EPTIs balance an impossible goal? - Page 3
| Forum Index > SC2 General |
|
Shifft
Canada1085 Posts
| ||
|
Kazam
Australia21 Posts
| ||
|
wherebugsgo
Japan10647 Posts
When I consider balance, I consider balance as it pertains to high level players, and how they perform against others. The way the rating system in SC2 works right now gets you to 50% overall winrate, in general. Most players actually have a higher than 50% overall winrate, other than those in bronze. No one could dispute the fact that reapers in TvZ made TvZ very imbalanced a few months ago, before the barracks after supply and reaper speed upgrade change. HOWEVER, most Z players would have winrates against T that were similar to their vP or overall winrates even in that patch, simply because of the way the rating system works. Also, for high level players, specialization matters less than for players of lower skill. Players of lower skill can be good in one matchup but suck hard at another, and it's never consistent across the vast pool of players. In GSL 1 we saw, though, that the only exception to the hardship of Zergs was Fruitdealer, and even he had to be very lucky to actually do what he did. Personally I think that balance is something that CAN be quantified, it just can't be quantified using the ladder system or most players' concerns. I think players should work on improving their game before actually claiming that something is imbalanced. | ||
|
whatthefat
United States918 Posts
1) There is no way of determining whether the players of one race are on average more or less skilled than the players of another race. In other words, if say all Z players switched to T, all T players switched to P, and all P players switched to Z, there is no reliable way to predict how things would shake out. 2) What is balanced at masters level may be significantly imbalanced at bronze level, and vice versa. Truly, I think Blizzard has done an incredible job balancing the game, and keeping it balanced at a wide variety of skill levels. | ||
|
AlphaIIOmega
Canada29 Posts
But I disagree that it's an impossible task to balance. Firstly, you need consistency in the map pool. Blizzard's maps are wildly bad. It's ok to have differently themed maps (i.e. more cliffs, more open spaces, etc.), but it's not OK to have these TINY TINY rush distances that some maps have. Bringing more consistency to the OPENING game will help a lot. Strategies like 2 rax/SCV pressure are ruining the TvZ data, because you can't deduce core long-term imbalances. As a reference, even Xel'Naga caverns is too small of a rush distance. You need maps minimally with the rush distance of Scrap Station (and preferably without the destructible rocks). Second, once you have balanced maps you must make a sample pool. Master's league is an excellent start point. Each race will be represented by MANY players within master's league. You also have to assume that the average player skill of each race is equal. Lastly, you can compare global player % to Master's league player %. If MMR is hiding imbalances, Master's league will show that. If 40% of players play race X globally on the NA server, but 50% of Master's league is race X, then you have yourself an imbalance. | ||
|
Shron
United States162 Posts
I know you mentioned maps, but the map pool truly is the only way to achieve "perfect" balance. | ||
|
palanq
United States761 Posts
1/3rd chance to make the game balanced 1/3rd chance to make Z underpowered in all matchups 1/3rd chance to make T overpowered in all matchups so your 1st disaster scenario can be averted and reduced to the 2nd disaster scenario | ||
|
Simberto
Germany11749 Posts
On January 20 2011 13:45 wherebugsgo wrote: I think the original post is flawed in that it doesn't actually take into account skill. That is the main point. How do you measure actual player skill? Obviously, it would be incredibly easy to find imbalances in matchups if you could give every player an exakt skill number, and then say "Ok, they have the equal skill number, so they should play 50/50. If it is not after a statistically significant number of games, there is an imbalance." Or "That guy has a much higher skill number, so they should play 70/30" However, the only way to determine the skill of a player is how much he wins. Which makes the whole logic circular, since if you base that skill number on win percentages, and then the expected win percentages on that skill number, you obviously get good results. One of the main points was that since you can not directly measure the player skill, stuff has impacts on other stuff, which could delude your conclusions regarding the skill of someone, or a matchup imbalance. When I consider balance, I consider balance as it pertains to high level players, and how they perform against others. The way the rating system in SC2 works right now gets you to 50% overall winrate, in general. Most players actually have a higher than 50% overall winrate, other than those in bronze. No one could dispute the fact that reapers in TvZ made TvZ very imbalanced a few months ago, before the barracks after supply and reaper speed upgrade change. HOWEVER, most Z players would have winrates against T that were similar to their vP or overall winrates even in that patch, simply because of the way the rating system works. Exactly hits the point. If there were a 10% handicap for one race in one matchup, this would just result in you getting respectively worse/better enemies of that race, and alter through the MMS the relative skill of matched players of all matchups. Which, in conclusion again means that it is hard to draw data of a potential imbalance out of any ladder stats. Personally I think that balance is something that CAN be quantified, it just can't be quantified using the ladder system or most players' concerns. I think players should work on improving their game before actually claiming that something is imbalanced. Then, how would you quantify it? I, too, am of the opinion that there probably is some way to do so, but it seems to be neither easy nor obvious, or at least i can not see it right now. | ||
|
oxxo
988 Posts
The problem right now is the maps. So far they've been balancing based on Steppes/Blistering type maps (Reaper nerfs, Zealot nerfs, depot before rax, etc.). Hopefully they move to bigger, non 2 player maps and balance around that. I do think balance is impossible on the small Bloodbath maps. | ||
|
ckunkel1
United States181 Posts
| ||
|
Rushme
Canada10 Posts
| ||
|
Samp
Canada783 Posts
| ||
|
usethis2
2164 Posts
I think anyone with decent knowledge with the game can get an idea of balance by looking at a game, and what's possible and what's not. If some of the strategies look relatively easy yet require overwhelming response from the opponent, then I know the strats are OP. | ||
|
wherebugsgo
Japan10647 Posts
On January 20 2011 13:59 Simberto wrote: That is the main point. How do you measure actual player skill? Obviously, it would be incredibly easy to find imbalances in matchups if you could give every player an exakt skill number, and then say "Ok, they have the equal skill number, so they should play 50/50. If it is not after a statistically significant number of games, there is an imbalance." Or "That guy has a much higher skill number, so they should play 70/30" However, the only way to determine the skill of a player is how much he wins. Which makes the whole logic circular, since if you base that skill number on win percentages, and then the expected win percentages on that skill number, you obviously get good results. One of the main points was that since you can not directly measure the player skill, stuff has impacts on other stuff, which could delude your conclusions regarding the skill of someone, or a matchup imbalance. You can't, generally. This is one reason why balance concerns should never be based around ladder. However, you CAN identify those who are very good at this game, and those who suck. This is why balance is not based around what happens in Bronze, or we'd see a very weird game indeed. You can, however, identify flaws in the game by looking at how professional players fare with their matchups. If, clearly, the majority of upper level players of one race are having difficulty with a certain matchup, something is wrong. This was obvious with ZvT before the big changes that happened a few months ago. On January 20 2011 13:59 Simberto wrote: Exactly hits the point. If there were a 10% handicap for one race in one matchup, this would just result in you getting respectively worse/better enemies of that race, and alter through the MMS the relative skill of matched players of all matchups. Which, in conclusion again means that it is hard to draw data of a potential imbalance out of any ladder stats. Yep, you can't use the ladder to balance. On January 20 2011 13:59 Simberto wrote: Then, how would you quantify it? I, too, am of the opinion that there probably is some way to do so, but it seems to be neither easy nor obvious, or at least i can not see it right now. Again, watch trends at high levels of play. Was it coincidence that every professional player prior to patch 1.1.2 had trouble with ZvT? Was it coincidence that it was almost always because of reapers? Clearly, Terran players had no problem using builds like 1/1/1 against Zerg (like TLO's hellion expand, Thor drops on Kulas and LT, banshee builds) but it seemed like reapers were just so imbalanced. There was an obvious early-game imbalance at the time. Now, right now, there's nothing really to suggest this. I don't think there is significant information yet, and right now there are no statistics that really back one argument up more than any other in regards to balance. It'll be a few months as the game matures before there are any issues, I think. And from here on out, I think the balance issues will be much smaller than the old ZvT issues. The only thing that could make the problems worse is a patch that changes too much. | ||
|
Angra
United States2652 Posts
On January 20 2011 12:50 Redmark wrote: Is that why every matchup has been perceived as imbalanced in both ways, shifting left and right pretty much every week? Is that why when they give Roaches a +1 range people flip the fuck out? Please. The game will not be balanced by armchair designers playing God. The game will be balanced incrementally, more by the community than by Blizzard. Do you honestly believe that adding +10 health here, taking away -5 damage there is going to eventually make the game have absolutely perfect balance while at the same time keeping it a dynamic, interesting, competitive game? | ||
|
Wonderballs
Canada253 Posts
1/3 = 0.33333333333333333333333333333... Technically the post could be longer. | ||
|
Paradice
New Zealand431 Posts
On January 20 2011 13:17 Whole wrote: To quote IdrA: + Show Spoiler + if statistics get you hard make one of those ladder analysis pages or something, but stop interfering with balance discussions. Ladder should never be mentioned in balance discussion. IdrA should never be mentioned in balance discussion. Great post OP! As a couple of other s have mentioned, Blizzard do have an 'adjustment' mechanism which they can apply to their match statistics, but it removes the influence of the *matchmaker*, not the effects of misrepresentative MMR defined by OP. There's little point in tracking distinct MMRs for each matchup too, as it still suffers the same effects. I agree there doesn't seem to be any mechanism by which Blizzard can use their aggregated statistics to prove or disprove balance, and that there's definitely no way individual players can do so. At best it can give hints. Resolving it statistically requires a known quantity to test each race against - the hypothetical "perfect play in all situations", but that's not currently feasible. What we can hope for in reality is for the game to be "balanced enough" - which I would define as there being no strategy that any one race can use that will win more than 50% of the time against another player of approx. equal skill that is prepared for it. The problem with this definition is that it can only be proven by counterexample or by waiting 10 years to see if anything emerges. To Blizzard's credit, some blatant counterexamples (TvZ mass reaper anyone?) have been identified, and have been removed. No doubt we'll find more. As long as the game ends up "balanced enough", the players can and should use their own innovation to resolve everything else. Just like they did with Brood War. | ||
|
Shaok
297 Posts
| ||
|
wherebugsgo
Japan10647 Posts
On January 20 2011 14:34 Angra wrote: Do you honestly believe that adding +10 health here, taking away -5 damage there is going to eventually make the game have absolutely perfect balance while at the same time keeping it a dynamic, interesting, competitive game? No, he's advocating that Blizzard do nothing for a few months, so we can examine what happens to the game as it evolves untouched for a while. The most recent patch was merely a week ago, we don't need to change anything right now. I might be in support of additional maps, but that's not a gameplay change as much as it is necessary just to phase out maps that aren't being used or being considered seriously right now, such as Steppes, Delta, Jungle Basin, and Blistering Sands. In the end, I have to agree with what he said. The game does NOT need core changes right now. It doesn't need changes to the statistics of units, either, but stat changes are more easy to predict than radical changes in how the game works on a fundamental level. | ||
|
Angra
United States2652 Posts
On January 20 2011 14:43 wherebugsgo wrote: No, he's advocating that Blizzard do nothing for a few months, so we can examine what happens to the game as it evolves untouched for a while. The most recent patch was merely a week ago, we don't need to change anything right now. I might be in support of additional maps, but that's not a gameplay change as much as it is necessary just to phase out maps that aren't being used or being considered seriously right now, such as Steppes, Delta, Jungle Basin, and Blistering Sands. In the end, I have to agree with what he said. The game does NOT need core changes right now. It doesn't need changes to the statistics of units, either, but stat changes are more easy to predict than radical changes in how the game works on a fundamental level. I'm not saying to all of a sudden rush out core changes to the game, sorry if I implied that. What I mean is, eventually down the road, all of these number changes most likely aren't really going to do anything to make the game better as a whole. It just seems like a temporary fix to adjust problems that arise, but there's probably going to always be those problems no matter how many times you tweak the numbers on units and stuff because the actual problem lies deeper than stuff like "should marauders do 18, or 20 damage?" | ||
| ||
