
This change would make it so that it becomes less efficient to not expand vs maynarding workers to fresh bases at one point, but you don't run out of minerals and are not neccessarily forced to expand because there is simply 4 nodes left as early.
It becomes progressively worse but you still have enough minerals to keep mining for a while. So while you can for example turtle or go for a 2 base timing, you will build up your army more slowly or recover more slowly from a failed two base, but still have a fighting chance while giving more benefits to the player who expands more and earlier. I feel that this is more easily accomplished than changing the way workers behave. It also prevents a complete overhaul of the game design due to massive changes in mining function for workers and economic values for units being skewed.
Then from the 3rd base onward mapmakers can also use gold nodes and high yield geysers so that everything beyond the 3rd base requires less workers to saturate a base providing similar income to what it is for a normal HotS base.
A gold base could have something like:
2x 2250
2x 1500
2x 750
1x High yield geyser
All in all, I think it is a lot better to work on improving the economy flow with existing properties that do not completely flip the game and require massive redesigns to values and such. This is easier to implement for Blizzard and a lot easier for casuals and viewers to understand. It also makes it easier for mapmakers to play around with it and for players to test and understand, coming from HotS.
I scrambled this down quickly before going off to university so I hope it makes enough sense. I will write more and respond later if people have feedback on this variation. Keep up the good fight and thanks for this great article that opened the discussion on the economy of StarCraft!
|
Great post, but wouldn't degrading patches be too confusing for the Casual?
|
You have the right idea, and I love the level of detail you put into your research.
However, I still think reducing the amount of patches at each base to 6, while keeping the same amount in each patch is a more obvious and reasonable change that will cause bases to saturate faster.. therefore providing a need to expand beyond 3 bases. Keep in mind like any idea it would need to be tested to see how it affects builds (like we've seen the current model extensively tested with much to be desired) but it's a very workable idea that changes the game at all levels for the better.
It's simple, just remove two patches at each base to make it 6 mineral patches with 1500 minerals per match, and 2 gas. Mineral income would decrease at many stages of the game from where it's currently at, but I believe this will work out for the better. MULEs will give Terran the ability to stay on 1-base a little longer than the other races, but the faster pumping of drones/probes should even this out.
To add to this, the 12 worker start has proven to discourage all cheese and encourage more early timings, while keeping very-early game stale, boring and predictable.
If we started with 9 workers, it would be right in the middle of HotS and the current LotV models, which should feel just right.. especially if we were give 200 minerals instead of 50 to start with as well. This provides a choice right away to make very-early game interesting again.. make workers, expand or start building tech/gas.
|
On May 15 2015 15:03 Geiko wrote:
Great post, this seems like a better alternative than the half/half option from blizzard.
from what I understood, at full worker saturation, first mineral node mines out at 2-3 minutes, so more bases is always a bit better.
However I fear that this system only encourages constant expanding and not holding numerous bases. For example, someone with 3 bases initially and expanding at the same time as his opponent on 6 bases initially would have the same income, even though he is 3 bases behind the whole game.
This is correct, but at least with this system there's stability to it (you lose 1/8 of your income at a time per base as opposed to having half your income drop out from under you), so the transition is more gradual / natural.
|
Italy12246 Posts
I personally prefer the DH model, but i like how you took blizzard's idea and actually improved it massively!
|
I prefer DH9 because I dont like being able to produce 40+ supply every minute from midgame upwards. I always feel I do play for wave to wave and just see where was better at that wave. Every fight except 200/200 is a pleasure for me (watch and play) but sadly this is so rare because of heavy (re-)produce.
|
I like this idea a lot. Only problem I see with this is a visual problem. The pathes must very clearly indicate their content.
One thing I have been thinking about is to mix one or two gold patches with normal patches for expansions. That way the players would get a small boost of income immediately for exapansion. This could be combined nicely with uneven patch values so that the gold patches would mine out faster. That would give you an insentive to have an expansion lead because your drones/probes/scvs would be more effecient with new bases..
|
If this doesn't convince Blizzard to change their beta, I fucking give up. Because graded patches is so obviously a superior arrangement of FRB, and I'm still dumbfounded they used 4x1500/4x750 instead. I mean, it's been around for years in BW, many of us (like every mapmaker) were saying this since they announced it.
You wouldn't think a bunch of super awesome (!!!) graphs would be required to see these things, but they are really cool (!!) and show the effect so well. I applaud your effort!
Makes me want to play around with different gradation arrangements and see if there are any major benefits to a linear distribution or shouldered plateau, etc. It'd be neat to use a "smoothness" metric and get a single number to represent each arrangement, sort of like your eye appraises the color distribution in the graphs.
Question: in building your graphs, did you use an algorithm or game data? Because for n < 2N (n workers, N patches), worker micro would significantly change mine-out times, and I think some of the results you point out would differ significantly. Most importantly perfect worker control in LotV would smooth things out a lot for 4+ bases, although it is unrealistic for real gameplay.
@frostalgia: While 6m/1hyg FRB has some desirable outcomes I don't think it's a panacea at all. For one, it does nothing about the economy cap, i.e. more bases doesn't improve income with the same #workers. So you just kick the can to 4base = 24 patches instead of 3base = 24 patches. It has a lot of balance issues as well, some major, such as zerg production being hugely buffed. It also makes it harder to expand when your total income is lower -- the price of a townhall as a proportion of income is higher. So in addition to delaying your expansion timing it also aggravates the liability of sinking money into a lower payoff instead of army. I think for it to work you'd need a lot of adjustments to balance things, and all townhalls would need to have their price reduced and possibly build time as well. At which point all you've really accomplished is +1'ing the number of bases players have. Is that the goal?
I assumed 6m/1hyg but you have 6m/2g. You realize this severely changes gas/min income ratio and therefore affects tech and balance in huge ways.
On May 15 2015 16:27 Teoita wrote:
I personally prefer the DH model, but i like how you took blizzard's idea and actually improved it massively!
We can have both! Wouldn't that be something.
|
Have you considered how mule usage would affect things for terran? (This is a failure in LotV as well.)
Forcing expansions because of less resources per base is not the right direction. The whole point of DHx (or DM) is that is REWARDS the player that expands. It only punishes the non-expanding player if the opponent expands and takes advantage of the extra income (which, if unchecked should be sufficient to trade until there is no 3 base deathball).
I also think BK had a good observation:
On May 15 2015 10:38 BronzeKnee wrote:
I don't understand why, but Blizzard erroneously believes that if people expand faster more action will result, and that will fix the three base turtle. That is only true if expanding isn't safe.
And isn't expanding rapidly part of the problem? People spend all their money in economy going up to three bases, that is why there is no action early and the game is so boring. If we make them go up to more bases rapidly, then even more time will be spent developing economy, unless expanding isn't safe. And if expanding isn't safe, won't we just be back to square one here, WOL Beta?
Rewarding the expansion (paying for itself quicker) is part of the goal of non-worker pairing models.
|
On May 15 2015 08:37 SC2John wrote:
I do have one concern though, and I would appreciate any correction on any point where I'm wrong. Basically, because this model is function of time rather than worker count, I feel that it encourages one expansion at a time rather than any kind of mass expansion. That is, you don't need to outexpand your opponent by much; for example, being on six bases does not actually grant any advantage over three bases in the short run, assuming each player is still mining optimally (or close to it). That said, I believe it may force a quicker fourth, but once each player has 32 nodes - X (mined out nodes), they will still be capped at a similar value to HotS, and while you can't turtle on 3 bases forever, you can still turtle relatively well and hit brutal timings with a hyper economy that can't be touched. This is a function of the time to mine out or significantly lose income to be too far into the game.
You are right in that we it is only the number of nodes which determines short-term gain. Thus 24 nodes (spread out over any number of bases) will mean that you are at income cap at 72 workers (max) or 48 workers (mining effective). Stated in other words: this is the HotS 3-base play. Taking more bases quickly will not improve this in the short term, but here is a key point: You actually want this effect. Defensive play with a small amount of bases should be a viable strategy and should put you on par with expanding players in the short-term. However, the advantage comes from long-term play.
Regarding your question of expanding quickly vs slowly, then the exact answer is given by the precise amount of mineral count chosen. Consider for instance the early-beta LotV verison with 4x750 + 4x1500 minerals. There was a magic number of 12 workers where you could pair workers on the large patches (8 workers) and put single workers on the small patches (4 workers). The patches would all disappear at the exact same time and you would experience no income drop whatsoever. Double-expanding and putting 2x12 workers was (and still is) significantly superior to single-expanding and putting 24 workers on your new expansion.
With the proposed DPM model, then this effect is smoothed out. You would not gain any immediate advantage, but do start gaining an advantage over time. The thing to note here is that the time-advantage does stack! By double-expanding you are able to significantly push back the income drop. For instance keeping a single harvester on the smallest patches (less than 16 workers per base) the first patch to disappear would be the second smallest. Using the numbers from of the TL;DR in the OP we would have
- 600 patch (2 workers, vanish after 7 minutes)
- 800 patch (2 workers, vanish after 10 minutes)
- 1000 patch (2 workers, vanish after 12 minutes)
- 1200 patch (2 workers, vanish after 14 minutes)
But only having 1 worker on the smallest nodes, for instance 13 workers the picture would be:
- 600 patch (1 worker, vanish after 14 minutes)
- 800 patch (1 worker, vanish after 20 minutes)
- 1000 patch (1 worker, vanish after 24 minutes)
- 1200 patch (2 workers, vanish after 14 minutes)
By double- or triple-expanding and keeping 13 workers at each base you are doubling the time of which you mine optimal with a given worker force.
|
|
|
|
|
|
EPT
