On January 14 2011 02:42 Kralle333 wrote:
I guess the peek of random players on platinum is because of cheesing not being that effective any more :D

That's a pretty ignorant thing to say.

Kralle333

Denmark301 Posts

January 13 2011 18:04 GMT

#44

On January 14 2011 03:03 colanderman wrote:

Show nested quote +

That's a pretty ignorant thing to say.

Why? Every random player i play is cheesing

terence158

Australia64 Posts

January 13 2011 18:40 GMT

#45

On January 13 2011 23:00 Faze. wrote:
Makes sense.
Except that I feel like I get A LOT more zerg and terran compared to protoss. Which is sad cuz I like ZvP

note: even though there are less zergs in numbers, they may play more games (not saying they do, but its possible); or their games may be shorter

also, procentage? i was confused for a while, thought it was like, some percentage of pro's or something.

EchOne

United States2906 Posts

January 13 2011 18:43 GMT

#46

On January 13 2011 23:00 Faze. wrote:
Makes sense.
Except that I feel like I get A LOT more zerg and terran compared to protoss. Which is sad cuz I like ZvP

It's because all the Protoss are being matched against each other and thrown into that inglorious clusterfuck of a matchup that is PvP.

Sonictonic

Sweden62 Posts

January 13 2011 18:46 GMT

#47

On January 14 2011 03:03 colanderman wrote:

Show nested quote +

That's a pretty ignorant thing to say.

Not really, From my own and my friends experiences random players cheese alot.

Mushin

United Kingdom15 Posts

January 13 2011 18:47 GMT

#48

Nice work with the graphs, but I don't think you can legitimately draw lines between the points for each league, since the leagues aren't a continuous distribution (they're discrete groups). It also looks like you used a curve for some of the lines, which is slightly misleading given the discrete data. Not to say they aren't illustrative or useful though!

It would be nice to multiply the percentage data by total players per league, to give the absolute numbes for each data point.

DrBoo

Canada1177 Posts

January 13 2011 18:51 GMT

#49

I think its kind of funny that terran win percentage is such ahead of all the other races at master league. I wonder what would contribute to that being so much higher then all the other races?
Maybe its because terran is better overall then all the other races at that level or maybe its because they can pull workers better then all the other races?
I'd love to see a chart with time in game and win percentage see which race fares the best in longer games and which race does the best in shorter game.
Though that kind of data I assume would be rather hard to find and attain.

thesmoosh

113 Posts

January 13 2011 18:56 GMT

#50

Now can we stop pretending zerg is underpowered and weak?

Edit: you people really need to learn how to read graphs. The difference between the top two win%'s in the masters league is 0.5%. That means Every 200 games one guy wins one more game than the other guy.

palanq

United States761 Posts

January 13 2011 19:05 GMT

#51

awesome, <3 graphs

pompey606

United Kingdom98 Posts

January 13 2011 19:12 GMT

#52

This is super awesome thanks for making this

Toxigen

United States390 Posts

January 13 2011 19:35 GMT

#53

On January 14 2011 03:56 thesmoosh wrote:
Now can we stop pretending zerg is underpowered and weak?

Edit: you people really need to learn how to read graphs. The difference between the top two win%'s in the masters league is 0.5%. That means Every 200 games one guy wins one more game than the other guy.

Yes, but I think it's still telling that Zerg is the winningest race in Silver (but the least, discounting Random, is within .5%), but at Master level Zerg is the least winningest (and the highest is greater than .5% above).

I think when you consider the scope and range of skill within a league, these results aren't necessarily incontrovertible but they are significant (i.e., it's not attributable to randomness).

I'm NOT saying Zerg is underpowered by any means, but I do feel like the race is harder to play optimally than Terran or Protoss. I admit I could be biased -- I am a Zerg player.

EDIT: Said Platinum, meant Silver.

idonthinksobro

3138 Posts

January 13 2011 19:43 GMT

#54

the first graph makes no sense unless the space inbetween 2 leagues is meant to show high and low gold for example but i doubt that. If you can give me the row numbers iam gonna look into it a little closer (i have maple and can make better graphs with it) but i think a simple bar chart would fit better for 1.

TeWy

France714 Posts

January 13 2011 19:48 GMT

#55

Hopefully these graphs will shred into pieces the notion that Zerg has the highest skill ceiling of all race, considering how 2* more Zerg players manage to get into the highest leagues than Protoss/Terran.

master beef

United States21 Posts

January 13 2011 20:02 GMT

#56

On January 14 2011 04:48 TeWy wrote:
Hopefully these graphs will shred into pieces the notion that Zerg has the highest skill ceiling of all race, considering how 2* more Zerg players manage to get into the highest leagues than Protoss/Terran.

Not trying argue anything about skill ceiling or be rude but did you even look at the graphs because that is not even remotely accurate. The only league where zerg doesn't have the lowest representation is masters and there is only 0.6% more zerg than terran.
http://sc2ranks.com/stats/league/all/1/all

Kralle333

Denmark301 Posts

January 13 2011 20:37 GMT

#57

On January 14 2011 04:43 idonthinksobro wrote:
the first graph makes no sense unless the space inbetween 2 leagues is meant to show high and low gold for example but i doubt that. If you can give me the row numbers iam gonna look into it a little closer (i have maple and can make better graphs with it) but i think a simple bar chart would fit better for 1.

It is a simple graph. You can only make better graphs if there's some correlation(which is only present at the zerg data).

bramapanzer

83 Posts

January 13 2011 22:15 GMT

#58

If anything this PROVES the assumption that zerg has the highest skill ceiling. But worst vs worst, terran is the strongest. This implies two things, that the terran race is "easier" as the worst players have better chance of winning than an equally skilled zerg. And also, at the very top level of play, best against the best, zerg players can "max out" or hit a higher "skill ceiling" much later.

Don't know if that was clear, so let me put it this way. If a race (Race A) had a very very very low skill ceiling, theoretically, anyone could max out the performance of the race. If another race (RACE B) had a higher skill ceiling, those that are more skilled and hit the "max" or ceiling of that race, they should win against maxed race A.

That being said, NO ONE has reached the skill ceiling for this game, so why even bring it up?

KevinIX

United States2472 Posts

January 14 2011 02:48 GMT

#59

On January 14 2011 03:51 DrBoo wrote:
I think its kind of funny that terran win percentage is such ahead of all the other races at master league. I wonder what would contribute to that being so much higher then all the other races?
Maybe its because terran is better overall then all the other races at that level or maybe its because they can pull workers better then all the other races?
I'd love to see a chart with time in game and win percentage see which race fares the best in longer games and which race does the best in shorter game.
Though that kind of data I assume would be rather hard to find and attain.

The graphs exaggerate the difference a bit. The difference between the Terran and the Protoss win percentage is 0.7% which is pretty reasonable, and simple map change or newly discovered strategy could easily make the graphs flip. And remember the matchmaker tries to keep everyone at a 50% win rate. Population is probably the better stat to look at when thinking about balance, but you can't determine balance solely on graphs like this. And of course, some races could simply be more popular than another.