Welcome to ZParcraft II - Page 56

On July 02 2014 02:30 Lunareste wrote:


Then go to the bnet forums and post suggestions. They asked for our feedback regarding this specifically.

On July 01 2014 07:19 Hider wrote:


It's all about how you interpret statistics though. You cannot look at naked win/rates, and that is unfortunately what David Kim is doing. The reason for this is that when one race is underpowered, the players of the race will drop down in leagues, which causes fewer players of the UP race in master/GM. Win/rates, however stay close to 50/50, because that's how the ladder system is designed. Thus, the only relevant metric for assessing ladder balance is looking at race distribution. According to Nios.Kr terran only has 24% players in GM and 28% in masters league, while they should have had 30% assuming the average terran player is equally good as the average zerg/protoss/random player.

Regarding Aliguac statistics

So if you go in and study Aliguac numbers you will notice that there only was a terran player in a non-mirror matchup in 57% of all games (in June so far). Compare that to the 70% of protoss/zerg, and the balance issue is quite significant.

Once against, be aware that win/rates significantly underestimates the real issue here.

Because Aliguac statistics only picks up the very best terran players against a higher number of protoss/zergs that doesn't neccsarily have the same skill.

As an example, if the 5th best terran player in the world is being faced up against the 50th best zerg/toss, the terran player should (If the game was balanced and average skill level of zerg/toss was = average skill of terran) have a w/r above 50%. However, looking at Aliguac statistics (for June at least), that is not the case.

So if you study the statistics in a way that makes sense, you cannot come to any other conclusion than terran being significiantly underpowered.

I do agree that it might be an indicator but this does not proof anything and you have to go way deeper than this to proof imbalance.

On July 02 2014 04:05 Hider wrote:


I don't disagree here. But statistics rarely give 100% certainty. This is by by far the best indicator out there. So if your going to assess balance by statistics, you need to look at race distribution.

The rest of your post just seems absolutely trollish without any arguments but lots of insults.

For instance you say I didn't proof win/rates underestimate the issue, but I just did when I demonstrated that if one race is UP, then fewer players will be able to qualify for competitive play, thus only the very best will show up on Aliguac rankings. That's, how the system works per definition. It sems to be you haven't really grasped that (somewhat basic) concept yet.

And it's obvious an extremely logical assumption to make that the average terran player = same skill level of toss/zerg. You cannot make any other type of assumption unless you have extremely good evidence for it.

On July 02 2014 05:04 Wombat_NI wrote:
I don't understand what you're arguing about here, nobody is saying that the Dwf's post is a masterpiece in statistical analysis.

It's the opinions and anecdotal experiences of a GM Terran, his personal take on the game's development, sprinkled with illustrative stats to accentuate the points he's making.

Again this can only be a joke right?
Well you can never proof something 100% but you could at least make a qualified assumption with a margin of error (not sure on the term) the size of about 5%.
Well my statistic courses are quite some years back but what i still do remember is that there are more qualified ways to judge something than to look at a single factor,

How is the racial distribution in GM? Or Masters?

What if the maps are just bad?

What if Terran has a steeper learning curve and people tend to think "well i'll just go with p/z than"?

So let me just go ahead and say there are just as many good white basketball players as there are black (everything else would be illogic), therefore the NBA must be racistic as there are way more black players.
Now you go ahead and disprove me and you better have extremely good evidence, because where would we be if we had to proof our facts instead of getting them disproven.

On July 02 2014 05:46 Wombat_NI wrote:
Why must everything related to SC2 have to be backed up with rigorous statistics, prevailing public opinion and intuition are enough for many legislatures (corruption aside).

Margin of error has nothing to do with the assumptions you are making. Your comparaing oranges with apples here.

24% of players in GM are terrans. 28% of master league players are terrans. 30% of all sc players are terran. Didn't I write that part in my inital post?


Whether it's the map or the race it self that is bad is irrelevant here. At least given the map pool, terran is underpresenting.

Then we would be able to see that with Code S terrans/PL terrans dominating. We actually saw that back in 2011 where terrans didn't do to well in the foreign scene/on the ladder, but performed very well at the highest level. If you read the OP, you would be aware that terran is underperforming at the highest levels, and thus you would conclude that the steeper learning curve isn't the answer.

No it can be quite obvious that there for cultural reasons are differences between how blacks and whites on average perform. But we have absolutely no information on the expected skill difference between protoss players, zerg players and terran players. Unless you have really good knowledge on this subject, you have to make the assumption of equal skill if your going to use statistics to assess balance.
Some have argued that terran is "supposed" to do worse on average becasue it's the campaign race and the human race is more intutive for casuals., however, that argument is becoming less and less relevant as casuals at this point in time are likely to have dropped off the game. So the casuals that played throughout the campaign and then tried a couple of ladder games aren't appearing on the Nios.Kr ladder statistic.

On July 02 2014 06:00 Poisonbox wrote:

Mind if i ask how i am comparing oranges with apples?
A confidence interval is a pretty common scenario for such a situation i guess.
And the margin of error (again if im correct on the term) would most likely be chosen with 5% or am i completely wrong here?



One would think that it's relevant in a terran UP thread if terran was UP but apparently it's not.




Please elaborate a little further, what are the cultural reasons and how do they affect the performance?
As this seems the be quite obvious you surely can easily explain it to someone like me.

Why do we have to make the assumption of equal skill if we have absolutely no information about it?
Seems a little unsatisfying to me.
Do we have to assume that everyone on this forum has the same race, gender and age?
After all we don't have any reliable information about it and might as well just assume it.

Your justification for the assumption of equal skill is that we have no information about it so we might as well just assume it, this might work for you but is in no way scientifically reliable.
Is it really that hard to just admit that the statistics are plain bs and no foundation for a discussion?
I mean go to anyone, your teacher, professor, someone knowing a little statistics, mb even take some time and learn a little about it, but stubbornly refusing to believe it, is that your way to go?
In the end i am not attacking you and you could easily back out if it hurts that much.

Why am I even arguing on the internet, might as well talk to my dog tt.

P.S reading through the posts in this thread everybody already seems to know how much skill each race takes :D

On July 02 2014 07:13 404AlphaSquad wrote:

So with your argument we could go back to wol 2010. It is as safe to say that zergs were simply not as skilled as terrans back then. I dont think anyone picks their race because of their skilllevel but based on how comfortable they feel with playing them.

On July 02 2014 07:13 404AlphaSquad wrote:

So with your argument we could go back to wol 2010. It is as safe to say that zergs were simply not as skilled as terrans back then. I dont think anyone picks their race because of their skilllevel but based on how comfortable they feel with playing them.