|
On May 03 2013 23:26 Sissors wrote: Besides that it lacks any kind of statistical significance (seriously people), WCS Korea Challenger shows we should nerf zerg. It is quite easy to cherrypick stuff until it shows what you want to show. ?
You said that if you took only the korean data, the results are 50/50. That is factually wrong. It has nothing to do with cherrypicking, nor statistical significance, nor nerfing anything.
|
|
I see. I just used the spreadsheet from the original post. The document you link and that one seem inconsistent in some places so who knows what the real statistics are... not interested enough to add it up all myself :p
|
It's not that confusing... All of the patch zergs are finally being removed from the game so of course there's going to be a huge drop in zerg play because the infestor BL combo is no longer the worst thing in SC2 to spectate as it would win games no one deserved won.
Also someone mentioned this earlier that Korean Terran's dominating foreign zergs means nothing on balance since foreign zergs are generally rather bad at the game (with exceptions of course in a few players)
|
I have just about a 100% winrate against zergs that haven't adopted a new strategy, but I haven't won a game against someone that goes roach hydra, so I think they just need to adapt a bit.
|
On May 03 2013 23:56 9-BiT wrote: I have just about a 100% winrate against zergs that haven't adopted a new strategy, but I haven't won a game against someone that goes roach hydra, so I think they just need to adapt a bit.
DISCLAIMER: This post means nothing regarding balance, he just made up a random percentage and placed it as so.
Link that 100% winrate please and why this would matter to a balance discussion : D
|
I think TvZ is the match up which needs fixing. When I look at TvZ match ups, sometimes I think what would have zerg have done to win? And then it just seems hopeless. The widow mine bio composition behind a solid terran macro is just so deadly and cost effective.
|
Catch 22:
If you use only the data from the top players, the "sample size isn't big enough." If you use data from all sorts of players, the "skill gaps are too large."
Will statistics and Starcraft ever pull an archon morph?
|
On May 04 2013 00:50 The_Pacifist wrote: Catch 22:
If you use only the data from the top players, the "sample size isn't big enough." If you use data from all sorts of players, the "skill gaps are too large."
Will statistics and Starcraft ever pull an archon morph?
Nah, problem is statistics are only used by people to back up their current beliefs on balance. If it supports your view, you embrace it; if not... well yeah you'll find any way to discredit it.
|
On May 03 2013 22:05 tomatriedes wrote:Show nested quote +On May 03 2013 21:37 Qwerty85 wrote:On May 03 2013 21:33 tomatriedes wrote:On May 03 2013 21:29 Qwerty85 wrote: It is a different game, zerg micro requirements are raised compared to TvZ in WoL. This will no doubt cause problems to many zergs that were considered very good before. I think TvZ is now much closer in terms of skill required on both parts and only the best of each race will do good.
Just look Innovation vs. Life. I dont remember watching such good TvZ in a long time. Just hope Blizzard doesn't interfere too soon and make it a "no rush 20min into deathball vs. deathball" matchup again.
As far as TvP goes, protoss seems a bit too strong in my opinion. Their early game is considerably stronger than in WoL and their lategame is still as good as ever before, if not better. TvP matchup is almost perfectly balanced at the moment according to these figures and you think that's a bad thing? What do you want? 60-40 terran favored, 80-20? What's going to make you happy? I am talking about my impressions about the matchup. It is still too early in my opinion to look too much into those stats. And you need to chill out a bit dude... There is frustration and anger all over your post. I dont see a reason for it. This is a thread about win rates. If you're not interested in discussing that you can post here instead: http://www.teamliquid.net/forum/viewmessage.php?topic_id=255254And I do get annoyed with the amount of whining I see from terran players on this web site, when at the moment they're the most winningest race at tournament level. edit- posted wrong link
At the moment, only whining comes from you. I was merely saying my opinion...
And Wol TvZ was zerg favored for months and months before Blizzard even admitted a problem with the infestor and tried to come up with a solution. And that was in a game that was already out for 2 years with basically all builds and timings figured out.
All I am saying is, these statistics are hardly reliable at the moment because game is so new that a lot of problems simply comes from players not being able to adapt to new circumstances.
If Blizzard took the sit and wait approach for the end of WoL, they should also wait a bit before making any bigger changes in Hots.
Right now, we can see much more about balance by actually watching the games. I suggest you watch/rewatch this season of GSL and see how TvP's played out.
Protoss has tons of options in the early game now, and it is very hard for terran to scout which build protoss will go for. At the same time, terran early game is much weaker due to the canon, msc etc..
Midgame is still good for terran and medivacs are certainly not easy to deal with, but late game again forces terran into same old high dps but low health bio that needs superb micro to withstand all protoss aoe in the late game...
Overal, yes I think it is slightly protoss favored even if it does not show on stats right now.
|
1561 Posts
On May 02 2013 02:54 BronzeKnee wrote:Show nested quote +On May 02 2013 02:52 Alryk wrote: 52/48 is almost definitely within margin of error, although we don't know exactly what the margin of error is.
Nice job of saying absolutely nothing =) If the sample size is large enough, then 52/48 could be a significant difference. Yes, it's not within margin of error of statistics. There is 2% difference. But fluctuations in this size are expected due to changing metagame and similar stuff. It doesn't really mean anything unless the matchup will stay the same for years at 52/48.
|
Data source?
If you want anyone educated to take it seriously, post a link to the data file.
|
On May 03 2013 15:22 Sissors wrote:Show nested quote +On May 03 2013 07:25 Jinky wrote:On May 03 2013 00:30 GTPGlitch wrote: @Jinky: That's a really dumb argument, because balance affects players no matter how bad they are. If i'm in bronze, and there's another guy in bronze, and protoss has a 99% winrate vs terran and he's playing protoss, it's very likely he beats me because we're at the same skill level and balance favors the protoss. Sure, I can just get better and win, but throwing the casual players under the bus isn't exactly the way to go about making star2 super popular First of all, there is no 99% win rate in SC2 now or ever in the past. That is an asinine way to argue. But even if there was a 99% win rate, if that win rate includes weaker players then it doesn't matter too much (as far as balance goes). The only win rates that matter are those with players who are so good that they push the limits of what the races can do (ie. top Koreans). Why is this? Because weaker players still aren't hitting their injects or setting up depot wall-offs or placing good force fields, and so they lose because they beat themselves, because they don't use their tools properly. It's only when a race is using all of its tools and still can't win that balance should be talked about. Sorry but thats besides horrible for blizzard sales also just not how it works. If at bronze level an average player who shifts from race X to race Y suddenly alot better, than race Y is too strong at bronze level. Now because of what you describe you can accept higher level of imbalance at bronze level than at code S level. However that doesn't mean it is irrelevant what the balance is at those levels. They are also paying customers, and if they have to play a horribly imbalanced game at their level they soon won't be paying customers anymore (not to mention they also simply deserve a balanced game for them). But to follow your logic, why should we then use the completely arbitrary line of top Koreans for balancing? Why not actually use a situation where those races can actually use all their tools instead of being limitted by our puny humans? So get the top players with best decission making, and let them play via something like automaton 2000 bot. Then the races truly aren't limitted anymore and used properly. And I now know already you aren't in favor of it, apparently you also agree that it should be balanced for regular use by Korean pro's. Then why shouldn't it be balanced for regular use of other players?
Imbalance at different levels? Haha, you can't be serious! Please explain to me how you can balance a game around a lower level of skill. There are hundreds of specific and varying reasons why someone isn't skilled at the game, and there might be hundreds of different reasons why another person of "equally low skill" also isn't skilled at the game. How can you balance for the plethora of people's widely various faults? If it's not a balance problem for the very best players, then it shouldn't be a problem for lesser-skilled players also because the tools are there for them to deal with the problem; they just need to improve their skill.
"Paying customers" does not factor in to pure balance. Regardless of how well the developer does with sales, balance is a separate issue. If people don't like the game, then that is just a design problem, not a balance problem.
I love your suggestion to use computer AI (ie. "automaton 2000 bot") to test balance vs top players. This would certainly be ideal... if it were possible. It is impossible because the game is so so so complex that no current computer AI can handle it as good as even a half-decent player can. (Seriously, the "Elite" AI in HotS is a joke.) The best that SC2 AI can do right now is, in general, follow a build order and A-move at certain timings. Maybe if someone invested a lot of time and money into something like IBM's Watson AI (which is able to learn and adapt) then this could be possible.
But SC2 is so complex, with so many variables that cannot be disregarded. Like I said in a previous post:
On May 03 2013 07:25 Jinky wrote: The game is way too complex (the possibilities of unit compositions, timings, army positioning, economic relationships, resource relationships, scouting knowledge or lack thereof, hidden costs, unit value per supply, unit per-hit cost from each specific enemy unit, etc. etc. etc.) for "balance discussions" to revolve around specific units/strategies that seem OP even though there are many other things that can be done to counter these (things that the top Koreans are probably doing, because they are winning versus those units/strategies).
So again, statistics threads like this one are irrelevant unless the data is about the very very best players in the world.
|
|
On May 04 2013 01:25 Protoss-Bah wrote: Data source?
If you want anyone educated to take it seriously, post a link to the data file.
It's located in the middle of the thread, page 5 or something, you might want to scan the thread before posting... you know just to be educated.
|
On May 03 2013 23:47 Sissors wrote:I referred to spreadsheet I linked earlier (someone elses, not mine): https://docs.google.com/file/d/0B76Yjrn8DAmfeWdpY0F0b1ByOEU/edit?pli=1. I don't follow all Korean stuff so if it is missing important ones I was indeed mistaken. In that spreadsheet it was pretty much 50/50 for TvZ. But as I said, if it is missing stuff I stand corrected.
GSTL is missing. The title is "Korea only" so WCS Challenger outside Korea shouldn't be counted. It is very misleading.
This set of data includes matches from March, not just from April. If you want to see win rates since HotS released I posted a more detailed collection categorized by tournaments here.
|
|
On May 05 2013 12:40 Emzeeshady wrote:Your missing GSTL. Terran is still dominating Zerg in the three highest level tournaments. WCS Korea, Proleague and GSTL. This isn't me commentating on balance, but a problem in my eyes with different races and looking at certain tournaments is that terran is such a good race in these tournaments where you can practice for the games, since in both MUs terran is the playmaker unless terran gets allin'd. That just gives terran such an advantage in the prep work. That's why I feel zerg especially(which is for the most part reactionary race) has it a lot easier in live tournaments and a lot harder in... scheduled tournaments. I'm not trying to say terran is balanced, I have no idea quite frankly, just food for thought.
|
|
Overall, I think tourney results make it very hard to tell unless they are consistent and ridiculously favoring one race over the other.
|
|
|
|