| 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 | ||
|
Cascade
Australia5405 Posts
On April 12 2015 15:14 Loccstana wrote: Increase mineral mining time to 5 seconds and increase minerals returned to base to 8. Also increase supply cap to 250. Voila, all of SC2's economy problems solved. I'm curious what would happen if the supply cap was increased a bit. | ||
|
Pontius Pirate
United States1557 Posts
I had a notion of how to deal with this efficiency problem without even removing worker pairing + Show Spoiler + Mineral staleness Each mineral node returns 6 minerals per trip if fresh. Once it has been mined, it becomes stale for between 2.7 and 5.4 seconds, during which time any mining returns only 4 minerals. I'll dig up some graphs later, but I remember the yields being on the whole quite similar to double harvesting, in that they had a really big boost compared to vanilla HotS at 8 workers, but wasn't a whole lot bigger at 16, and wasn't far off from parity at 24 per base. | ||
|
lastride
2390 Posts
| ||
|
Grovbolle
Denmark3805 Posts
| ||
|
ZenithM
France15952 Posts
Haha yeah I thought it was a joke at first. Typically the kind of graph you would have found in the Lings of Liberty article. | ||
|
Lunker
9 Posts
I agree with the complaints about that one graph. A bar chart with 2 bars where you can't tell the size of the bars relative to either 0 or a third bar doesn't help inform people any more than just giving the numbers. One way log scales can be useful is that changing the numbers represented by two bars by the same percentage changes the length of the two bars by the same amount. It doesn't matter as much, but ideally the line graphs should be bar or dot graphs since you can only have a whole number of workers. It would be easier to read. Colorblind people might have problems with some of them too. There doesn't seem like there'd be much income advantage in taking more than 6 bases with only 48 workers mining minerals, and I'd like to be able to be on 8 mining bases and have a big army. Would this proposal work well with a raising of the supply cap to 250 or 300? | ||
|
timchen1017
37 Posts
On April 12 2015 13:21 SC2John wrote: No, it isn't. There are still 8 workers mining at 100% efficiency on 4 nodes. That is a 2:1 ratio. The scaling is a direct result of worker pairing, as we PROVED in the article. Worker pairing creates a linear economic growth up until 16 workers on 8 patches, and then plateaus. Past 3 bases (24 nodes), you need at least 9 extra workers on minerals to see gains on the investment. In a No Worker Pairing model, you immediately have more income by transferring workers from mineral lines that have more than 8. This produces a stronger scaling model than the current HotS economy. No, I am not saying there's no 2:1 scaling, just saying the scaling itself does not mean much. It really does not matter whether it's optimum to have 2 workers per patch or one in the vacuum. The problem is not that it scales linearly until 16 workers/base; in BW it scales linearly up to 8 workers a base as well. It is this couples with the 200 supply upper limit, as well as the poor scaling after 2 workers/patch. In the above example with 4 patches, I am just saying that yes you still have a 2:1 ratio, but now you will NOT have a 3 base saturation as a result. I dunno whether you want to say your point is then it is a 6 base saturation, but in any case it will not happen in practice as no one will expand that much. So the bottom line is, in the vacuum it is obvious 2:1 ratio alone will not cause 3-base saturation. You need the supply limit, poor scaling after optimum saturation, and 8 patches a base to reach that conclusion. LotV sort of modifies how there is 8 patches in a base, thus the conclusion is questionable, and IMO not well explained in the article. This is like the riddle of the disappearing coin: 3 men go into a shop to buy an item, which costs $30. Each of the men contributes $10. The shopkeeper realizes the price is actually $25 and he's overcharged them, so he gives each of the men back $1, but since they can't divide the last $2 evenly, they tip the shopkeeper. Because each of the men got back $1, it's as if they spent $9 each. 9x3 = $27 Add the other $2 that the shopkeeper kept, and you have $29. Where did the other dollar go? The answer is that trying to change 10x3 to 9x3 is faulty logic. Your example is based off of trying to turn the economy into something it is not, creating fuzzy math. Not sure what you want to say, but here you should say $27-$2=$25, so every one paid $9 for the $25 price and $2 tip. Nothing strange here. If you want to add up to $30, it is $9*3+$1*3=$30, where $1 is the money everyone gets back. There is nothing fuzzy here. If anything, it's your mind that is fuzzy, not the math. $9*3+$2=$29 is faulty, not fuzzy. First of all, BW had differing mineral patch numbers per bases, so just try to keep that in mind when you're critically comparing the two; otherwise you sound like someone who never played BW at all and is just trying to make a nostalgia argument. As a matter of fact I have never played BW. I only know about its economy through articles like this. How does that matter? A theoretical question can be asked whether I have played with it or not. Actually, now that I think about it, what IS the argument you're trying to make here? You went out of your way to try and disprove that mining efficiency had an effect on expanding, and then say that the current system is flawed. So what then is your actual proposal? If it's not that mining efficiency prevents players from being rewarded by taking expansions, what IS preventing those players from being rewarded? What I am saying is that now I do think the current system is indeed flawed. But I don't think the argument written in the article is sufficient to show it. Also the way the article spent time on various topics makes it very easy for Blizzard engineers to think it is another what if article that spent little time really thinking about the current system. As I have said, you need 8 patches, 200 supply, optimal 2:1 ratio, and horrible scaling afterwards to reach that conclusion. It is not obvious at all from the argument and data provided in the article such that the current LotV system fits that description. | ||
|
Hider
Denmark9376 Posts
| ||
|
Uvantak
Uruguay1381 Posts
On April 12 2015 16:10 Hider wrote: Good article, but I do wonder what happens to income rate past the 16h worker? I told Zeromuss that there should be a higher focus past the first 16 workers and I kind of got that, compared to how it was before at least. To check that info you must download the Excel sheet, OR check this link. https://docs.google.com/spreadsheets/d/1ecOUFukYut7X9mEnsb1RGoadA80ImwbQfKUH_tTPcp8/edit?usp=sharing | ||
|
Kruxxen
United States149 Posts
The pacing and income felt much more natural, and early game income boost seemed to accelerate your build just enough to get it developing quickly. As I said, I wasn't playing against opponents good enough to know the advantages of having more bases, and I only played a few games. However, the flow of the economy felt significantly better, and maynarding made a noticeable difference in income. I'm not seeing any significant drawbacks to this model. The work put into this article is astounding, well done. I really hope Blizzard is willing to test these changes instead of the current LotV model. We saw early on in the beta a positive response to the way LotV economy was working, but it very quickly turned around and many players are saying that the changes to units are more positive than the changes to the economy. Obviously the economy was forcing a playstyle that we'd like to see more of. These changes encourage that play style, which is what we really need. Turtle play has a place in SC, and should continue to be viable to the extent that you aren't completely overwhelmed economically by your opponent. If the other player can get a significant economic advantage while you turtle you will be more encouraged to harass to stay in the game. Other members can/have better explained how defensive play had a place in BW, so I'll leave that to them, but I just wanted to say that I'm all for these changes. And from what I've played it feels like a much more interesting economic model both than the proposed LotV changes, as well as the current HotS model. Edit: A lot of people have already touched on the Bob vs. Chris graph, so I won't harp on it too much, but I will say this. Having the y axis start from 0 would actually better serve to prove the point that you are trying to make. By seeing the tiny difference in income from a 4 base to 3 base player (with the 3 base player actually ahead in this situation) you better illustrate the flaw in the current system. Where I think we all agree the player with 4 base mining should have noticeably higher income. | ||
|
Hider
Denmark9376 Posts
I still hold the opinion that rewarding/punishing is a conception issue, rather than a real one. What I read from your reply is that the 'punishing' feeling comes from the non-sustainability or the requirement to expand at a timing. I can agree to that, but that is different from all economic considerations you are talking about in the article. The "reward vs punishing"-phrase is definitely meaningless in itself, and I am not sure how many people actually understands what it means. However, the difference between a BW'ish economy and a "force-bases" economy is that the former gives the players an opportunity to stay on fewer bases. Remember that 3 vs 4 vs 5-base income graph in the article? Think about the consequences here. Let's say you are zerg and supermobile, would you be interested in trying to get 5 quick bases? Yeh why not?. However, let's say you play an immobile protoss style or tank-heavy mech. Are you gonna go for a quick 5 bases? Unlikely as you cannot defend that many bases. In fact, you are more likely to stay on 3 bases as the extra benefit of being on 4 bases vs 3 bases isn't that signifciant. So this means you get scenarios where 3 vs 5 base is very possible, and as a consequence, the game gets balanced around this --> Immobile units more cost efficient than mobile units. In LOTV, however, there is a signficiant difference as everyone is forced to take bases a ton faster. One problem with this scenario can be illustrated with the following example: --> You fall behind in army strenght --> You can't maintain the current 3-base rate --> Forced down on 2 bases --> 33% less econ --> Can't reclaim 3rd base (no way of cathing up) --> Snowball effect into GG In BW, however, if you temporarily get behind as the mobile player you can still be up on bases (since its difficult for the mobile player to punish it immediately). As the immobile player, you can get your 3rd delayed and still have a solid income on 2 bases while getting more units out which allows you to reclaim 3rd (since you scale better as the immobile player). Thus, BW has a significant lower snowball effect than the LOTV-economy (which is good). With a LOTV economy, the entire game is balanced around everything being mobile!!! Thus, there are no reasons to expect you will be able to reclaim an expansion. And let's think about what would happen to the gameplay if we attempted to balance it around the immobile player constantly being on 2 bases vs mobile player constantly being on 3 active bases. First of, that would require a significant difference in cost efficiency. Secondly, it would create every incentive in the world for the immobile player to turtle as much as possible. The reward for investing into harass is very small as he in that proces delays his own expansion timing. E.g. if you play mech and want to secure a new base, you need a critical tank account. But investing into Hellions and Medivacs delays that. So the mech player is focussing all of his ressources on massing defensive Siege Tanks and dumping minerals into Turrets + walling off everywhere to prevent any type of armytrading. In the midgame, such a gameplay style has never been interesting as it simply results in very stale gameplay (in the late game it can be cool though). In a BW'ish econ, the immobile player can stay on fewer bases which allows him to invest into aggressive options. He isn't as heavily rewarded for acquiring new bases as fast possible as the real income assymetry is between a 5base vs 3 base econ. (An immobile player could perhaps acquire a faster 4 base, but he cannot spread him self out over 5 bases) On top of that, his aggressive options are actually more likely to do damage than under a HOTS-economy as the enemy is spread thinner (more opportunites to do damage). FYI: : I do agree this isn't a treatise. Its instead a well-written assesment of income rates between different economic system. But it definitely goes quickly over the whole incentive/reward/punishment-part, which I think is the most complicated part of the economy. | ||
|
GUNZx5
10 Posts
| ||
|
PharaphobiaSC2
Czech Republic85 Posts
If you want expending to be rewarding, than turtle players would never expand like now... | ||
|
solidbebe
Netherlands4921 Posts
Barring that however this is a very good article. The current state of expanding is very strange. And the lotv way of doing it solves nothing. I hope blizzard pays some attention to this idea. | ||
|
yuzisee
Canada10 Posts
Then, the most efficient allocation of workers is the most extreme ("all my workers are on gold patches") and the least efficient allocation of workers is the opposite ("all my workers are in one base"). If you have 40 workers, your income is highest on 10 bases. (Two per gold mineral patch x two gold mineral patches per base x 10 bases = 40 workers). If you have the same 40 workers, the _least_ efficient allocation of workers is on one or two bases (you'll be mining the most normal patches and the fewest gold patches). In this partial gold patch system, we are still rewarding those who expand (relative to the same number of workers), but we are only somewhat rewarding those who build more workers (relative to the same number of bases). Double Harvesting, on the other hand, may allows you to _almost always_ benefit from building additional workers, as well as _almost always_ benefit from building additional bases. There is one more consequence of all this discussion that hasn't really been brought up yet, and that's harassment. Blizzard wants harassment to always be viable, because it creates more non-stop action (in their words). One interesting property of the Heart of the Swarm economic model is that if you are able to kill off a couple of your opponent's bases but the workers escape, it is no longer as worthwhile to harass their worker lines (e.g. Hellion/Banshee harass is no longer incentivised) because they are oversaturated anyway. The reverse is also true. If you are able to kill off a bunch of the opponent's workers, it is no longer nearly as valuable to try to kill their Nexii because they won't be able to saturate them anyway. This means that in Heart of the Swarm, strategic depth narrows as you make progress, rather than broadens -- and that's the opposite of a good strategy game. | ||
|
Musicus
Germany23576 Posts
| ||
|
Dirtyharry
Germany171 Posts
| ||
|
ShiQuRas1
2 Posts
I was just wondering if you have any thoughts on a 8 worker start with the Double Harvest model. Have you done any experiments on that? You would have a 1:1 ratio right form the beginning and take away like 20-30 sec of the build time you would need for the initial 2 workers. I wonder if this is somewhat of a compromise as well as your 1400 mineral-patch idea in the spirit of Blizzards' new policy to speed the game up a bit. Or do you think the build time of the first 2 workers is crucial to completely shutting down 6 (now 8) pool strats or proxies like the LotV-12-worker start does? Greets ShiQuRas | ||
Whitewing
United States7483 Posts
On April 12 2015 16:04 Lunker wrote: Nice post. I agree with much of what's been said in the OP and elsewhere about the three base cap, and hope Blizz looks at the issue closer. I agree with the complaints about that one graph. A bar chart with 2 bars where you can't tell the size of the bars relative to either 0 or a third bar doesn't help inform people any more than just giving the numbers. One way log scales can be useful is that changing the numbers represented by two bars by the same percentage changes the length of the two bars by the same amount. It doesn't matter as much, but ideally the line graphs should be bar or dot graphs since you can only have a whole number of workers. It would be easier to read. Colorblind people might have problems with some of them too. There doesn't seem like there'd be much income advantage in taking more than 6 bases with only 48 workers mining minerals, and I'd like to be able to be on 8 mining bases and have a big army. Would this proposal work well with a raising of the supply cap to 250 or 300? Unfortunately, that's outside the realm of an economic system, as the base count you can secure is limited by maps and map design, which is rather prohibitive in that regard. Having 6 bases quickly enough that none of them are mined out by the time you've taken the 6th is already rather unlikely, 8 is pretty impossible to design a map for as is, unless you're taking your opponent's bases or something ridiculous (in which case, win the game already). Raising the supply cap is pretty much out of the question: it's set at 200 deliberately by Blizzard because they want the game to run on lower end machines which wouldn't be able to handle more units. I would expect that to remain non-negotiable. On April 12 2015 17:44 ShiQuRas1 wrote: Hi, i have read thorugh the whole article (but not thrpugh the comments) and i think the study is really interesting and your suggested chagnes are really good. When i heard about the LotV mineral changes and the 12 worker start my biggest concern as well was to how it would super early game strategies away from the game... i know we all hate cheese :-P but it's a part of the game that especially in tournaments is super fun and exciting to watch. I was just wondering if you have any thoughts on a 8 worker start with the Double Harvest model. Have you done any experiments on that? You would have a 1:1 ratio right form the beginning and take away like 20-30 sec of the build time you would need for the initial 2 workers. I wonder if this is somewhat of a compromise as well as your 1400 mineral-patch idea in the spirit of Blizzards' new policy to speed the game up a bit. Or do you think the build time of the first 2 workers is crucial to completely shutting down 6 (now 8) pool strats or proxies like the LotV-12-worker start does? Greets ShiQuRas This is something I've done a fair amount of thinking on. Firstly, we'll have another article coming regarding the 12 worker start, so this will wait for that. However, if blizzard very much wants to start with more workers than the previous 6 to accelerate the start, I would absolutely advocate for 8 as the best change to make. An 8 worker start, I think, is the best decision. 12 is too many, and we'll go over why. Look for that article! | ||
|
hewo
Norway119 Posts
+ Show Spoiler + On April 12 2015 15:11 Cascade wrote: Ok, so I went back through the thread and saw that you already discussed the cut y-axis on the bar plots. I didn't read the thread before posting, sorry about that. I assume that is why you ignore the last 4 or 5 people that pointed out the mistake.  I feel that your most elaborate reply to the issue, that a lot of people raise, is this post: So while it almost always is a bad idea to cut a bar chart (I really can't come up with any exceptions tbh), as I explained in my previous post (people read the area as the "amount of stuff"), there are ofc always exception and you are right that it depends on what point you want to make. So let me explain in detail why this plot is a bad idea in this specific case. First, what people see when they look at the plots, is "whoa, that red bar is a lot larger than the blue bar!". This goes for anyone, no matter the experience or background or anything. The first thing you notice when you look at the plot is that there is a lot more red than blue. Is that the point you want to get across? Probably yes, as that would fit with the story you are trying to tell, so you are probably happy with the plot giving that impression. It is definitely misleading though, as the difference is like 5%. So it gives the impression that there is a huge relative difference, while it is actually very small, which is misleading no matter how you see it. You say that they point is to show that they are 100 minerals different, focusing that it is important that they are different at all, and that the fact that it is a difference between 2000 and 2100 is of less importance. Ok, that's fine. But if you are just after the difference, not the absolute values, then plot the difference... A single number. With two bars you need to read the y-axis and take the difference to find the difference. So it doesn't do a good job of showing how big the difference is. It shows that there is a difference, but the plot would look exactly the same if the difference was 1000 minerals, 100 minerals or 1 mineral, so the visual bars don't help at all to distinguish that difference. So the point you actually want to make, or SHOULD want to make at least, is to show how big the difference is compared to other things. For example, how many mining drones does it correspond to? How big is the difference compared to a saturated base? How big is the difference compared to the other comparisons you are doing in the analysis? So a much better way to illustrate your point would be a plot, bar-chart is fine, of "gained income" from different things. I threw one together in an online plotting program I just googled, but something along the line of this ![[image loading]](http://imagizer.imageshack.com/img673/3591/dIrcVy.png) You now get a clear understanding of how big the difference is, and you would immediately spot if the difference is 1 mineral/minute or 1000 minerals per minute. Thanks. ^Huge spoilered quote... This is what I was trying to get at in my original posts. Choosing a graph type that is not really displaying what you want it to display is in itself contradictory. Cascade illustrates this very well, and I think this sentence also is key: "It shows that there is a difference, but the plot would look exactly the same if the difference was 1000 minerals, 100 minerals or 1 mineral, so the visual bars don't help at all to distinguish that difference". Zooming in makes the actual numerical difference irrelevant, but that's what you want to illustrate (as you stated). Yes you label the axis, but reading and understanding the labels is just as big of a barrier for people as reading the text and understanding it that way. Thanks Cascade for your effort! I found it hard to put my thoughts down in words last night. | ||
| ||
|
|
Epic.LAN
Big Brain Bouts
sebesdes vs SpeCial
Harstem vs YoungYakov
GgMaChine vs uThermal
CranKy Ducklings
Epic.LAN
CSO Contender
BSL20 Non-Korean Champi…
Bonyth vs Sziky
Dewalt vs Hawk
Hawk vs QiaoGege
Sziky vs Dewalt
Mihu vs Bonyth
Zhanhun vs QiaoGege
QiaoGege vs Fengzi
Sparkling Tuna Cup
Online Event
BSL20 Non-Korean Champi…
Bonyth vs Zhanhun
Dewalt vs Mihu
Hawk vs Sziky
Sziky vs QiaoGege
Mihu vs Hawk
Zhanhun vs Dewalt
Fengzi vs Bonyth
Esports World Cup
ByuN vs Astrea
Lambo vs HeRoMaRinE
Clem vs TBD
Solar vs Zoun
SHIN vs Reynor
Maru vs TriGGeR
herO vs Lancer
Cure vs ShoWTimE
[ Show More ] |
|
|
EPT![[image loading]](http://www.teamliquid.net/staff/ZeromuS/Economy/fourbasev3basehots.png)
![[image loading]](https://pbs.twimg.com/media/B3ZPiyZCMAAjU0s.jpg)
