|
|
|
|
Personally this isn’t drastic enough. I’d use the last expansion to change SC2 into something like this:
- 6 nodes per base
- change mining saturation to 2 workers maximum instead of 3 workers
- compensate by increasing minerals per trip (normal node 7 minerals, high yield 10 minerals)
- use high yield nodes and half capacity nodes to balance expansion and tech requirements at certain stages of the game (up for consideration)
In other words, the reason to expand beyond 3 bases shouldn’t be just that you’re mining out the oldest one but that you want larger income.
|
Canada13378 Posts
On April 23 2015 01:16 Blackfeather wrote:Show nested quote +On April 23 2015 01:11 ZeromuS wrote:On April 23 2015 00:06 Blackfeather wrote:+ Show Spoiler +On April 22 2015 23:57 OtherWorld wrote:Show nested quote +On April 22 2015 23:53 Blackfeather wrote:On April 22 2015 23:46 [PkF] Wire wrote:On April 22 2015 23:37 Blackfeather wrote:On April 22 2015 23:35 [PkF] Wire wrote:
I actually think the downsides of 12 workers at the start are massively underestimated. By accelerating the decisions in the early game so much you create a lot of subtle problems (P spend a chrono less on probes, overlord scouts come very late, eco grows quickly and out of control) that you wouldn't have had with a 8 workers start for instance. I'm surprised no one is really questioning this decision. I think everybody agrees with Blizzard that the two minutes Blizzard took away were pretty boring for viewers and players alike. So while people see these disadvantages, those are mostly small racial balance issues that Blizzard could work out in time. Tbh I'm surprised Overlord didnt get a speed buff, but since most early all-ins need more preparation time than the twelve worker start gives them it currently doesnt seem to be a problem. The race which profits most of it is probably Terra because they produce worker at the slowest pace and have multiple no gas openings. Yeah but isn't there a happy medium between having those full two minutes during which nothing happens and literally skipping that time ? 8-9 workers at the start would massively reduce the downtime while not being as problematic IMO. I'm pretty sure Blizz did some testing on the numbers. Tbh from what I see on streams I dont see any balance problems aside from toss early game weakness, which they tried to solve with Adept buffs. The only real "problems" I see with this change is that all the early all-ins like double 10/10 rax bunker rush, two gate zealot and six and ten pools and super gas heavy timings like banshee rush are gone/weaker/later. But tbh I have yet to find a person who thought that these added anything to the game. Lol So many of competitive SC2's greatest moments involve early cheese... A matter of taste I think. 80-90% of them end in either side loosing immediately, which is freaking terrible for viewers and players. The ones where both side end up beaten up but not beaten usually transition into good games because they throw build orders out of the window and we get to see the minds behind the machines. On April 23 2015 00:05 ZeromuS wrote:Show nested quote +On April 22 2015 23:57 OtherWorld wrote:On April 22 2015 23:53 Blackfeather wrote:On April 22 2015 23:46 [PkF] Wire wrote:On April 22 2015 23:37 Blackfeather wrote:On April 22 2015 23:35 [PkF] Wire wrote:
I actually think the downsides of 12 workers at the start are massively underestimated. By accelerating the decisions in the early game so much you create a lot of subtle problems (P spend a chrono less on probes, overlord scouts come very late, eco grows quickly and out of control) that you wouldn't have had with a 8 workers start for instance. I'm surprised no one is really questioning this decision. I think everybody agrees with Blizzard that the two minutes Blizzard took away were pretty boring for viewers and players alike. So while people see these disadvantages, those are mostly small racial balance issues that Blizzard could work out in time. Tbh I'm surprised Overlord didnt get a speed buff, but since most early all-ins need more preparation time than the twelve worker start gives them it currently doesnt seem to be a problem. The race which profits most of it is probably Terra because they produce worker at the slowest pace and have multiple no gas openings. Yeah but isn't there a happy medium between having those full two minutes during which nothing happens and literally skipping that time ? 8-9 workers at the start would massively reduce the downtime while not being as problematic IMO. I'm pretty sure Blizz did some testing on the numbers. Tbh from what I see on streams I dont see any balance problems aside from toss early game weakness, which they tried to solve with Adept buffs. The only real "problems" I see with this change is that all the early all-ins like double 10/10 rax bunker rush, two gate zealot and six and ten pools and super gas heavy timings like banshee rush are gone/weaker/later. But tbh I have yet to find a person who thought that these added anything to the game. Lol So many of competitive SC2's greatest moments involve early cheese... they add a lot to the game honestly. The GSL finals of MKP (Boxer/Fake Boxer/Foxer back then) vs Nestea was amazing. Every game a 2 rax bunker rush. Slowly nestea figured out how to hold MKPs variant on the popular push and won the series in a 4-2 after losing i think 2 games early? Was brilliant. Fantastic series i suggest EVERYONE watches. On top of this relative timings for cheese still exist for T and Z. They still have the super early pool (12 pool for example) which is relative to P and T timings a stronger version of an 8 or 9 pool I think. T still has proxy rax, you can afford three now. P has nothing. There is no viable proxy pylon based play early like proxy gates because the pylon just starts so so so late thanks to 12 worker start and the time to scout it gets increased. I guess you could proxy 3 gate adept or stalkers but thats it. Proxy oracle play is far far weaker due to the no engi bay turret and it was never done PvZ after the no evo requirement for spores was put in place. You will still see it in pvp because relative timings in pvp dont change at all. I have seen proxy 4 gate zealot into adept vs Zerg on desrows stream once and he destroyed, so they arent totally out of the game for toss either. It's just not as much of an immediate loss anymore. On April 22 2015 23:50 ZeromuS wrote:Show nested quote +On April 22 2015 22:19 KaZeFenrir wrote:
I think you guys need a tldr for the differences between dh10 and lotv, and why you are bothering with all of this in the first place? Seems like more than a few people don't understand.
From what I've read (correct me if I'm wrong):
Lotv: you are PUNISHED for not expanding since you mine out faster. 2 base with 16 workers is the same income as having 1 base with 16 workers.
Dh10: you are REWARDED for expanding because more than 8 workers mine innefccienctly. 2 base with 16 workers is MORE income than 1 base with 16 workers.
Maybe focus on simplifying it for the tl;dr people's so they have a general idea and aren't arguing about something they don't understand/misunderstood. The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post. The issue is that the mining income isn't linear as the worker count goes up. Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20. Show nested quote +On April 22 2015 21:25 timchen1017 wrote:
I am again sad to see this response. 2 points here:
1. why optimal saturation for both are 16-20 workers? HotS I understand, but DH10? how is optimal defined anyway? If you just look from the graph to get optimal income/worker, should it not be 8? You sure have reasons, but if you want to state this in a meaning-to-clarify article, you had better spell them out clearly.
2. Your comments on the total mineral amount. You failed to realize why Blizzard changes half of the patches to half of the amount. This is because when half of the patches are mined out, it creates a similar worker inefficiency problem if an expansion is not taken: the base can accommodate up to 12 workers to give max income, but 8 is optimal for income/worker. So there is again some expand vs. turtle choices to be made. If all mineral patches are same, no such effect to be seen.
In the end, the major problem I find is that LotV is never treated properly here. This line of thought that BZ should at least try the DH10 or similar model is really weird. Sure, if a theoretical/numerical treatement of LotV is done and compared with DH10 and we found that DH10 has lots of benefits then sure why not try it, but no. Actually most of the work done is comparing HotS and DH10. How half of the patches only contain half of the mineral amount affect things is always dealed with in a hand-waiving way, if at all. + Show Spoiler +In response to 1) we use the word "optimal" to mean - the most workers you would reasonably make to ensure you get good income from a base. In HotS the "optimal" amount is 16 because workers begin to bounce after that and you see diminishing returns. Or you should but the efficiency of up to about 19 workers is almost the same as 16, so putting 19 instead of 16 is still very good. This same income point of minerals per worker is retained in DH so if you happen to not have a free mineral line, you aren't hurt by going up to 16. If you do not have free mineral lines to put workers on then stopping at 8 is not in any way optimal. You are only making a small % of your peak income for that base. 16-20 gets you much closer to the peak income for the base without investing into workers who will take a long time to pay for themselves. 2) This post isn't responding to the LotV model specifically. I touched on it in my previous article and will make a comparison between DH and LotV in the followup more specific. But at its core we are arguing against worker pairing. Not specifically against LotV. We dislike the fact that bases mine out half way at first because it creates odd situations in gameplay and interaction between players. A consistent income that does not change is non-punishing. Sure you can say that everyone loses the half patches but the value of half patches is so much higher early in the game and so much less later in the game that early game contain to take a third strategies become extremely powerful. If you can contain someone to 2 base while taking a third in the LotV model the half patches mining out means the contained player loses 25% of their available income. If they can't break the contain with full econ then they instantly lose because they now are on 3/4 econ and still contained. If they break the contain on 3/4 income they are a full base behind but its more than just 1 base at that point, because the player who took the third could have unpaired the workers on their half patches to put them on the new base increasing the overall lifespan of a higher income rate (which while interesting decision making, puts them in an even more powerful position to the opponent). I'll have it in better writing later but at its core I am against half patches as a design decision to achieve the design goal of limiting the number of resources available on a map. I am also against limiting the resources available on the map by about 25% because it forces a timer on the entire map to be mined out and creates perhaps too much of a skirmish feel to the game. I can understand quicker games being better for broadcast, but when you lose 25% of the maps resources you are limited in how small a map can even get in terms of mapmaking beyond rush distances. you begin to have to worry about minimal number of available bases. And you begin to constrict the time of a match for broadcast benefit at the detriment of artfully played long macro games with lots of back and forth. I can again understand the design goal to reduce available resources mapwide in the effort to encourage more expanding and punish players who NEVER expand or play TOO defensively. I get that. It puts a pressure on the player whose sole goal in the game is cost efficient trades. You need to balance this alongside our goal of giving the cost inefficient player more income to make those inefficient trades in our model. Thats our goal as well. So you could combine the two if you are blizzard. Our model mines 5% quicker to begin with so if you couple that with 5% or 10% less resources on the map you have a more subtle but still impactful timer on the players that is less punishing than half patches and continues to reward expansion based play by providing so much more income that inefficient trades are made. Again, I feel very strongly that removing worker pairing from sc2 (however blizz wants to do it) alongside a less extreme implementation of the blizzcon LotV reduced mineral model (keep in mind fewer resources overall was ALSO an approach taken by the community in the past), would be possibly, hopefully, the silver bullet in this game for making it truly amazing and better than it is now. On April 22 2015 23:37 Blackfeather wrote:Show nested quote +On April 22 2015 23:35 [PkF] Wire wrote:
I actually think the downsides of 12 workers at the start are massively underestimated. By accelerating the decisions in the early game so much you create a lot of subtle problems (P spend a chrono less on probes, overlord scouts come very late, eco grows quickly and out of control) that you wouldn't have had with a 8 workers start for instance. I'm surprised no one is really questioning this decision. I think everybody agrees with Blizzard that the two minutes Blizzard took away were pretty boring for viewers and players alike. So while people see these disadvantages, those are mostly small racial balance issues that Blizzard could work out in time. Tbh I'm surprised Overlord didnt get a speed buff, but since most early all-ins need more preparation time than the twelve worker start gives them it currently doesnt seem to be a problem. The race which profits most of it is probably Terra because they produce worker at the slowest pace and have multiple no gas openings. Show nested quote +On April 22 2015 14:01 Plexa wrote:On April 22 2015 13:48 Blackfeather wrote:On April 22 2015 13:12 ZeromuS wrote:On April 22 2015 12:33 Blackfeather wrote:On April 22 2015 12:17 coolman123123 wrote:
Does anyone feel it's kind of inappropriate for the writers of a fansite to use it as a soapbox for their ideas? I'm okay with the initial post but now it's become a campaign to have the idea tested. We can sit around and pretend like we know for certain that this is the best course of action for the game, but we don't. It's just an idea, and unless something is positively the right thing to do, I don't think this proactive approach should be taken.
I feel that it's more inappropriate that the original post takes a small detail David Kim got wrong and uses it to say that he misunderstood the entire post, completely ignoring everything else DK said. As I wrote earlier I think David Kim perfectly understood the initial post aside from some small numerical things he didnt calculate thoroughly and gave his reasoning or at least sugarcoated understandable reasons why he doesnt want to do it.
Coming from Dota where every year the meta changes entirely because of balance patches Blizzard always patched way to defensively for my taste, Sc2 could be a way better game than it is. So I dont think it's wrong to try to force Blizzard to try this, because hell, they can just take the HotS map and release it as a LotV test map to see if they like the replays. If they dont, they dont have to implement it. We are talking about a working time of like 10 hours.
I just think that the way it is done is a little bit insulting, although that might not be intentional. Its possible we misunderstood but the baseline is the first point: - In the HotS resourcing model, the 2nd player has almost no econ advantage (due to it being difficult to fully saturate every base + how the mining works per base) which implies that there are 16 workers on the first two bases and 16 spread across 4 bases (in hots). This is the only way that the economy advantage is zero as per: the "no econ advantage" comment. Extrapolating this number as similarly applying to the following point: - In the community suggestion model the 2nd player will have near double the econ advantage (due to it being pretty easy to fully saturate every base) we assume that he is discussing the model in itself and not a comparison of DH to SC2 econ because of the final point: - In the Void model, we have something in between the above Which implies that they are comparing the current hots model to our recommended model. The only way for double income to happen in our model with 4 base vs 2 is double the workers. This implies that they understand that the 8 workers is the optimal number similar to 16 in the current model. alternatively they might think that 8 is the saturation point (similar to 24 in the current model). Which implies, to me that they are comparing 2 bases on saturation (8 being their assumption) compared to 4 bases on optimal cap (8 again across 4 bases being their assumption). If we take the idea that 8 is the saturation point/cap then they could be comparing 16/16 split (which makes no more money than 8/8 if the 8 point is cap in their understanding) to a 4 base 8/8/8/8 split. Their assumption that 8 is the cap (similar to 16 in current or maybe 24) is especially apparent (or can at least be logically understood to be their assumption) when you consider this part of their comment: (due to it being pretty easy to fully saturate every base) In the second part in discussing our model (again as a compare contrast where LotV falls in between) it makes sense to believe they see the cap at 8 and not being at 24 or optimal around 16, which Plexa clarifies in the OP. It makes sense to say that yes the 4 base player will EASILY double the income of the 2 base player because it is VERY easy to ONLY make an extra 16 workers for the other 2 bases and have double the income (because the bases are capped at 8 workers). This probably due to the fact that the talking point of "8 is better on each mineral line" of the first article was not clear enough in its distinction of: optimal counts and saturation points as plexa cleared up in this OP. In the points of summary Dkim presented, it is difficult to read into them as a comparison of the DH model to the HotS model. It reads to me like its a quick description of what hots does, a quick description of what they think DH does and how LotV's current econ fits in between. IMO the LotV econ is NOT in between the DH and HotS model. BUT if you consider the DH model as setting a cap on 8 workers you can begin to see exactly how LotV IS perfectly in between. Eventually 8 workers IS the optimal count for a LotV base (DH in their understanding of it) and it DOES begin just like HotS at the start of mining the base. In this sense the LotV econ IS exactly perfectly a mix between the two with regards to worker counts and "optimal" worker counts per base. What we felt was lost in translation somewhere along the line was that our change is more of a mid point between LotV trying to give players expanding more money (since they get to replace the mined out half patches + 4 more patches for a short period -- meaning reward for expanding), and HotS (consistent mineral mining). Except our consistency remains all the time from start of base to full mine out and the additional mineral patches are not "bonuses" alongside a replacement of patches but a full reward - you get it ALL if you can hold it and bigger benefits come from splitting properly across bases (which is another WAY of applying base management you will eventually see in LotV - ensuring only 8 workers are on a half mined out base). I hope that this breakdown describes why we thought they misunderstood. We aren't trying to insult them, we are trying to be as respectful as possible but taking the statements they made as a whole and when examining the bullet points specifically you can see how the only logical train of thought to their conclusion is one in which they believe 8 is the worker cap, which is unfortunately, not the case. Your statement makes perfectly sense and I agree that he probably misunderstood DH10. He explains however in his post that he dislikes the "imbalance" of teching vs expanding in the current version already and that DH10 would make expanding even better, which I assume still stands true with the right numbers (its far to early in the morning for me to calc that through, I might do so when I'm awake again). The original post of Plexa totally ignored this point and continued going about his wrong understanding, while even with the right numbers his point remains (I assume). So Plexa doesnt give him a single reason to rethink his position but only proves to the community that DK didnt understand your original post. Which can be easily taken as trying to make him look like an idiot. I might be just overthinking, it's pretty late here. Now you could argue that balancing tech vs expand can also be achieved by buffing tech while changing expand, but as I said earlier if we created a scenario where both sides need to be able to fight about map control not dependent of the mu or the tech level we are talking about half a year of rebalancing at least. On a side note: I'm going to bed, catching up with you later. The intention behind my post was to correct the misunderstanding. I don't want to do DK's analysis of the situation for him, it'd largely look like zeromus' first set of analysis or the analysis coming up in the second article in development. By correcting the misunderstanding I feel that Blizzard can re-assess the situation and determine whether they still think the "imbalance" exists to the same degree they otherwise thought. tldr; there's little value in taking down analysis which was based on a faulty premise to begin with. EDIT: I also posted this on reddit which deals with the same subject. Loomismeister
I don't think you addressed his main concern with completely adopting this efficiency idea. You saw one fallacious bullet point he made and immediately pointed out why the numbers were wrong, but the core point he was making still stands.
To truly crush his argument, point out why your solution does fit into blizzards proven game development strategy of being easily tuned and iterated. Point out why it is better to put pressure on expanding from the other player rather than from the map itself. Actually make a good argument and David Kim will see it.
If it isn't easily iterated, if it's not better pressure, then David Kim will never yield to community pressure simply because you've shown solidarity. The comment section of your response post is troubling to me because it seems like if blizzard doesn't implement your idea then the community is already preparing the doom and gloom. Plexa
Here's the logic behind my post. Zeromus's original post was analysed through the lens of faulty information. By correcting that misinformation and getting them to re-read the article they can then reform their opinion on the model.
There's little use in me re-writing zeromus's articles since they already explain (or will explain) the points that you raise.
EDIT: I also don't think its about one fallacious bullet point, under the assumption that 8 workers saturates a mineral line (which is roughly in line with the numbers in the Blizzard example) everything is completely different. It looks like they took the 1:1 efficiency ratio as saturation which is a fundamental misunderstanding of the model. No reasonable discussion can come from Blizzard thinking we're arguing one thing, when we're actually arguing something different. By first correcting this misunderstanding (first section) and the using their own example to illustrate the mechanics of the model (the second section) we feel like we put the discussion in the right place to continue. Alright I didnt know that you had another article coming on the topic, I read your OP as a single reply in the same way Loomismeister read it. Tbh I still fail to see how this rewards teching over expanding. Mining faster doesnt help teching if the enemy mines even faster because they are expanding. In the current version you would just get steamrolled. No doubt their current solution seems a bit forced, but the main thing in both versions is that expanding is largely beneficial while teching higher than t2 is most of the time not. DH10 just makes expanding past three bases even more rewarding compared to high tech. While I would prefer DH10 I can totally understand that DK sees the rebalancing problems coming with it and points out that they already have to do rebalancing with the current model or rebalance the current model. The way I read out your science.png adding a tenth worker adds 10 mins per minute compared to 55-57 before. The eleventh is way better but still over 20% inferior to workers before the ninth. So expanding once you hit 9 workers gives you massive advantages and things like zerg double expand creates economies that have 30-50% more economy than one base on the same worker count, making expanding pay itself after less than two minutes even if you stay on the same worker count. That is way shorter than LotV's half mineral patch system, essentially taking options like tech in the current balancing away. For Zerg DH10 means that if you have 24 workers going from two to three bases is better in terms of economy than using those mins to build more workers. For T and P it's pretty close. Totally disregarding the fact that mains have other advantages like being the main production building for Zerg and having chrono, scan and mule for Terra and that you can pump worker even faster, which accelerates the advantage even more. So yeah DK's concern that this overbuffs expanding vs teching still is totally valid. Have you tried to tune down the numbers a bit? Like what happens if the worker collects overall 8 minerals instead of 10? Just something that could make the jump to DH10 less extreme for Blizzard. Just because your diminishing returns start on the 9th worker doesnt mean you dont make that 9th worker. Also in the science.png image each line is calculated per worker. (math in HotS time below) So getting the 9th worker on one mineral line means each worker now makes 55 mpM compared to the 58 before -- on average. The 9th worker adds almost 40 minerals per minute to your income. IF he had a free mineral node to go on you would get almost 60 minerals a minute more. But this doesn't invalidate the income the 9th onward workers make for you. It only encourages you to spread them to a new base as soon as possible IF you are able to take and hold a base. The key interaction here is that it makes new bases MORE vulnerable because there will be workers there right away. In the current SC2 econ you dont put workers on the new base other than fresh workers you JUST made to transfer. So while this spreading of workers will benefit zerg immediately with their ability to spread out quicker, it will also provide a new challenge to zergs because they will have more mineral lines open to harass. In HotS players might come to a zerg third with a small hellion reaper push and find 4 drones meanwhile 32 drones are in the main and nat mining peacefully with a wall in with queen to block hellions. In DH10 the zerg has the option to spread to the third line sooner but open themselves up to hellions as a result. So if terran can for a short while contain those drones from getting their personal mineral patches for more money the terran can keep up with both mules and a 3rd cc being built. Zerg on the other hand might cut some drones (again keeping even with terran for it) and try to get an army to gain map control so that the hellion push cant deny saturating the third as they drone up behind the aggression. These kinds of small shifts in thought as to where you put the workers, how fast you put them there, what kind of defense you have for them, because you want their income to be better, this is all subtle small interactions that create more decision making and more action on the map. I didnt point out the 9th, I pointed out how inefficient the tenth is and as a result going over 9. The ninth is still ok (60% of the workers before) and once you overcome the fact that the tenth is barely existent (17% of a new worker on a new expansion) because you have reached like 14 workers going up to 16 is probably a good idea. However going from nine to eleven is less useful than adding a single worker on another base. It is close to ideal to have 9 on every base and then max out one base to 16. So overall ideally in the early game you want to have 24 to 27 workers on 3 bases and add a 4th once you have saturated one of the bases up to 16 because the tenth worker on a base will need 5 minutes to start generating value while the first nine need less than one. Having 27 mineral workers on three bases gives you close to 238 minerals per minte (19%) more compared to two bases, paying for the base in less than two minutes even without producing workers. I doubt that we need to discuss that this model is a strong buff for the expanding and map controlling race and would need heavy rebalancing of at least two of the three races. As I said before I think that a constant fight for map control and multi-pronged harassment is more interesting than deathball doom pushes and all-ins, but I doubt that Blizzard is going to go so far as to redevelop their different racial balancing entirely. They might just make starcraft 3 But you don't want only 9 workers. Even if the extra guys after that are inefficient, you don't want to do that because your overall income is still very poor compared to a person who has the extra workers. If you cut at 9 your gas income is much higher than your mineral income and you have a hard time affording anything properly. The games between CatZ and Iaguz showed that stopping at 9 workers as Z (the xpanding race) vs T (the more workers race) meant that anytime before catz was able to secure 5/6 bases he was behind in army workers and income. The entire time. And Iaguz rode that to victory. In basically every game. That is pretty interesting. Are you sure that that isnt a result of Mule? Because in a scenario with the low workercount Mule is going to have an immense impact. As I calculated through earlier expansions pay for themselves in less than two minutes and often are preferable to building more workers. However if you stay at 27 workers early on Mule is bound to have a massive impact. I might be really overlooking something, so if I was wrong that's all the better.
Mules on paper are weaker because overall scv mining is more in our model than standard hots. So relative to the SCVs mules are a bit weaker on low worker count bases (comparing same situation across the 2 models), and the same as scv counts reach 16+
Zerg was actually able to keep up with the mule income on more bases which is often not the case in HotS. If you look at standard sc2 the mule puts terran far ahead of the other 2 races in mineral income normally. Now the mule will still do this at early levels since a mule is better than the 9th 10th etc worker so early game income spike due to mules remains to retain build orders or help them along. But in the end the overall power of a mule compared to the other two races is a bit lower IF the terran gets out expanded by 2 bases. If they are on even or just 1 base down T is benefiting from mules much the same way they always have.
GREAT!
Now on the topic of expansions paying for themselves, whitewing did a differential analysis of some sort a DID i think? And he found that bases in our model pay themselves back somewhere between 15-25% faster than HotS depending on a number of factors like workers available for transfer, race, and base number. Natural bases will be on the lower end of the spectrum and thirds fourths fifths etc will be on the higher end of the pay themselves back quicker spectrum.
So this makes dedicated kill your opponent with big army timing windows a bit smaller, disincentivizing that over just trying to deny bases or workers through harass instead of big punch in the face to kill them outright since their econ investment put them SUPER behind in army for example.
|
I very strongly disagree that 11/11, proxy gates or early pools didn't bring a lot to the game. Most memorable series included disgusting all-ins. Hell, the most memorable game of that legendary Squirtle vs Mvp series is probably the Atlantis Spaceship proxy rax...
|
On April 22 2015 23:50 ZeromuS wrote:
The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post.
The issue is that the mining income isn't linear as the worker count goes up.
Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20.
Maybe instead of words, throw a graph, with some arrows and red letters "this is good", "this is bad", "this is what we achieve" 
|
On April 23 2015 02:11 BlackLilium wrote:Show nested quote +On April 22 2015 23:50 ZeromuS wrote:
The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post.
The issue is that the mining income isn't linear as the worker count goes up.
Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20. Maybe instead of words, throw a graph, with some arrows and red letters "this is good", "this is bad", "this is what we achieve"
Yeah, while I dislike having everything boiled down to a few pictures, I think it's necessary at this point. Too many people still seem to not get it. =/
|
On April 23 2015 01:44 [PkF] Wire wrote:
I very strongly disagree that 11/11, proxy gates or early pools didn't bring a lot to the game. Most memorable series included disgusting all-ins. Hell, the most memorable game of that legendary Squirtle vs Mvp series is probably the Atlantis Spaceship proxy rax...
Would DH make cheese more powerful? Because your first workers mine more, you have an advantage vs players that keep building workers normally compared to the HotS model.
|
Aotearoa39261 Posts
On April 23 2015 01:16 Blackfeather wrote:Show nested quote +On April 23 2015 01:11 ZeromuS wrote:On April 23 2015 00:06 Blackfeather wrote:+ Show Spoiler +On April 22 2015 23:57 OtherWorld wrote:Show nested quote +On April 22 2015 23:53 Blackfeather wrote:On April 22 2015 23:46 [PkF] Wire wrote:On April 22 2015 23:37 Blackfeather wrote:On April 22 2015 23:35 [PkF] Wire wrote:
I actually think the downsides of 12 workers at the start are massively underestimated. By accelerating the decisions in the early game so much you create a lot of subtle problems (P spend a chrono less on probes, overlord scouts come very late, eco grows quickly and out of control) that you wouldn't have had with a 8 workers start for instance. I'm surprised no one is really questioning this decision. I think everybody agrees with Blizzard that the two minutes Blizzard took away were pretty boring for viewers and players alike. So while people see these disadvantages, those are mostly small racial balance issues that Blizzard could work out in time. Tbh I'm surprised Overlord didnt get a speed buff, but since most early all-ins need more preparation time than the twelve worker start gives them it currently doesnt seem to be a problem. The race which profits most of it is probably Terra because they produce worker at the slowest pace and have multiple no gas openings. Yeah but isn't there a happy medium between having those full two minutes during which nothing happens and literally skipping that time ? 8-9 workers at the start would massively reduce the downtime while not being as problematic IMO. I'm pretty sure Blizz did some testing on the numbers. Tbh from what I see on streams I dont see any balance problems aside from toss early game weakness, which they tried to solve with Adept buffs. The only real "problems" I see with this change is that all the early all-ins like double 10/10 rax bunker rush, two gate zealot and six and ten pools and super gas heavy timings like banshee rush are gone/weaker/later. But tbh I have yet to find a person who thought that these added anything to the game. Lol So many of competitive SC2's greatest moments involve early cheese... A matter of taste I think. 80-90% of them end in either side loosing immediately, which is freaking terrible for viewers and players. The ones where both side end up beaten up but not beaten usually transition into good games because they throw build orders out of the window and we get to see the minds behind the machines. On April 23 2015 00:05 ZeromuS wrote:Show nested quote +On April 22 2015 23:57 OtherWorld wrote:On April 22 2015 23:53 Blackfeather wrote:On April 22 2015 23:46 [PkF] Wire wrote:On April 22 2015 23:37 Blackfeather wrote:On April 22 2015 23:35 [PkF] Wire wrote:
I actually think the downsides of 12 workers at the start are massively underestimated. By accelerating the decisions in the early game so much you create a lot of subtle problems (P spend a chrono less on probes, overlord scouts come very late, eco grows quickly and out of control) that you wouldn't have had with a 8 workers start for instance. I'm surprised no one is really questioning this decision. I think everybody agrees with Blizzard that the two minutes Blizzard took away were pretty boring for viewers and players alike. So while people see these disadvantages, those are mostly small racial balance issues that Blizzard could work out in time. Tbh I'm surprised Overlord didnt get a speed buff, but since most early all-ins need more preparation time than the twelve worker start gives them it currently doesnt seem to be a problem. The race which profits most of it is probably Terra because they produce worker at the slowest pace and have multiple no gas openings. Yeah but isn't there a happy medium between having those full two minutes during which nothing happens and literally skipping that time ? 8-9 workers at the start would massively reduce the downtime while not being as problematic IMO. I'm pretty sure Blizz did some testing on the numbers. Tbh from what I see on streams I dont see any balance problems aside from toss early game weakness, which they tried to solve with Adept buffs. The only real "problems" I see with this change is that all the early all-ins like double 10/10 rax bunker rush, two gate zealot and six and ten pools and super gas heavy timings like banshee rush are gone/weaker/later. But tbh I have yet to find a person who thought that these added anything to the game. Lol So many of competitive SC2's greatest moments involve early cheese... they add a lot to the game honestly. The GSL finals of MKP (Boxer/Fake Boxer/Foxer back then) vs Nestea was amazing. Every game a 2 rax bunker rush. Slowly nestea figured out how to hold MKPs variant on the popular push and won the series in a 4-2 after losing i think 2 games early? Was brilliant. Fantastic series i suggest EVERYONE watches. On top of this relative timings for cheese still exist for T and Z. They still have the super early pool (12 pool for example) which is relative to P and T timings a stronger version of an 8 or 9 pool I think. T still has proxy rax, you can afford three now. P has nothing. There is no viable proxy pylon based play early like proxy gates because the pylon just starts so so so late thanks to 12 worker start and the time to scout it gets increased. I guess you could proxy 3 gate adept or stalkers but thats it. Proxy oracle play is far far weaker due to the no engi bay turret and it was never done PvZ after the no evo requirement for spores was put in place. You will still see it in pvp because relative timings in pvp dont change at all. I have seen proxy 4 gate zealot into adept vs Zerg on desrows stream once and he destroyed, so they arent totally out of the game for toss either. It's just not as much of an immediate loss anymore. On April 22 2015 23:50 ZeromuS wrote:Show nested quote +On April 22 2015 22:19 KaZeFenrir wrote:
I think you guys need a tldr for the differences between dh10 and lotv, and why you are bothering with all of this in the first place? Seems like more than a few people don't understand.
From what I've read (correct me if I'm wrong):
Lotv: you are PUNISHED for not expanding since you mine out faster. 2 base with 16 workers is the same income as having 1 base with 16 workers.
Dh10: you are REWARDED for expanding because more than 8 workers mine innefccienctly. 2 base with 16 workers is MORE income than 1 base with 16 workers.
Maybe focus on simplifying it for the tl;dr people's so they have a general idea and aren't arguing about something they don't understand/misunderstood. The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post. The issue is that the mining income isn't linear as the worker count goes up. Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20. Show nested quote +On April 22 2015 21:25 timchen1017 wrote:
I am again sad to see this response. 2 points here:
1. why optimal saturation for both are 16-20 workers? HotS I understand, but DH10? how is optimal defined anyway? If you just look from the graph to get optimal income/worker, should it not be 8? You sure have reasons, but if you want to state this in a meaning-to-clarify article, you had better spell them out clearly.
2. Your comments on the total mineral amount. You failed to realize why Blizzard changes half of the patches to half of the amount. This is because when half of the patches are mined out, it creates a similar worker inefficiency problem if an expansion is not taken: the base can accommodate up to 12 workers to give max income, but 8 is optimal for income/worker. So there is again some expand vs. turtle choices to be made. If all mineral patches are same, no such effect to be seen.
In the end, the major problem I find is that LotV is never treated properly here. This line of thought that BZ should at least try the DH10 or similar model is really weird. Sure, if a theoretical/numerical treatement of LotV is done and compared with DH10 and we found that DH10 has lots of benefits then sure why not try it, but no. Actually most of the work done is comparing HotS and DH10. How half of the patches only contain half of the mineral amount affect things is always dealed with in a hand-waiving way, if at all. + Show Spoiler +In response to 1) we use the word "optimal" to mean - the most workers you would reasonably make to ensure you get good income from a base. In HotS the "optimal" amount is 16 because workers begin to bounce after that and you see diminishing returns. Or you should but the efficiency of up to about 19 workers is almost the same as 16, so putting 19 instead of 16 is still very good. This same income point of minerals per worker is retained in DH so if you happen to not have a free mineral line, you aren't hurt by going up to 16. If you do not have free mineral lines to put workers on then stopping at 8 is not in any way optimal. You are only making a small % of your peak income for that base. 16-20 gets you much closer to the peak income for the base without investing into workers who will take a long time to pay for themselves. 2) This post isn't responding to the LotV model specifically. I touched on it in my previous article and will make a comparison between DH and LotV in the followup more specific. But at its core we are arguing against worker pairing. Not specifically against LotV. We dislike the fact that bases mine out half way at first because it creates odd situations in gameplay and interaction between players. A consistent income that does not change is non-punishing. Sure you can say that everyone loses the half patches but the value of half patches is so much higher early in the game and so much less later in the game that early game contain to take a third strategies become extremely powerful. If you can contain someone to 2 base while taking a third in the LotV model the half patches mining out means the contained player loses 25% of their available income. If they can't break the contain with full econ then they instantly lose because they now are on 3/4 econ and still contained. If they break the contain on 3/4 income they are a full base behind but its more than just 1 base at that point, because the player who took the third could have unpaired the workers on their half patches to put them on the new base increasing the overall lifespan of a higher income rate (which while interesting decision making, puts them in an even more powerful position to the opponent). I'll have it in better writing later but at its core I am against half patches as a design decision to achieve the design goal of limiting the number of resources available on a map. I am also against limiting the resources available on the map by about 25% because it forces a timer on the entire map to be mined out and creates perhaps too much of a skirmish feel to the game. I can understand quicker games being better for broadcast, but when you lose 25% of the maps resources you are limited in how small a map can even get in terms of mapmaking beyond rush distances. you begin to have to worry about minimal number of available bases. And you begin to constrict the time of a match for broadcast benefit at the detriment of artfully played long macro games with lots of back and forth. I can again understand the design goal to reduce available resources mapwide in the effort to encourage more expanding and punish players who NEVER expand or play TOO defensively. I get that. It puts a pressure on the player whose sole goal in the game is cost efficient trades. You need to balance this alongside our goal of giving the cost inefficient player more income to make those inefficient trades in our model. Thats our goal as well. So you could combine the two if you are blizzard. Our model mines 5% quicker to begin with so if you couple that with 5% or 10% less resources on the map you have a more subtle but still impactful timer on the players that is less punishing than half patches and continues to reward expansion based play by providing so much more income that inefficient trades are made. Again, I feel very strongly that removing worker pairing from sc2 (however blizz wants to do it) alongside a less extreme implementation of the blizzcon LotV reduced mineral model (keep in mind fewer resources overall was ALSO an approach taken by the community in the past), would be possibly, hopefully, the silver bullet in this game for making it truly amazing and better than it is now. On April 22 2015 23:37 Blackfeather wrote:Show nested quote +On April 22 2015 23:35 [PkF] Wire wrote:
I actually think the downsides of 12 workers at the start are massively underestimated. By accelerating the decisions in the early game so much you create a lot of subtle problems (P spend a chrono less on probes, overlord scouts come very late, eco grows quickly and out of control) that you wouldn't have had with a 8 workers start for instance. I'm surprised no one is really questioning this decision. I think everybody agrees with Blizzard that the two minutes Blizzard took away were pretty boring for viewers and players alike. So while people see these disadvantages, those are mostly small racial balance issues that Blizzard could work out in time. Tbh I'm surprised Overlord didnt get a speed buff, but since most early all-ins need more preparation time than the twelve worker start gives them it currently doesnt seem to be a problem. The race which profits most of it is probably Terra because they produce worker at the slowest pace and have multiple no gas openings. Show nested quote +On April 22 2015 14:01 Plexa wrote:On April 22 2015 13:48 Blackfeather wrote:On April 22 2015 13:12 ZeromuS wrote:On April 22 2015 12:33 Blackfeather wrote:On April 22 2015 12:17 coolman123123 wrote:
Does anyone feel it's kind of inappropriate for the writers of a fansite to use it as a soapbox for their ideas? I'm okay with the initial post but now it's become a campaign to have the idea tested. We can sit around and pretend like we know for certain that this is the best course of action for the game, but we don't. It's just an idea, and unless something is positively the right thing to do, I don't think this proactive approach should be taken.
I feel that it's more inappropriate that the original post takes a small detail David Kim got wrong and uses it to say that he misunderstood the entire post, completely ignoring everything else DK said. As I wrote earlier I think David Kim perfectly understood the initial post aside from some small numerical things he didnt calculate thoroughly and gave his reasoning or at least sugarcoated understandable reasons why he doesnt want to do it.
Coming from Dota where every year the meta changes entirely because of balance patches Blizzard always patched way to defensively for my taste, Sc2 could be a way better game than it is. So I dont think it's wrong to try to force Blizzard to try this, because hell, they can just take the HotS map and release it as a LotV test map to see if they like the replays. If they dont, they dont have to implement it. We are talking about a working time of like 10 hours.
I just think that the way it is done is a little bit insulting, although that might not be intentional. Its possible we misunderstood but the baseline is the first point: - In the HotS resourcing model, the 2nd player has almost no econ advantage (due to it being difficult to fully saturate every base + how the mining works per base) which implies that there are 16 workers on the first two bases and 16 spread across 4 bases (in hots). This is the only way that the economy advantage is zero as per: the "no econ advantage" comment. Extrapolating this number as similarly applying to the following point: - In the community suggestion model the 2nd player will have near double the econ advantage (due to it being pretty easy to fully saturate every base) we assume that he is discussing the model in itself and not a comparison of DH to SC2 econ because of the final point: - In the Void model, we have something in between the above Which implies that they are comparing the current hots model to our recommended model. The only way for double income to happen in our model with 4 base vs 2 is double the workers. This implies that they understand that the 8 workers is the optimal number similar to 16 in the current model. alternatively they might think that 8 is the saturation point (similar to 24 in the current model). Which implies, to me that they are comparing 2 bases on saturation (8 being their assumption) compared to 4 bases on optimal cap (8 again across 4 bases being their assumption). If we take the idea that 8 is the saturation point/cap then they could be comparing 16/16 split (which makes no more money than 8/8 if the 8 point is cap in their understanding) to a 4 base 8/8/8/8 split. Their assumption that 8 is the cap (similar to 16 in current or maybe 24) is especially apparent (or can at least be logically understood to be their assumption) when you consider this part of their comment: (due to it being pretty easy to fully saturate every base) In the second part in discussing our model (again as a compare contrast where LotV falls in between) it makes sense to believe they see the cap at 8 and not being at 24 or optimal around 16, which Plexa clarifies in the OP. It makes sense to say that yes the 4 base player will EASILY double the income of the 2 base player because it is VERY easy to ONLY make an extra 16 workers for the other 2 bases and have double the income (because the bases are capped at 8 workers). This probably due to the fact that the talking point of "8 is better on each mineral line" of the first article was not clear enough in its distinction of: optimal counts and saturation points as plexa cleared up in this OP. In the points of summary Dkim presented, it is difficult to read into them as a comparison of the DH model to the HotS model. It reads to me like its a quick description of what hots does, a quick description of what they think DH does and how LotV's current econ fits in between. IMO the LotV econ is NOT in between the DH and HotS model. BUT if you consider the DH model as setting a cap on 8 workers you can begin to see exactly how LotV IS perfectly in between. Eventually 8 workers IS the optimal count for a LotV base (DH in their understanding of it) and it DOES begin just like HotS at the start of mining the base. In this sense the LotV econ IS exactly perfectly a mix between the two with regards to worker counts and "optimal" worker counts per base. What we felt was lost in translation somewhere along the line was that our change is more of a mid point between LotV trying to give players expanding more money (since they get to replace the mined out half patches + 4 more patches for a short period -- meaning reward for expanding), and HotS (consistent mineral mining). Except our consistency remains all the time from start of base to full mine out and the additional mineral patches are not "bonuses" alongside a replacement of patches but a full reward - you get it ALL if you can hold it and bigger benefits come from splitting properly across bases (which is another WAY of applying base management you will eventually see in LotV - ensuring only 8 workers are on a half mined out base). I hope that this breakdown describes why we thought they misunderstood. We aren't trying to insult them, we are trying to be as respectful as possible but taking the statements they made as a whole and when examining the bullet points specifically you can see how the only logical train of thought to their conclusion is one in which they believe 8 is the worker cap, which is unfortunately, not the case. Your statement makes perfectly sense and I agree that he probably misunderstood DH10. He explains however in his post that he dislikes the "imbalance" of teching vs expanding in the current version already and that DH10 would make expanding even better, which I assume still stands true with the right numbers (its far to early in the morning for me to calc that through, I might do so when I'm awake again). The original post of Plexa totally ignored this point and continued going about his wrong understanding, while even with the right numbers his point remains (I assume). So Plexa doesnt give him a single reason to rethink his position but only proves to the community that DK didnt understand your original post. Which can be easily taken as trying to make him look like an idiot. I might be just overthinking, it's pretty late here. Now you could argue that balancing tech vs expand can also be achieved by buffing tech while changing expand, but as I said earlier if we created a scenario where both sides need to be able to fight about map control not dependent of the mu or the tech level we are talking about half a year of rebalancing at least. On a side note: I'm going to bed, catching up with you later. The intention behind my post was to correct the misunderstanding. I don't want to do DK's analysis of the situation for him, it'd largely look like zeromus' first set of analysis or the analysis coming up in the second article in development. By correcting the misunderstanding I feel that Blizzard can re-assess the situation and determine whether they still think the "imbalance" exists to the same degree they otherwise thought. tldr; there's little value in taking down analysis which was based on a faulty premise to begin with. EDIT: I also posted this on reddit which deals with the same subject. Loomismeister
I don't think you addressed his main concern with completely adopting this efficiency idea. You saw one fallacious bullet point he made and immediately pointed out why the numbers were wrong, but the core point he was making still stands.
To truly crush his argument, point out why your solution does fit into blizzards proven game development strategy of being easily tuned and iterated. Point out why it is better to put pressure on expanding from the other player rather than from the map itself. Actually make a good argument and David Kim will see it.
If it isn't easily iterated, if it's not better pressure, then David Kim will never yield to community pressure simply because you've shown solidarity. The comment section of your response post is troubling to me because it seems like if blizzard doesn't implement your idea then the community is already preparing the doom and gloom. Plexa
Here's the logic behind my post. Zeromus's original post was analysed through the lens of faulty information. By correcting that misinformation and getting them to re-read the article they can then reform their opinion on the model.
There's little use in me re-writing zeromus's articles since they already explain (or will explain) the points that you raise.
EDIT: I also don't think its about one fallacious bullet point, under the assumption that 8 workers saturates a mineral line (which is roughly in line with the numbers in the Blizzard example) everything is completely different. It looks like they took the 1:1 efficiency ratio as saturation which is a fundamental misunderstanding of the model. No reasonable discussion can come from Blizzard thinking we're arguing one thing, when we're actually arguing something different. By first correcting this misunderstanding (first section) and the using their own example to illustrate the mechanics of the model (the second section) we feel like we put the discussion in the right place to continue. Alright I didnt know that you had another article coming on the topic, I read your OP as a single reply in the same way Loomismeister read it. Tbh I still fail to see how this rewards teching over expanding. Mining faster doesnt help teching if the enemy mines even faster because they are expanding. In the current version you would just get steamrolled. No doubt their current solution seems a bit forced, but the main thing in both versions is that expanding is largely beneficial while teching higher than t2 is most of the time not. DH10 just makes expanding past three bases even more rewarding compared to high tech. While I would prefer DH10 I can totally understand that DK sees the rebalancing problems coming with it and points out that they already have to do rebalancing with the current model or rebalance the current model. The way I read out your science.png adding a tenth worker adds 10 mins per minute compared to 55-57 before. The eleventh is way better but still over 20% inferior to workers before the ninth. So expanding once you hit 9 workers gives you massive advantages and things like zerg double expand creates economies that have 30-50% more economy than one base on the same worker count, making expanding pay itself after less than two minutes even if you stay on the same worker count. That is way shorter than LotV's half mineral patch system, essentially taking options like tech in the current balancing away. For Zerg DH10 means that if you have 24 workers going from two to three bases is better in terms of economy than using those mins to build more workers. For T and P it's pretty close. Totally disregarding the fact that mains have other advantages like being the main production building for Zerg and having chrono, scan and mule for Terra and that you can pump worker even faster, which accelerates the advantage even more. So yeah DK's concern that this overbuffs expanding vs teching still is totally valid. Have you tried to tune down the numbers a bit? Like what happens if the worker collects overall 8 minerals instead of 10? Just something that could make the jump to DH10 less extreme for Blizzard. Just because your diminishing returns start on the 9th worker doesnt mean you dont make that 9th worker. Also in the science.png image each line is calculated per worker. (math in HotS time below) So getting the 9th worker on one mineral line means each worker now makes 55 mpM compared to the 58 before -- on average. The 9th worker adds almost 40 minerals per minute to your income. IF he had a free mineral node to go on you would get almost 60 minerals a minute more. But this doesn't invalidate the income the 9th onward workers make for you. It only encourages you to spread them to a new base as soon as possible IF you are able to take and hold a base. The key interaction here is that it makes new bases MORE vulnerable because there will be workers there right away. In the current SC2 econ you dont put workers on the new base other than fresh workers you JUST made to transfer. So while this spreading of workers will benefit zerg immediately with their ability to spread out quicker, it will also provide a new challenge to zergs because they will have more mineral lines open to harass. In HotS players might come to a zerg third with a small hellion reaper push and find 4 drones meanwhile 32 drones are in the main and nat mining peacefully with a wall in with queen to block hellions. In DH10 the zerg has the option to spread to the third line sooner but open themselves up to hellions as a result. So if terran can for a short while contain those drones from getting their personal mineral patches for more money the terran can keep up with both mules and a 3rd cc being built. Zerg on the other hand might cut some drones (again keeping even with terran for it) and try to get an army to gain map control so that the hellion push cant deny saturating the third as they drone up behind the aggression. These kinds of small shifts in thought as to where you put the workers, how fast you put them there, what kind of defense you have for them, because you want their income to be better, this is all subtle small interactions that create more decision making and more action on the map. I didnt point out the 9th, I pointed out how inefficient the tenth is and as a result going over 9. The ninth is still ok (60% of the workers before) and once you overcome the fact that the tenth is barely existent (17% of a new worker on a new expansion) because you have reached like 14 workers going up to 16 is probably a good idea. However going from nine to eleven is less useful than adding a single worker on another base. It is close to ideal to have 9 on every base and then max out one base to 16. So overall ideally in the early game you want to have 24 to 27 workers on 3 bases and add a 4th once you have saturated one of the bases up to 16 because the tenth worker on a base will need 5 minutes to start generating value while the first nine need less than one. Having 27 mineral workers on three bases gives you close to 238 minerals per minte (19%) more compared to two bases, paying for the base in less than two minutes even without producing workers. I doubt that we need to discuss that this model is a strong buff for the expanding and map controlling race and would need heavy rebalancing of at least two of the three races. As I said before I think that a constant fight for map control and multi-pronged harassment is more interesting than deathball doom pushes and all-ins, but I doubt that Blizzard is going to go so far as to redevelop their different racial balancing entirely. They might just make starcraft 3 But you don't want only 9 workers. Even if the extra guys after that are inefficient, you don't want to do that because your overall income is still very poor compared to a person who has the extra workers. If you cut at 9 your gas income is much higher than your mineral income and you have a hard time affording anything properly. The games between CatZ and Iaguz showed that stopping at 9 workers as Z (the xpanding race) vs T (the more workers race) meant that anytime before catz was able to secure 5/6 bases he was behind in army workers and income. The entire time. And Iaguz rode that to victory. In basically every game. That is pretty interesting. Are you sure that that isnt a result of Mule? Because in a scenario with the low workercount Mule is going to have an immense impact. As I calculated through earlier expansions pay for themselves in less than two minutes and often are preferable to building more workers. However if you stay at 27 workers early on Mule is bound to have a massive impact. I might be really overlooking something, so if I was wrong that's all the better.
It basically comes down to a few factors like having 16 workers before expanding allows you to instantly transfer 8 to your natural which is better. Additionally while you may only gain 30% more minerals/minute at 18 workers compared to 9 workers (~300minerals/minute), that's still money you would otherwise not have. What it means is that it takes longer to get a return on investment of the workers between 10-24 than 1-9.
|
Italy12246 Posts
That's something that needs to be evaluated, but it's unlikely. You have to remember that because the first 8 workers are more efficient and cause builds to develop quicker, sending out workers (proxy gate, proxy rax) or sacrificing their production (early pools) can hurt you a good deal. That definitely requres further testing.
|
On April 23 2015 03:11 Grumbels wrote:Show nested quote +On April 23 2015 01:44 [PkF] Wire wrote:
I very strongly disagree that 11/11, proxy gates or early pools didn't bring a lot to the game. Most memorable series included disgusting all-ins. Hell, the most memorable game of that legendary Squirtle vs Mvp series is probably the Atlantis Spaceship proxy rax... Would DH make cheese more powerful? Because your first workers mine more, you have an advantage vs players that keep building workers normally compared to the HotS model.
On the other hand, the defender has more resources to raise buildings, set up a wall, and fighting units... Things go both ways.
I don't think theorycrafting will answer this one. We need actual games to show if cheese is a (bigger) problem or not.
|
The confusing part of this analysis is that all the shown graphs compare DH10 to HotS as a function of available workers.
Who cares about HotS? We need to compare against LotV.
It would make more sense to compare DH10 to LotV by plotting mineral income as a function of time assuming constant worker production until base saturation.
This way you could show when exactly 1 base runs out of minerals in both models.
And you could see how much money is available at any given point in time.
It would also take into account differences in initial worker count, mining efficiency and available minerals patches.
|
On April 23 2015 02:11 BlackLilium wrote:Show nested quote +On April 22 2015 23:50 ZeromuS wrote:
The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post.
The issue is that the mining income isn't linear as the worker count goes up.
Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20. Maybe instead of words, throw a graph, with some arrows and red letters "this is good", "this is bad", "this is what we achieve"
and remove words like optimal, efficiency and saturation, because many people find them confusing. avoid mentioning the "24 mineral patch cap" because people think in terms of bases, not patches. avoid complex statements that link multiple concepts in one sentence. only use one graph/image maximum.
|
|
|
I think you misunderstood what David meant by "saturated" and then redefined it your own definition. Your definition made him, and the devs, really look as though they did not understand. In normal parlance, an SC2 base is considered "saturated" when you get to 16 mining workers -- not 24. Let's use that as our definition and evaluate.
In Davids example he is really saying that the hots player on 2 bases needs 32 workers but the 4 base player would need 64 and that is pretty difficult to reach. Given that fact, he really wouldn't get that big of an econ advantage.
In DH10 though a base saturates (again, using the same meaning of the word) at 8 workers. That is where you start losing efficiency. So a 4 base player WOULD have an econ advantage over the 2 base player. He only needs 32 workers to fully saturate all 4 bases compared to the 16 of the two base player.
This means, in DH10 the 2 base player would have 16 workers over 2 bases. Thats an income of about 900mm (450x2). The 4 base player would have 32 workers over 4 bases. Thats an income of about 1800mm (450x4). Double the income.
Of course, the same would be true for Hots, if you could have ever get to that many workers. The two base 32 and the four base 64. But the point is, you can't. At some point you have to cut rope and start making an army.
|
Aotearoa39261 Posts
On April 23 2015 03:21 Grumbels wrote:Show nested quote +On April 23 2015 02:11 BlackLilium wrote:On April 22 2015 23:50 ZeromuS wrote:
The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post.
The issue is that the mining income isn't linear as the worker count goes up.
Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20. Maybe instead of words, throw a graph, with some arrows and red letters "this is good", "this is bad", "this is what we achieve" and remove words like optimal, efficiency and saturation, because many people find them confusing. avoid mentioning the "24 mineral patch cap" because people think in terms of bases, not patches. avoid complex statements that link multiple concepts in one sentence. only use one graph/image maximum.
We need to be detailed, at least in our original exposition of the material, since there are plenty of people on TL who hold PhDs on the economy of SC2 (or think they do) and will tear you to shreds if your analysis isn't comprehensive. Can we make a tldr version for those not wanting to get into the details? Probably, but communicating that is difficult!
|
On April 23 2015 03:21 Grumbels wrote:Show nested quote +On April 23 2015 02:11 BlackLilium wrote:On April 22 2015 23:50 ZeromuS wrote:
The tl;dr is actually difficult to explain in a few words and I'm trying to work on it for my next post.
The issue is that the mining income isn't linear as the worker count goes up.
Its a straight line to 8, and slopes from 9 to 24, and this slope means each worker becomes less useful the closer you get to 20. Maybe instead of words, throw a graph, with some arrows and red letters "this is good", "this is bad", "this is what we achieve" and remove words like optimal, efficiency and saturation, because many people find them confusing. avoid mentioning the "24 mineral patch cap" because people think in terms of bases, not patches. avoid complex statements that link multiple concepts in one sentence. only use one graph/image maximum.
I think not putting any words or images or graphs at all is better ; that way people won't have to use their brains to understand things and they will be happy
|
I think one thing being missed here is the value of stopping worker production. Your examples assume constant worker production from both players, but does't account for strategic cutting of workers. Typically on a two base economy a player will stop producing workers in order to produce the maximum amount of units for some sort of timing usually at the point where return on worker cost between starting it and the timing is less than a significant increase on army at the start of that battle..
I haven't seen much in the analysis mentioned on gas so assume everything below is pure mineral income
So the 2 base player will be at (LOTV numbers as half the patches become depleted)
HOTS: 32 workers ~1300 income
DH10: 16 workers ~950 income
LOTV: 32 workers ~1300 > ~1250 > ~900
Compare that to a 4 base player
HOTS: 48 workers ~1950 income
DH10: 32 workers ~1750 income
LOTV: 48 workers ~1950 > ~1950 > ~1950
This is the income disparity in those cases
HOTS: ~650 or 1.5x
DH10: ~800 or 1.8x
LOTV: ~650 1.5x > ~700 1.6x > ~1050 2.2x
Currently LOTV is a mix between the two as far as an income disparity, but rapidly increases to be more impactful as the natural base goes down to 8 patches.
The other question I have relates to income rates before saturation. I'm not sure how to measure it for an in game senario, but it will be easier to saturate new expansions meaning those income disparities will be hit quickly as opposed to taking more time to ramp up.
My other concern is that the lower worker saturation makes worker disparities from harass more extreme. There are questions of harassment units only killing a few workers and it just being too crippling of an economic difference. Taking out 4 workers is suddenly half a base of income as opposed to 1/4th. Of course these are questions that would need to be solved through play testing.
I'm just mainly concerned that this model makes things an extreme instant impact as opposed to having a gradual impact. You feel the economic increase more instantly from easier worker saturation and you feel the decrease as more extreme from worker deaths.
|
Again, unless you are a genius, it is close to impossible to compare the DH10 model to the LotV model with the shown graphs.
Please consider plotting earned minerals as a function of time for both DH10 and LotV for one base until saturation.
This takes into account the differences in worker count, efficiency and mineral halfings.
|
Great article, I can't wait to see more data come together. I have a suggestion/small critique, both articles don't really have a clear, concise thesis that definitively outlines the goals and expectations of the system. There are well written conclusions and plenty of useful data, but I personally had to spend a lot of time piecing through the information. As Plexa mentioned, it is entirely possible that someone other than DK read the article and possibly misconstrued the arguments made by TL writers. I think in the future it would be wise to have a nice, pretty thesis near the top of the article with html5 implemented to provide easy links to the meat of the article.
Again though, I think this is a brilliant effort and I can't wait to test this out with friends. Thanks again for the passion to do this!
|
On April 22 2015 15:05 Plasmid wrote:
Using the classic sc2 harvest model of 5 per trip, could the standard patch of 8 minerals be reshaped so that the closer ones are mined perfectly with 2 harvesters, then some at 2.5 then some at 3 and then some at 3.5 for example?
That would reward base expanding strats in the short run by putting workers on the closer patches across bases earlier, being a short term incentive to expansion.
Then the closer patches could have less minerals to start with (like the current LotV), which would provide the long term incentive to expand because the low minerals patches would vanish relatively ast.
(I am not remotely a good player or a balance specialist, so maybe this is terrible.)
This was investigated at some length during WoL beta and early WoL. It's a great solution except for one fault, which is that the workers don't automatically pair up on closer patches, they still go to open patches even if they are extremely inefficient due to extra distance. This necessitates a significant amount of player attention to properly pair up workers on the close patches at a new base. And imagine if you pull workers due to harass -- you have to set up the pairing when you go back to mine each time.
FYI tidbit: a patch 5 squares away is almost perfectly at 3 worker saturation with linear buildup, i.e. 2 workers' mine rate is 67% of 3 workers'. And consequently with a standard "close" patch (3 squares) being >90% saturated with 2 workers, a 5squares-away patch with 2 workers gives ~70% the income.
One of the reasons why "worker bouncing" is better is because it delivers the desired economy dynamics using just the default worker behavior once the player has distributed them roughly evenly between bases (a relatively simple task).
|
|
|
|
|
|
EPT
