| Forum Index > SC2 General |
|
Offhand
United States1869 Posts
Korean numbers show that the game is ridiculously volatile there. One bad GSL for protoss has them at the historic balance low for the past 6 months, which is kind of absurd given everything leading up to that. Obviously the lack of events is to blame for the small data size. No race really has any ground to complain about balance at the moment. Yes, even zergs. | ||
|
Erasme
Bahamas15899 Posts
On May 03 2011 02:03 Alejandrisha wrote: Zergs could have 80% winrates in every matchup (including zvz) and they'd still find something to whine about. It's just the way the game works. Have you ever beaten a zerg who actually thought they deserved to lose? It is just the way Zerg play has developed; apparently being beaten into the ground by 5rax reaper for so long didn't sit well with them and they've grouped any loss they've suffered with a similar abuse. It is sad, really. A lot of zergs I know are losing because of their mindset rather than their play; they are beating themselves. That's kinda stupid. You can see that a reaper nerf was needed. That TvZ was (or is) imbalanced in favor of terran. And you still say 'whiny whiny zergs'. That's beyond my mind. Stop bashing on zerg. | ||
|
Silkath
United Kingdom102 Posts
PS That's sqrt(no. of games)/(no. of games) if you're not familiar with Poisson statistics. | ||
|
Alejandrisha
United States6565 Posts
On May 03 2011 02:20 Erasme wrote: That's kinda stupid. You can see that a reaper nerf was needed. That TvZ was (or is) imbalanced in favor of terran. And you still say 'whiny whiny zergs'. That's beyond my mind. Stop bashing on zerg. I said the reapers were abuse. They were imbalanced. I'm not saying they were not. "whiny whiny zergs" was not found in my posted. How is that beyond your mind. Even Nestea, the zerg hero says "Zerg TT." It looks like you look past every other sentence in my post and just harp on 2 words here, three words there. | ||
|
Ctuchik
Sweden91 Posts
On May 03 2011 01:39 Befree wrote: So are the data points you used to for the graph every month? So like 6 values for each race? And when you say you only tournaments/leagues, are you excluding all the KOTHs and qualifiers that are listed in the TLPD The Korean statistics I'm most critical of. If we take for example the April statistic, I assume you're using the GSL Code A/S and the SK Gaming Asia Master tournament as the sole source of data. Now when you also consider the GSL code A and S are not even finished and that a large portion of games are mirror match ups, I just can't see how we can find value in a sample like this. I think the huge variations we see are much more indicative of the volatility that comes from the tiny sample, rather than patches or "metagame" changes. Do you think you could give your numbers more specifically and calculate the error? I don't think you just noting a small sample size is enough (As shown by how seriously people are taking the korean data in this thread). I think otherwise this thread is abusing statistics and giving people bad information. Yeah, there is one data point per race per month. I can very easily make it per week (or day) if you are interested in seeing that. The data set IS the TLPD. I am loading everything in it. The only restrictions on the data set is the time period (6 months) and the exclusion of mirrors. I would probably agree that the data set is to small to make any real analyses on the Korean side, but it is pretty interesting anyway. I would not make make the conclusion that P is weak based on only one data point. | ||
Chill
Calgary25987 Posts
On May 02 2011 23:59 DNB wrote: This is super interesting, I especially like the graphs where the winrates switch places oppositely and then start to even out. It really tells you about the math behind the metagame... There is no math behind the metagame. That's like saying "the math behind anticipation". | ||
|
Alejandrisha
United States6565 Posts
On May 03 2011 02:28 Chill wrote: There is no math behind the metagame. That's like saying "the math behind anticipation". Whenever some one says meta game I just wait for it xD | ||
|
Kaonis
United States243 Posts
Ah, Chill isn't even reading this thread anymore, unless he has some special search function for the word "metagame" If metagame isn't an acceptable term, then "the widely accepted as standard set of strategies as well as possible cheeses all seem to be fairly volatile at this point in the game". Also, protoss not dominating? Am I reading this wrong? | ||
|
Ctuchik
Sweden91 Posts
On May 03 2011 02:20 Silkath wrote: Any chance of error bars on each data point? Just Poisson noice based on the number of games contributing to that point would be fine. Would be very useful to compare how reliable a given data point is. Alternatively, any indication of how many games contribute to each datapoint would be useful. If TvZ is up at 70% one month with only 10 games played then back down at 50% the next month with 100 games played the number of games is really important in interpreting the data. PS That's sqrt(no. of games)/(no. of games) if you're not familiar with Poisson statistics. I would love to do this, going to see if I can get this added! | ||
|
Thojorin
Germany162 Posts
On May 03 2011 01:43 Mafe wrote: Hm. Unless there is additional information about the data (number of matches per month, especially april), I am convinced this does not proove anything (if scientific criteria would apply). In terms of balance I mean. Exactly. It's a nice plot, giving food for thought, showing exactly what the original poster said it shows. Inferring and/or interpreting any interesting general statements from this plot, like "as of now, races are balanced", requires a lot of work (and is impossible with these plots alone). | ||
|
caradoc
Canada3022 Posts
On May 03 2011 02:28 Chill wrote: There is no math behind the metagame. That's like saying "the math behind anticipation". I disagree with this, I think there is math behind metagame-- To simplify-- you have a strategy that is perceived to be powerful, people flock to it. As more people flock to it, strategies which are strong vs. said strategy gain in popularity, people flock to said counter. As more people flock to said counter, initial strategy becomes less popular. The curves representing the number of people using both of those strategies approach bell curves. This is idealized and simplified, and the reality is much more complex of course, but its disingenuous to say there is no math. But I understand your hostility towards the term, it gets bandied around and used in a million different ways. | ||
|
Chill
Calgary25987 Posts
On May 03 2011 02:38 Kaonis wrote: So.. The metagame seems to be unstable. Ah, Chill isn't even reading this thread anymore, unless he has some special search function for the word "metagame" If metagame isn't an acceptable term, then "the widely accepted as standard set of strategies as well as possible cheeses all seem to be fairly volatile at this point in the game". Also, protoss not dominating? Am I reading this wrong? It's not special, I just used the search function. How about "Protoss strategies are dominant"? There's no need to try to overcomplicate it. Using the wrong word is more convenient but it's still wrong. Compare the following: 1. The shuttle failed when the fuel inside initial booster rocket #6 exploded. 2. The shuttle failed when the table exploded. #2 is simpler but wrong. | ||
|
Chill
Calgary25987 Posts
On May 03 2011 02:42 caradoc wrote: I disagree with this, I think there is math behind metagame-- To simplify-- you have a strategy that is perceived to be powerful, people flock to it. As more people flock to it, strategies which are strong vs. said strategy gain in popularity, people flock to said counter. As more people flock to said counter, initial strategy becomes less popular. The curves representing the number of people using both of those strategies approach bell curves. This is idealized and simplified, and the reality is much more complex of course, but its disingenuous to say there is no math. But I understand your hostility towards the term, it gets bandied around and used in a million different ways. Very true. My comment was a defensive exaggeration. | ||
|
Thrombozyt
Germany1269 Posts
On May 03 2011 01:45 Saechiis wrote: You would quote anything that says Protoss is hard to play, it's getting tiring to read your replies to every thread that has "balance" in it. We get it: Terran OP, Protoss UP, sympathy for Zerg to get them to agree and ending on the note that you're not implying imbalance. Thank god I'm not the only one that noticed.. | ||
|
xbankx
703 Posts
Also, people got to stop thinking higher apm=playing better. I just hate it whenever I hear someone say "I played better but I still lost". No, you lost because you played worse than your opponent saying the prior is merely trying to justify your loss through balance/imbalance discussion. Some player in a few post said a lot of zerg whiners. It is mainly because pro zerg like Idra, and artosis (before he switched) tend to whine a lot more and the other zerg just follow. Protoss players have pros like Tyler and Incontrol. When protoss were dying left and right to terran/zerg in GSL1,2, and 3 they never said imba. They just said the protoss were playing bad. Lets all remember the show Imbalance with Idra/artosis that said supposedly the "2base void ray colossus zealot build" was supposed to be unstoppable. How many PvZ pro games have people seen where toss went that. | ||
|
Jayrod
1820 Posts
On May 03 2011 01:45 Saechiis wrote: You would quote anything that says Protoss is hard to play, it's getting tiring to read your replies to every thread that has "balance" in it. We get it: Terran OP, Protoss UP, sympathy for Zerg to get them to agree and ending on the note that you're not implying imbalance. Protoss IS hard to play though and that is going to become very clear with the next patch when all the nubs 4-gating start fucking up the TL strategy forums trying to figure out why they can't win anymore. | ||
|
entropius
United States1046 Posts
On May 03 2011 02:20 Silkath wrote: Any chance of error bars on each data point? Just Poisson noice based on the number of games contributing to that point would be fine. Would be very useful to compare how reliable a given data point is. Alternatively, any indication of how many games contribute to each datapoint would be useful. If TvZ is up at 70% one month with only 10 games played then back down at 50% the next month with 100 games played the number of games is really important in interpreting the data. PS That's sqrt(no. of games)/(no. of games) if you're not familiar with Poisson statistics. I was just about to post the exact same thing. I tell my students the first week of lab: "No data is useful without an error bar." | ||
|
Jermstuddog
United States2231 Posts
On May 03 2011 02:51 xbankx wrote: Personally, people should just stop complaining about imbalance and just play to improve. Let the designer of the game do the game tweeking. Also, people got to stop thinking higher apm=playing better. I just hate it whenever I hear someone say "I played better but I still lost". No, you lost because you played worse than your opponent saying the prior is merely trying to justify your loss through balance/imbalance discussion. Some player in a few post said a lot of zerg whiners. It is mainly because pro zerg like Idra, and artosis (before he switched) tend to whine a lot more and the other zerg just follow. Protoss players have pros like Tyler and Incontrol. When protoss were dying left and right to terran/zerg in GSL1,2, and 3 they never said imba. They just said the protoss were playing bad. Lets all remember the show Imbalance with Idra/artosis that said supposedly the "2base void ray colossus zealot build" was supposed to be unstoppable. How many PvZ pro games have people seen where toss went that. So many issues with the logic in this post... I'll start with your comparison of pro Zergs vs pro Protoss. Zerg has been buffed repeatedly since release and is only now coming to equal win ratios vs other races. Protoss has been nerfed and now they're doing better than ever. When Tyler and Incontrol said Protoss were just playing bad, that's true. When IdrA, Artosis, and every korean pro said Zerg was less capable, that was also true. Then you move on to talk about 2 base Void Ray/Colossus implying that it isn't imbalanced when the infestor was specifically buffed by Blizzard to deal with this unit composition and even AFTER that, Zergs are STILL having trouble with it. When they made that episode so many months back, it definitely WAS imbalanced and had to be addressed by the creators of the game... *sigh* why bother... | ||
|
joyeaux
United States169 Posts
On May 03 2011 01:49 Cloak wrote: Sample size is smaller for Korea but deviation from average is ridiculously huge. If you do a chi squared, I'm sure it would reject the null at 95% confidence, probably even 99% confidence. 700 games is pretty damn sizeable. PvT and PvZ are looking like absolute garbage right now. That would be relevant if we were 1. looking at a random sample of a population and 2. drawing conclusions about the whole population. These statistics measure a non-random section of the population (results of specific tourneys and league) and from those statistics you aren't trying to draw conclusions about the whole population, but about the ephemeral subject of "game balance." TL;DR statistics do not work that way. | ||
|
whatthefat
United States918 Posts
Perhaps the raw number of players of each race who qualify for tournaments, or make it to a particular round would be a more reliable statistic for assessing balance. But ultimately the term balance isn't very well defined. Would we consider the game balanced if the 3 races were objectively equal when played close to perfection, but one race won 80% of the time between equally skilled players at all lower levels? Or what if each player has a particular race that stylistically suits them best, but one race is only the best fit for 5% of the playing population. Since more than 5% of the population plays that race, they may appear to perform worse than they should. Would that be defined as imbalance? | ||
| ||
|
|
OSC
Replay Cast
OSC
LAN Event
Korean StarCraft League
CranKy Ducklings
LAN Event
IPSL
dxtr13 vs OldBoy
Napoleon vs Doodle
BSL 21
Gosudark vs Kyrie
Gypsy vs Sterling
UltrA vs Radley
Dandy vs Ptak
Replay Cast
[ Show More ] |
|
|
EPT
