|
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 |
Canada13378 Posts
On April 12 2015 10:47 timchen1017 wrote:
I get your point of worker/patch ratio. Still I think, in the end the result is similar. I mean, if 4 patch/base, it is still pretty much impossible to get to 24 patches., or 6 bases. This is the same as 48 patches and 8 patches per base. And as long as the optimum patch number is larger than one can realistically get to, it always rewards expanding. I mean, with some number tuning and barring worker kill issues for now, 4 patch/base can pretty much work in the same way as your double harvest mining model, regarding to expand vs turtle issues, right?
If you use my point of view, what I am saying about the current protoss is that they are still relying the "extra" patches that will mine out sooner. Or the current balance forces them to do so. And that is what is wrong, instead of the fundamental economic model.
Or to put it another way, LotV actually adds a new layer of the strategy, that you may do a "timed" turtle using those small patches. Maybe the problem is then if you make the "real" turtle viable the "timed" turtle becomes too strong; so "real" turtle becomes not viable. But maybe that is what Blizzard intended. There is still a build of a bit turtling, but that build can not turtle forever.
From this perspective, another different thing is probably that the reward for mass expanding will be masked by this temporary bonus. That is whenever you expand to a new base, in this point of view you actually expand into two, before those patches mine out, and that is probably too much.
I think the issue is that the "timed" optimal base count becomes very different for different races. Zerg in LotV can hit an optimal worker ratio to patch ratio with only 4 hatches. Terran might be able to hit a 4 CC? Protoss imo cannot get to 24 patches easily as it is now and it partially is responsible for the difficulties protoss is having.
I think its better to give everyone the same tools and allow for turtle strategies honestly. Now, we could and probably should drop below 1500 mineral patches. I think 1400 is a good compromise tbh in double harvest (since the time is not just influenced by total minerals but total mining time going up too).
I agree that we can balance it so protoss doesn't need "extra patches" but what if they are able to get the "extra" patches, it becomes a bit of an issue. Its a lot harder to balance protoss around a reduced economy than it is to balance around a slightly faster economy (mineral only, gas is the same).
With regards to your "timed turtle" I argue double harvest achieves this same level of strategy with a little extra depth. The defending player will need to scout workers being cut or not, and will have a greater income on same worker (or higher) count. The sheer income advantage can translate to a better defense from the player if the turtling player waits too long - which achieves the same purpose - you are on a timer except its controlled by the opponents economy.
Timed turtling is also too short IMO. The issue there is that tech becomes difficult to obtain. If it takes 2 minutes to obtain blink and you mine out at 6:30, and you need blink to hold a third base, AND blink can only start at 3 mins in (in order to not die to unscoutable builds) it punishes the blink player for teching to blink.
I don't think reducing research times across the board is a good option either. So in this sense - players are given fewer opportunities, or options to their choice to expand. Tech is more difficult to obtain if you expand a lot as you should be spending the gas on units to hold and minerals on the town hall and workers.
The relative value of each mineral in LotV is a lot higher, and the value of tech much lower is how i see the economy of LotV.
I guess we just fundamentally disagree but honestly, i think removing worker pairing is just a more effective way to remove the mining cap and unlock more bases while allowing the highest variety of strategies to be viable (while increasing counterplay to turtle strats).
I could even begin to go into the importance of higher income and its relation to trading against turtle play once the mineral cap is unlocked by removing worker pairing but that involves related net incomes and value in building banks etc.
Suffice to say if there was no 2:1 cap, i think turtle swarmhost or mech would be far far less problematic than we saw in HotS.
P.S. While we may disagree, i do like this discussion. Got me thinking on other things more and more
|
I'm curious how Double Harvesting looks beyond the 16 workers per base mark. Does net income continue to increase, but very slowly? Or does adding more than a 16th worker add no income as it happens in HotS with more than 24 workers?
Edit: This seems like the sort of thing that the article covered, but I have double checked and am indeed not finding it. If I'm missing something, let me know.
|
Canada13378 Posts
On April 12 2015 11:06 AmicusVenti wrote:
I'm curious how Double Harvesting looks beyond the 16 workers per base mark. Does net income continue to increase, but very slowly? Or does adding more than a 16th worker add no income as it happens in HotS with more than 24 workers?
Edit: This seems like the sort of thing that the article covered, but I have double checked and am indeed not finding it. If I'm missing something, let me know.
It increases very slowly from workers 17 to 24. At 24 (full mineral saturation) it mines the same as the Worker Pairing model (WoL, HotS, LotV if full base). The diffrence is negligible.
So in short - going up to 24 you will always have a net increase. The net increase is less and less worth the investment of more workers (no point in getting 17 unless you want to expand and spread the workers optimally to begin with or lost a base and need to transfer workers to save them).
See the excel file I uploaded and linked to in the conclusion or the screencap of my excel file in full res for the stats at 24.
I didn't do the actual curve 17-24 in full since at 16 it begins to match HotS, and at 17 it might be more than HotS but its efficiency is significantly worse the close you get to 24 which balances out to be the same in the end
|
I only skimmed so far (sorry), but you really want the bar chart y-axis to go all the way down to 0, or they are very misleading.
Also don't approve of the random inserted pictures that are only there to break the text.
You know I agree with your point though.
Edit:
Let me post the plot I posted in the other thread, that sums up the key concept as I think about it. You touch on it in the last few plots, but it's not quite there for me.
![[image loading]](http://i.imgur.com/f0pUpWS.jpg)
Point being that once on 3 (or "sufficient" amount of) bases, there is no economical gain from taking bases, and thus map control. Turtling up on three bases gives me as much income as taking every base on the map.
So in sc2 it isn't really possible to play "Sauron Zerg", where you stay low tech, but expand like mad and build ridiculous amounts of low-tech units. Point was that by maintaining map control your drones were more valuable over many bases, and I could afford inefficient trades. In hots, once your opponent is on 3 bases, I can never afford inefficient trades as I cannot mine faster no matter how many bases I take.
|
Very, very nice article. Thank you so much for putting the work into it! I noticed that gas mining was not discussed, and while I understand that the minerals side of things is what is motivating the changes, it would seem odd to have a different algorithm for gas mining than for mining minerals. Would this create problems with teching too fast, or was analyzing the effects of double harvesting gas simply too much trouble for something that isn't causing trouble in HotS?
|
I would like Zeromus to know that I read the entire article and absolutely loved it. Well done. I've been watching this game and playing it for over 12 years and can honestly say this is some of the best stuff I've seen written about an inherent flaw and strategy stifling aspect of SCII.
|
Canada13378 Posts
On April 12 2015 11:23 feanaro wrote:
Very, very nice article. Thank you so much for putting the work into it! I noticed that gas mining was not discussed, and while I understand that the minerals side of things is what is motivating the changes, it would seem odd to have a different algorithm for gas mining than for mining minerals. Would this create problems with teching too fast, or was analyzing the effects of double harvesting gas simply too much trouble for something that isn't causing trouble in HotS?
Gas can mine the same way, changing the way it interacts with patch is a different set of behaviours than the gas geysers.
Also to the person who doesn't like the random pictures to break up text: after years of TL articles I and others have learned, if you dont break up the text no matter how interesting people will get real bored and stop reading really fast haha.
They also help show how the mining looks for some visual aids for more visual learners.
Starting at 0 is not always best, but rather putting enough information to allow people to not be mislead so long as they remain critical of the images is the best. I am putting a little bit of I guess, respect to you and assuming you won't just look at pics but rather read what goes along with it. The graphs are for support, not replacement. Also 5 years of grad school has taught me some basic use of images and graphs to support points without being purposefully misleading.
On April 12 2015 11:32 bo1b wrote:Great post, but this graph is a little dodgy in its implications tbh ![[image loading]](http://www.teamliquid.net/staff/ZeromuS/Economy/fourbasev3basehots.png)
The graph clearly states "bob vs chris example" and should be viewed as a visual representation (for ppl who like images) to help them better understand the long text. Its intended for visual learners to have a little support alongside what is otherwise, IMO, convoluted text. Its not meant for people who "get" the complicated text 
|
Great post, but this graph is a little dodgy in its implications tbh
|
Canada11219 Posts
Very nice article. I would definitely go for the Double Harvest model as the first experiment. I like the paradigm shift from 'three base is the problem' to '24 nodes is 100% efficiency is the problem.' I hadn't thought of it quite like that, but it is so true.
However it is done, I think it is very important to create inefficiency into the economic system. If you have X patches, then 1:1 ratio should be 100% and every worker added after that should still gain income but at a noticiable less than 100% efficiency.
I think this is an absolute requirement for a worker-model RTS game. Because we have to think, if workers always mine at 100% efficiency, why exactly do we have workers at all rather than simply build income generating buildings? Buildings that passively generate income is a MUCH simpler economic model if we simply want 100% efficiency. (Or static inefficiency- like Battle for MiddleEarth.) Is retaining the worker-model just a nostalgic hold-over that Blizzard never bothered to update?
One reason to have a worker-model RTS is centred around harassment. Yes you can kill buildings as part of harassment. But being able to dodge storms, reaver shots, disrupter shots, tank shots or being able to flee form hellion or vulture harassment or lurker or DT drops is inherently more interesting.
But the second reason is to incentivize expanding. You could have passive resource collection buildings built around the map on nodes like the SupCom series. But an inefficient worker system above X minerals creates greater depth in an worker-model RTS. I think it maximizes the use of the particular economic system, rather than try borrow (and thereby cheapen, what a system that it is not.) It's not just that the system creates inefficiencies- it is a DYNAMIC inefficient system. Depending on how many bases and how you distribute your workers and how you save them from harass, your income will fluctuate. This is a FAR more interesting system than 100% efficient passive generation OR inefficient but static income generation. (If you are not familiar, depending on where you place buildings in Battle for Middle Earth, the income can be anywhere from 100% efficient to 0% efficient. But once placed, the efficiency will never change.) The dynamic changes is what makes the worker model so interesting and creating inefficiencies capitalizes on this dynamic economic model.
Hypothetical Example:
Player 1 has 40 workers on two bases, but Player 2 has 32 workers on 3 bases. Player 2 has better income because their spread out workers are mining more efficiently. There is so much more flexibility in play- Player 1 got stuck on two bases, but can power workers and units to bust out, while Players 2 makes use of their map control to increase their mineral count. However, if Player 1 can bust out and expand, their superior worker count can spring them ahead once they spread the workers out. The Half Mineral method (I think) puts the trapped player on too much of a ticking time bomb- especially if they get trapped on one base. I definitely think inefficient workers allows for better comeback potential (which is always a good thing, in my opinion.)
|
On April 12 2015 11:32 bo1b wrote:Great post, but this graph is a little dodgy in its implications tbh
2:nd this. It looks like it's a huge difference from a glance, then you look at the y-axis...
Or worse: most don't.
In general, when you make bar charts, the area is intuitively interpreted as "how much" of whatever you are plotting, so to give a fair impression, you should ALWAYS let the bar chart y-axis go all the way down to 0.
|
I love this article but I still just cannot appreciate the over emphasis on mineral income.
zerg is grabbing quick forth in zvt because of the gas income. (and if not they could just go for a macro hatch)
Protoss grabbing new bases in sh zvp is for the boosted economy which is more on the gas.
Terran mech needs 8 gas for sky mech transition (also more for the gas)
The weaker (or none of some would say) mineral advantage for getting more than 3 base doesn't appear to be a problem to me.
Sh style still needs to move out for a 4th so they can maintain the 3 base mining efficiency for example.
It just doesn't lead to bases all around the map, which is supposed to mean engagement everywhere but in reality it just destroys the unit interaction for sc2 because mobility becomes a much bigger deal.
|
Excellent read, excellent data. I think a change along these lines will provide more incentive to expand and defense those expansions. I hope Blizzard takes this into consideration.
|
Canada13378 Posts
On April 12 2015 11:33 Falling wrote:+ Show Spoiler +Very nice article. I would definitely go for the Double Harvest model as the first experiment. I like the paradigm shift from 'three base is the problem' to '24 nodes is 100% efficiency is the problem.' I hadn't thought of it quite like that, but it is so true.
However it is done, I think it is very important to create inefficiency into the economic system. If you have X patches, then 1:1 ratio should be 100% and every worker added after that should still gain income but at a noticiable less than 100% efficiency.
I think this is an absolute requirement for a worker-model RTS game. Because we have to think, if workers always mine at 100% efficiency, why exactly do we have workers at all rather than simply build income generating buildings? Buildings that passively generate income is a MUCH simpler economic model if we simply want 100% efficiency. (Or static inefficiency- like Battle for MiddleEarth.) Is retaining the worker-model just a nostalgic hold-over that Blizzard never bothered to update?
One reason to have a worker-model RTS is centred around harassment. Yes you can kill buildings as part of harassment. But being able to dodge storms, reaver shots, disrupter shots, tank shots or being able to flee form hellion or vulture harassment or lurker or DT drops is inherently more interesting.
But the second reason is to incentivize expanding. You could have passive resource collection buildings built around the map on nodes like the SupCom series. But an inefficient worker system above X minerals creates greater depth in an worker-model RTS. I think it maximizes the use of the particular economic system, rather than try borrow (and thereby cheapen, what a system that it is not.) It's not just that the system creates inefficiencies- it is a DYNAMIC inefficient system. Depending on how many bases and how you distribute your workers and how you save them from harass, your income will fluctuate. This is a FAR more interesting system than 100% efficient passive generation OR inefficient but static income generation. (If you are not familiar, depending on where you place buildings in Battle for Middle Earth, the income can be anywhere from 100% efficient to 0% efficient. But once placed, the efficiency will never change.) The dynamic changes is what makes the worker model so interesting and creating inefficiencies capitalizes on this dynamic economic model.
Hypothetical Example:
Player 1 has 40 workers on two bases, but Player 2 has 32 workers on 3 bases. Player 2 has better income because their spread out workers are mining more efficiently. There is so much more flexibility in play- Player 1 got stuck on two bases, but can power workers and units to bust out, while Players 2 makes use of their map control to increase their mineral count. However, if Player 1 can bust out and expand, their superior worker count can spring them ahead once they spread the workers out. The Half Mineral method (I think) puts the trapped player on too much of a ticking time bomb- especially if they get trapped on one base. I definitely think inefficient workers allows for better comeback potential (which is always a good thing, in my opinion.)
Comeback potential is a really big thing we saw in our analysis as well.
The fact that 8 workers makes 63% as many minerals as 16 is huge for comebacks. You dont lose 50% of your income for losing 8 workers (like you would in hots) you lose 37.
Now RELATIVE to your opponent you are still behind, just by not as much. Also, it makes harassment want to be greedier and it has to do more damage to be worth it. Trading 4 hellions for 8 workers is bad on a 16 full line. Trading them for 8 workers on an 8 mineral line is HUGE.
More dynamic imo.
On April 12 2015 11:45 ETisME wrote:+ Show Spoiler +I love this article but I still just cannot appreciate the over emphasis on mineral income.
zerg is grabbing quick forth in zvt because of the gas income. (and if not they could just go for a macro hatch)
Protoss grabbing new bases in sh zvp is for the boosted economy which is more on the gas.
Terran mech needs 8 gas for sky mech transition (also more for the gas)
The weaker (or none of some would say) mineral advantage for getting more than 3 base doesn't appear to be a problem to me.
Sh style still needs to move out for a 4th so they can maintain the 3 base mining efficiency for example.
It just doesn't lead to bases all around the map, which is supposed to mean engagement everywhere but in reality it just destroys the unit interaction for sc2 because mobility becomes a much bigger deal.
Hmmm. think of it this way.
If the three base player is turtling and making bank, the other player might throw resources at them to try and pick away at their bank. The problem is that the turtle player making bank and spending it is probably trading more efficiently in the trades.
Since both players make the same income per minute, the turtle player is more efficient and eventually starves out the player who isn't turtling because of efficiency in defense. That is, the aggressor who is expanding to take half the map is using their bank to harass instead of letting the bank grow appropriately. You see this happen a lot in PvZ vs old Swarmhosts. The protoss slowly starves themselves if they attack the Zerg.
If your income is actually larger on more bases, you can still build a bank while being inefficient in trades against the opponent. The Protoss could spend 1k on a bunch of zealots and instead of losing 1k, they could make that 1k back FASTER than the turtle player. This means trades slowly benefit the player with a lot of bases. With more of an income, the player with more bases can afford to throw units away and maybe, slowly, create a timing for themselves. They might be able to whittle enough away from the opponent then with more money and production hit again before the opponent can fully stabilise.
Doing the "hit again" in SC2 right now happens by spending an entire bank all at once and hoping you do it before the other guy does. With much more income you might get more chances to hit again, or have so much more production that you can effectively reload faster (if not more often).
|
Canada11219 Posts
On April 12 2015 11:45 ETisME wrote:
I love this article but I still just cannot appreciate the over emphasis on mineral income.
zerg is grabbing quick forth in zvt because of the gas income. (and if not they could just go for a macro hatch)
Protoss grabbing new bases in sh zvp is for the boosted economy which is more on the gas.
Terran mech needs 8 gas for sky mech transition (also more for the gas).
Well, that was true in BW as well- Zerg tended to expand a lot more because they are greedy gas guzzlers, but because of the ever increasing inefficiencies of a fully saturated mineral line, Zerg could actually have a substantially better (efficient) mineral income because they didn't need to have 20 drones per mineral line. So you'd actually tons of expansions and expansion hunting (or else drop play to punish over-expanding zergs.) It also meant they didn't mine minerals as fast- rather the opposite direction Blizzard is going with their Half Mineral idea. The drive for gas, complements the inefficient mineral system as spread out workers will garner more income.
(And then there was the fact that vespene gas never truly ran out, so into the late game you could actually be mining gas from 3-4 depleted bases and 2 fresh ones. I don't know how that would change SC2, but I still like that idea.)
|
One small critique of the original article. IMO whenever you have graphs you should make them start at 0 unless there's a very good reason not to. Anything else kind of defeats the purpose of having a graph in the first place. At first glance it seems like some of the differences in mining rates are WAY larger than they actually are because of this.
Awesome article though, I have expressed elsewhere that I think it's very well written. I hope Blizzard at least tries this system. Honestly my only reservation is that it's a bit haphazard and doesn't look as neat as the current system. Might make it very confusing for new players. Don't think this is a big issue though.
Also, what would the effect on the early game timings be due this change, because of slightly higher income at the start? Would you still get a gateway on 13 etc? Is this a big deal or not?
|
@ZeromuS
I only came from sc2 background and I don't really see any issue where mech can also starve out the mobile player which is what wol was like, it lead to interesting situation where map gets mined out and I like how it slowly gain more grounds and th re mobile player slowly loses the momentum or needs to tech switch to a strong deathball.
(just like bio vs mech or sh in zvp where protoss goes warp prism into colossus deathball into skytoss)
the true problem of turtling in sc2 imo is solely that turtling player usually is either defending perfectly and doesn't break or collapsed extremely quickly because of a hole in defense.
Snute zvp for example if played perfectly is very hard to break but when protoss like hero who always find a crack, it collapsed right away.
I think lots of economy proposal are merely a change in the gameflow of sc2 and not really necessarily a good change if we don't get the units to compliment with the changes.
I guess I am just not convinced until I see the game played out with other proposed economy.
|
United States7483 Posts
On April 12 2015 06:51 y0su wrote:
The double harvest method just seems... "unnatural"
Just curious, what kind of impact would it have if all expansions (after natural) were gold? (6 patch). Worker pairing would mean you'd get optimal harvesting at 12 instead of 16 so your 4 base model would be 51 on minerals (24 on gas). 16+16+12+7 would be better? (perhaps the gold mineral return per trip and overall resources per gold patch could be modified as needed)
edit:maths fail
This isn't a 4 base model though, with this system, optimal efficiency is 1 worker per patch, and thus you continue to see gains for every base you take until such a time as you have more mineral patches than you have workers mining minerals. If you can manage to take 5, 6, or 7 bases you will see a return on them with this.
Maps don't really allow for it very easily, and holding those bases will be very hard, but the fact that there is a return means it will sometimes be worth the risk of spreading out, and make for more interesting strategic diversity.
Also, in case this gets brought up, the spike growth at the start of claiming a new base early on does lead to a faster return on expansions, but the math demonstrates that the return is only around 15-20% faster than it previously was, and isn't too much of a concern in terms of being overly rewarding too quickly.
|
Canada13378 Posts
On April 12 2015 12:17 ETisME wrote:
@ZeromuS
I only came from sc2 background and I don't really see any issue where mech can also starve out the mobile player which is what wol was like, it lead to interesting situation where map gets mined out and I like how it slowly gain more grounds and th re mobile player slowly loses the momentum or needs to tech switch to a strong deathball.
(just like bio vs mech or sh in zvp where protoss goes warp prism into colossus deathball into skytoss)
the true problem of turtling in sc2 imo is solely that turtling player usually is either defending perfectly and doesn't break or collapsed extremely quickly because of a hole in defense.
Snute zvp for example if played perfectly is very hard to break but when protoss like hero who always find a crack, it collapsed right away.
I think lots of economy proposal are merely a change in the gameflow of sc2 and not really necessarily a good change if we don't get the units to compliment with the changes.
I guess I am just not convinced until I see the game played out with other proposed economy.
I mean you can still starve out the other person as mech, it just creates a bigger window for the opponent to try and kill you after a trade and more options to do so starting from the mid game is all
|
P.S. While we may disagree, i do like this discussion. Got me thinking on other things more and more
Actually I don't think I disagree with you entirely, I certainly agree that worker inefficiency can work out nicely. I just find that even though your proposal may work, you didn't address the real problem in current LotV economy enough to persuade designer to start from scratch again. Oh on this point maybe I initially disagree with you, in the sense that I think all the problems you mentioned can be tuned away via some number changing with the current setup, but in the end I think I realize what fundamentally the problem is, albeit in a different way.
As I have said, you think along the lines that the economy in LotV starts the same as SC2 but the smaller patches just creates some artificial pressure to expand-or-die. I think it is beneficial to think the other way around though, as I said, think of those small patches as extras (so that the long term optimum worker per base is 8) and argue from there.
From our discussion, I think the key problem probably is not the 2:1 pairing (if you just halve the number of the patches, it should work the same as 1:1 pairing, which is the case you are suggesting), but probably the scaling beyond the optimum ratio. I think in BW this scaling is better and this is what allowed turtle player to stay in fewer bases.
In contrast, the scaling beyond 2:1 in SC2 is horrible. But from my perspective, the LotV economy compensates the turtling player by giving them temporary access to more patches, thus increase the effectiveness of having more than 8 workers in a base. In fact it is the same effective before the extra patches mine out-- that is too much according to your argument I believe. But you also say the "timed" turtle window is too short (which I agree), which states the opposite I think... so right now I still think this is a conception issue.
I mean, I view the problem this way. In BW, aggressively expanding player will have 8 workers per base. turtle player will have up to 16. In the new LotV scheme, to first order the effect is actually the same: the aggressive player if he wants to completely avoid mining out issues he should also expand after 8 workers per base. (Probably more, like 12 so the extra patches mine out at the same time as the regular ones). The turtling player can have 16, the reward is linear (so better than BW), but is on a timer. There are severe issues in the current beta, say how protoss should take 3rd base, but the problem is not fundamentally different when switching to a different economy.
The fundamental difference, in my opinion, is just that current scheme creates a new time scale when you go full turtle mode-- which may or may not create a more interesting dynamic, once the imbalances are ironed out. Ideally I imagine it should be the case such that comparing to BW, the turtling player gains more advantage initially but that advantage has to be used to push out not long after.
Oh, but from this point of view the real problem probably manifests itself in a different way: the reward for the aggressive player to mass expanding is lost. It just gains the privilege of not mining out any time soon, but this is not something that will snowball. Probably this is the real issue you are talking about. Since the reward of aggressively expanding is delayed in this scheme, it is hardly effective in a exponential growing economic world.
So in the end, everyone is pushed to expand more, but only up to a certain number of bases. The potential mining out issue is not a reason for the aggressively expanding player to expand more. That is the core problem of the current scheme I guess.
|
Canada13378 Posts
Yeah, basically I feel that expansions should reward people from the start.
If you want greater rewards and scaling on turtling bases you can still do that with a similar scheme to Double Harvest, just apply double mining and accept the tradeoff of pulling early workers being really negative and harmful especially in mirrors (pvp 1 gate proxies, early pools ZvZ).
|
|
|
|
|
|
EPT
