There's also the problem with these stats that players don't stay the same in terms of ability. I mean... Incontrol was able to take the series off Whitera in season, he sure as hell won't be able to this season. TLO, no offense to the guy since he's awesome... just isn't as competitive as he was season 1, same can be said of other players like Moman.

I also wanna point out that there are people talking about how exciting the games are now that the Koreans are gone, but viewer numbers don't exactly agree that NASL2 is more popular by far as compared to viewer numbers in NASL1.

jakethesnake

Canada4948 Posts

November 30 2011 14:48 GMT

#22

On November 30 2011 20:07 Talin wrote:
Honestly, every time somebody makes conclusions about SC2 based on statistics I die a little inside. -_-

This is just one example of how statistics can always be (and often is) wrong.

I completely agree. As someone who does statistics for a living at a university, I sometimes just want to pull my hair out at the gross over simplification that happens with statistics on TL that leads to baseless conclusions.

Grumbels

Netherlands7032 Posts

November 30 2011 14:58 GMT

#23

Obviously when you remove the 10 worst players of the league, the win percentages of the top should go down a bit.

Coeus1

Finland160 Posts

November 30 2011 15:05 GMT

#24

Statistics calculations lose meaning when your starting assumptions can be questioned. In this case the base of your argument is the "returning players win ratio". And that can be questioned.

Itsmedudeman

United States19229 Posts

November 30 2011 15:08 GMT

#25

"overall" player quality I'd believe that. NASL 1 was based off of invites and some players just didn't deserve it and now the lower end has better players. But the top players are steam rolling through the tournament as if they should be in NASL code S and this is NASL code B.

caradoc

Canada3022 Posts

November 30 2011 15:11 GMT

#26

You absolutely cannot draw those conclusions from that data.

Xoronius

Germany6362 Posts

November 30 2011 15:26 GMT

#27

On November 30 2011 20:42 cascades wrote:
[...]It is interesting to note that the "garbage" who remained, NASL managed to retain qxc, hasu and haypro [...]

How can you put Hasu in this? Hasu was far away from elimination in both seasons and was never considered bad at any point of time. He got 5-4 in Season 1 and made it through the playoffs and in season he qualified directly for the grand final, beating the champion of season 1.

@topic: I share the opinion, which was stated before, the average skill level has maybe increased because of many bad players droping out, but the overall strengh was most certainly bigger in Season 1.

phiinix

United States1169 Posts

November 30 2011 15:27 GMT

#28

You can conclude that returning players did worse this season but I don't think you've accounted for enough outside variables to conclude that the flaying field is stronger. A big one would be who they played against; you can't have a high win rate in a group that plays among themselves.

Bold conclusion. I think too bold.

s4life

Peru1519 Posts

November 30 2011 15:34 GMT

#29

Statistics were not wrong, their interpretation was a bit too optimistic though because it didn't take into account interdependencies between events. f.e., groups were different, players improved, latency, etc.

Kira__

Sweden2672 Posts

November 30 2011 15:38 GMT

#30

I mean yeah you could use your statistics and say that they prove that the players are now in a tougher competition, but they could simply also mean that these players have gotten worse since season 1.

I'm not saying it's either one or the other, but I don't think you could make the assumptions you are

JOJOsc2news

3000 Posts

November 30 2011 15:42 GMT

#31

You can't draw these conclusions from these data and say they are robust. I am sure you are aware that this is only an interpretation that by no means holds statistical significance. It includes so much noise and the factor cannot simply be determined to be the player field.

shabinka

United States469 Posts

November 30 2011 15:52 GMT

#32

Just had to comment on this. Statistics are not wrong, its how you interpret them. Don't blame the math for an interpretation that you think is 'invalid'. Its difficult to predict games and with a game was volitle as starcraft its difficult to interpret the statistics.

paintfive

785 Posts

November 30 2011 15:53 GMT

#33

I don't even see how the overall win rate dropping implies "stronger players"

Ricemagical

270 Posts

November 30 2011 15:55 GMT

#34

You're trying to maintain that the statistics from a single tournament represent the entire scene but that's clearly not the starcraft scene is composed of MUCH MUCH more than just NASL. For example, if you take a look at the mlg pool play brackets, machine and incontrol were consistently last in the mid-late mlg events while axslav and strifecro failed to reach the championship bracket in the events they did participate in.

BarbieHsu

574 Posts

November 30 2011 16:04 GMT

#35

But there was no Artosis...

Hypertension

United States802 Posts

November 30 2011 16:24 GMT

#36

The problem is that the players will always either overperform or underperform their potential. By selecting the winners from Season 1 you are biasing your results towards players that overperformed. This is why in any study that shows promise, follow up studies usually show a regression towards the mean, i.e. less impressive results. So it is not surprising that the players that had a slight winning edge in season 1 had a worse performance in season 2. Look at any sport, the top players/teams in any year are almost always closer to average or even below average the next year.

Inky87

United States533 Posts

November 30 2011 16:56 GMT

#37

A 5% difference in a sample size this small means diddly, for starters. Not to mention every other factor besides the players competing like... the rested state of all the players at the time of their matches or the shift in the game. Look at Zergs 6 months ago. They were bemoaning the state of ZvP... and then they figured out how do scary pressure builds while getting 70 drones and now they win loads against the current protoss builds. It sounds like you put some work into this, but it doesn't sound like you've looked at it from all the angles.

Kraznaya

United States3711 Posts

November 30 2011 17:07 GMT

#38

proof that the decision to include artosis in season 1 was even worse than the decision to exclude koreans in season 2

CosmicSpiral

United States15275 Posts

November 30 2011 17:11 GMT

#39

I like the OP takes one article to task for making questionable conclusions off of relatively scanty data, and then he does the exact same thing in an article he writes. Let's ignore the fact that overall everyone played less games this season, individual player skill across the board has heavily fluctuated between S1 and S2 for a host of reasons, the abundance of S1 players who were literally free wins, etc.etc.