|
On April 26 2011 03:49 Azuzu wrote:Show nested quote +On April 26 2011 03:21 Day[9] wrote:
Does anyone have statistics on the overall race usage percentage?
Something like
"40% of people who own SC2 play Terran on ladder."
I certainly know that, for new players, there's a skew towards Terran / Protoss as the single player campaign introduces both of those races, but not zerg. That may be true for some very new players, but I'd wager competitive players in GM didn't pick their race based on the campaign. Also consider "zerg" has become a very popular term that people who haven't even played played starcraft will understand. There could even be people that pick zerg BECAUSE they didn't get a chance to play them in the campaign. Why people play what race is baffling. Take World of Warcraft... when the game first launched almost every server HEAVILY alliance favored. And now? Most servers are horde favored. Why? There was minor racial balancing or whatever, but nothing that really explains the shift from a balance perspective. You could say that when people started out, they wanted to be more humanlike, but as the game evolved, the horde become the "cool" race. This is probably too much of a tangent, and mostly deals with the masses not the top gamers.... In my experience with games, the most top players play the (perceived) best race/character. There's always different reasons/justifications like "this race/character fits my playstyle" but what it really boils down to is that people play with the best chances of winning because that's what makes you feel good/justified in spending your time doing something. For the masses, it's like picking a favorite color. There's no real "reason" to like orange or maroon. For top players, it's like picking a color to put on your product at the store. There's a very real, statistical impact this decision makes on your sales.
I agree, orange and maroon both suck.
|
This is only the case because:
1) Terrans are less susceptible to cheese/allins
2) Mirror is more figured out
3) Ladder maps are smaller than tournament maps. Terrans do a lot better on small distance maps.
|
Terran is the least likely to lose to build order poker and because of small mistakes. Therefor, in an enviorment where you are playing lots of games, Terran will do better, especially since you're playing lots of games in a row, and this makes you likely to lose to small mistakes. For a best of 3 or best of 5, Terran isn't better, but for 100 games, I think they are.
This doesn't mean they're imbalanced, I'm just saying. What do you think of this? Am I right or wrong? and if I'm wrong, please tell me why.
|
Keep in mind that this includes all GM ladders, and I know for a fact SEA GM isn't that competitive.
|
United Arab Emirates492 Posts
Narrow it down to Kr + EU ladder only ! others are not that competitive or representative of highest skill level.
|
the map pool ladder : DQ, close positions on metalo and ST, water gulch. It speaks for itself.
|
On April 26 2011 09:50 Ruscour wrote:
Keep in mind that this includes all GM ladders, and I know for a fact SEA GM isn't that competitive.
oh please, you want to have an accurate competitive picture but then you want to leave out players like idra, kiwikaki, and select?
|
On April 26 2011 10:10 Jayrod wrote:Show nested quote +On April 26 2011 09:50 Ruscour wrote:
Keep in mind that this includes all GM ladders, and I know for a fact SEA GM isn't that competitive. oh please, you want to have an accurate competitive picture but then you want to leave out players like idra, kiwikaki, and select?
they don't play on SEA..
|
Terran is clearly over powered.
|
Nice one OP, you successfully trolled Terran players. This thread should've been insta-locked. It's too blatant.
|
Alright, I have a few problems with this sort of analysis.
1) Skill level is not certain to be equal across ladders. A win on the Korean ladder could be more difficult than the NA ladder for instance. So mixing all of the GM leagues together is probably a mistake.
2) Your method of showing the race distribution of the Top XXX players doesn't produce very informative results. A better strategy would be to split the players into "bins" of 10. Or 50, as I had done some time ago. If I plug the new leaderboard into my old spreadsheet, I get something that looks like this:
![[image loading]](http://img199.imageshack.us/img199/8597/racialpreferenceapril20.png)
No obvious winner there, is there?
3) There is no test for statistical significance. This has been brought up before, and people have provided their own significance tests, both ANOVA and chi-square, so I won't dwell on this. For what it is worth, my spreadsheet also showed no statistically significant anomalies in racial distribution.
I believe it is possible to put together a statistical analysis of balance using the ladder. I don't think this is it, though.
|
+ Show Spoiler +On April 26 2011 10:32 Scarmath wrote:Alright, I have a few problems with this sort of analysis. 1) Skill level is not certain to be equal across ladders. A win on the Korean ladder could be more difficult than the NA ladder for instance. So mixing all of the GM leagues together is probably a mistake. 2) Your method of showing the race distribution of the Top XXX players doesn't produce very informative results. A better strategy would be to split the players into "bins" of 10. Or 50, as I had done some time ago. If I plug the new leaderboard into my old spreadsheet, I get something that looks like this: No obvious winner there, is there? 3) There is no test for statistical significance. This has been brought up before, and people have provided their own significance tests, both ANOVA and chi-square, so I won't dwell on this. For what it is worth, my spreadsheet also showed no statistically significant anomalies in racial distribution. I believe it is possible to put together a statistical analysis of balance using the ladder. I don't think this is it, though.
Um, I'd have to disagree. On that chart protoss is clearly the winner. The blue line is at the top almost the whole way and only at the bottom in 1 spot....
|
Terrans have a 40% representation in the top 100 in both Korean and European ladders and have the most people with very high win rates...and not by a small margin.
NesTea had the right idea when he said it is much easier to do well with Terran, Protoss takes a long time to get good at because there are so many things you just lose to without being able to do much, a lot of timings to learn and a lot of games you must lose to figure it all out. Even a Protoss player as good as oGsMC still gets crushed by simple Terran build order pushes, like the Banshee,Tank,Raven,Marine done by MKP or the Banshee/Raven/Marine done by plot.
In the GSL, how many people are confident in a Toss player as they are confident in a player like MMA, SuperNova or MVP? MC, possibly Alicia, and the scary thing is there is a big difference between those two players still, maybe in a few years this might all even out, when everyone gets really good.
|
Um, I'd have to disagree. On that chart protoss is clearly the winner. The blue line is at the top almost the whole way and only at the bottom in 1 spot.....
You misunderstand. Racial distribution itself is not an indication of balance. The question is if the very top bins are disproportionate to the whole. Mathematically, they are not.
|
On April 26 2011 10:55 Scarmath wrote:Show nested quote +Um, I'd have to disagree. On that chart protoss is clearly the winner. The blue line is at the top almost the whole way and only at the bottom in 1 spot..... You misunderstand. Racial distribution itself is not an indication of balance. The question is if the very top bins are disproportionate to the whole. Mathematically, they are not.
I didn't state anything about balance. All I was trying to say was I disagree with the statement at the bottom of the graph. My interpretation of that graph is that protoss is clearly the most preferred and zerg the least. In 13 of the 20 bins, protoss is the most, in 2 more it is tied for first and in 1 it is tied for second. There is only 1 bin on the chart that has protoss in last (not including ties)
|
Terran has a lot of sick players that's for sure. Just like BW, I think Terran rewards players with the highest skill set. I've thought Terran was the best race played perfect even in BW days. That's why I picked them, no race rewards skills more. Terran will always attract the best players because of this.
|
On April 26 2011 11:07 Reborn8u wrote:Show nested quote +On April 26 2011 10:55 Scarmath wrote:Um, I'd have to disagree. On that chart protoss is clearly the winner. The blue line is at the top almost the whole way and only at the bottom in 1 spot..... You misunderstand. Racial distribution itself is not an indication of balance. The question is if the very top bins are disproportionate to the whole. Mathematically, they are not. I didn't state anything about balance. All I was trying to say was I disagree with the statement at the bottom of the graph. My interpretation of that graph is that protoss is clearly the most preferred and zerg the least. In 13 of the 20 bins, protoss is the most, in 2 more it is tied for first and in 1 it is tied for second. There is only 1 bin on the chart that has protoss in last (not including ties)
My point is that which race has the most players is largely irrelevant, what matters is that it doesn't significantly change from bin to bin.
|
On a global level, GM is
38% Protoss
30% Terran
29% Zerg
The terran's have the most average points. Potential explanations for this include:
Theres alot of Protoss in the GM league that don't deserve to be.
Terran rewards mechanics the most/has the highest skill ceiling.
Co incidence.
|
Coincidence.
I think we have a winner in that one.
|
Nice statistics, always enjoyed looking at the race and win distributions for the respective races.
Of course, this is the ladder, which is not an exact representation of balance and many other factors affect the state of the ladder, which lest we forget is in constant flux.
Also, a lot of top pros don't feature prominently on the ladder, and there's differences in skill levels in general across the ladders, especially with the newly established Chinese ladder and the godlike Korean ladder...
Just a few caveats to keep in mind, I guess, when enjoying on this veritable statistical feast.
|
|
|
|
|
|
EPT
