|
Really nice article. The fit doesn't appear to be right on the graph: any chance you could post a list of how many comments there were for each month for the whole period (both broodwar and sc2)?
I'm not sure if region locking is inferred from the data: even if you had a better fitting, more sharply diving regression, you would have to be careful in picking out causation from correlation. Less may be written about foreign players not simply because of the region lock, but because there are fewer viewers to start with due to an external reason. Because we have had a partial region lock this year, it would be very interesting to see the data up through July of 2015, instead of cutting of at December 2014.
Also would be interesting if you could collaborate with Conti's streaming numbers to see if they tell a similar story in terms of region locking. (I took a cursory glance, but the evidence did not look convincing in terms of showing a foreigner inspired uptick, so I would love to see what you would make of it.)
|
Why linear interpolation, why?
|
Ok all wow! I'm REALLY happy with the response this is getting and I'm really happy that I've sparked some discussion. I won't be able to quote everyone who's commented but I'm going to address some stuff that people have mentioned.
Graphs - Yeah I know, they're kinda shitty. I just googled 'make scatter plot' or 'make line graph' and used online websites. All I did was plug in the data. Speaking to the first scatter plot directly, take away the first 24 pieces of data, and remove the 'Idra outlier' (official name now) and there's still an upward trend. As far as I can tell, the math in calculating the linear regression is legit.
SC2 needs 'x' to be popular - I am ONLY commenting on the professional starcraft scene. I'm not trying to make any claims about game balance, or price point, or features, or anything at all besides the professional SC2 scene, the data that I've collected, and what conclusions I feel can be drawn from it. I recognize that there are a lot of factors in play, and I was hoping that a few of them would come up when I was doing my research, but I have NO data on those other factors. So the statements I make in the article are the only things I really feel comfortable commenting on based on this data.
The data you collected is bad because... - I thought long and hard about what sort of data I wanted to collect. It wasn't possible for me to look at the forums. Poles didn't work at all. SCreddit wasn't possible as a source for this type of project. There were no other websites that have existed for as long, and have had the prominence that TL.net has had. The most unbiased set of data that I could look at, that still showed the trends I needed to see, was the news articles on TL.net. The best way to measure growth is obviously community activity. So I looked at the comments posts were getting. It seems to make sense to me that the news articles that received the most attention in the form of comments were the stories that drove interest in StarCraft one way or another. There isn't really anything that I can say beyond that. I think some people are misunderstanding how I determined which articles were 'most popular' as well.
With each scatter plot, there is a linear regression. That's a line that measures where data would be EXPECTED to fall, creating a sort of average. If you want to make a prediction about future data, you extend that line and you should have an idea of where data will fall!
So from 2004-2009, the scene and TeamLiquid.net was growing. That means that there was an upward trend. Data was increasing in size as time went on. The stories that I marked as 'most popular' are stories that fall above the linear regression. That means they were more popular than the data would suggest they'd be. In theory, logic says that those are the stories that helped drive growth.
From 2010-2014 the scene and TeamLiquid.net shrunk. There was a downward trend where data decreased over time. More posts fell below the linear regression, forcing it to form that downward slope. Still, the stories that fall above that line are stories that were particularly interesting in some way shape or form. They had more people interacting with them than one would assume by looking at the data. So those are the stories that are most 'important' for the scene in this time period.
That's how I marked articles and stories as 'most popular' for their respective time periods. That's why they are relevant.
Will you post the raw data? - Probably not unless there's some SERIOUS interest. Its just a ton of numbers and I don't really love the idea of typing it all out (I have it stored in a physical notebook atm). I will if there's enough interest though. I'd do anything for you, SC2 community <3
I disagree with your conclusion/I think your conclusion is hella biased - I did warn you more than once that personal bias would come into play!  I believe I even prephased my conclusion with such a warning as well!
All I can say is that numbers don't lie. You're more than welcome to disagree with me and I'm so happy that you read the article and looked at the data to the point where you feel you can form an opinion!
To be honest, I wasn't sure that I was going to get this result when I started! I was actually quite worried that I would find that the scene cares more about top level play and the KR region than anything else! I'm a huge foreigner fan! But that isn't what I found. I think that at this point, you have to have blinders on to say that the lack of a region lock on WCS in 2013 had no effect on the decline of SC2. That's a little harsh and I mean no disrespect, but everything I've seen says otherwise.
At the end of the day, I think Olympic style region locking is the way to go for WCS. I think that people who want to see the highest level of play aren't watching WCS anyway. I think that at no point were WCS NA or EU representative of 'the highest level of play' because they were featuring the Koreans who didn't feel totally comfortable competing with the best of the best in Korea. I think that if the scene cares most about Koreans, no one loses with pure region locking, only foreign fans will win when their favourite teams/players get to come first in WCS for their region.
There's my big shpeal for now, I hope I've appropriately responded to some of the comments I've seen! Thank you everyone who's been offering such kind words of support and everything! Thank you all for reading and for participating in the conversation! I think that it is a relevant topic!
|
cyan just plot your data in a spreadsheet like LibreOffice Calc, Google Spreadsheet, Microsoft Excel or a math software (even wolframe alpha), but please get in quadratic or something and with the option of not zeroing.
Also would be nice to indicate year/month on the graph, right now we have to "calculate" the start date and add the number of months (up to 60) to know when are the results in order to link 'em with the little information snippet on popular stories.
As for the data I dunno why you went back in time to get inconsistent post counts (time period missing, some data few days old, others nearly same day) instead of taking all your results from today. Yes you would get few more results because of "late" posters, but usually a subject dies out withing a month and TL is quite diligent about their necro policy. It would at least get the same "cooling" period to each news.
As for how you interpret your results and even your "solutions", well let's say you "assume" quite a bit while your only point demonstrated was about news only, and as discussed before they themselves might have been skewed by the way they were collected and analysed.
I'm sorry if this community is a little hars, but some have scientific (or autistic) backgrounds. Hell just look at how we demolish all the "math behind xxx" threads when there's one comma missing.
Thank for the effort, but like in SC2 it takes way more to reach perfection.
|
On August 18 2015 13:55 varsovie wrote:
cyan just plot your data in a spreadsheet like LibreOffice Calc, Google Spreadsheet, Microsoft Excel or a math software (even wolframe alpha), but please get in quadratic or something and with the option of not zeroing.
Also would be nice to indicate year/month on the graph, right now we have to "calculate" the start date and add the number of months (up to 60) to know when are the results in order to link 'em with the little information snippet on popular stories.
Yeah I'm not great with ....well.....any of those programs. I don't know how to use them. I'm really sorry about the por quality of the graphs. I thought this would be a-ok but it could have been better. A lot better.
[B]As for the data I dunno why you went back in time to get inconsistent post counts (time period missing, some data few days old, others nearly same day) instead of taking all your results from today. Yes you would get few more results because of "late" posters, but usually a subject dies out withing a month and TL is quite diligent about their necro policy. It would at least get the same "cooling" period to each news.
Oh man I totally forgot to talk about this! The reason why I didn't just look at the site as it stands today is twofold!
1. I wanted to get a picture of the stories and the community reaction to them as they happened. Not two months later. Not two years later. Within one month of each article being posted. There's much less room for fudged numbers that way, and its a unique chance to see a snapshot of the community frozen in time.
2. I would have NO idea of what to be looking for. I'm not a frequent TL user right now, I'm not familiar with navigating the website, and I don't even know if I would be able to get as far back as 2004. Right now if I go to the News section and click to the last page available, I get to 2008. The Wayback machine offers a deeper timeline with less hassle.
[B]As for how you interpret your results and even your "solutions", well let's say you "assume" quite a bit while your only point demonstrated was about news only, and as discussed before they themselves might have been skewed by the way they were collected and analysed.
I'm sorry if this community is a little hars, but some have scientific (or autistic) backgrounds. Hell just look at how we demolish all the "math behind xxx" threads when there's one comma missing.
Thank for the effort, but like in SC2 it takes way more to reach perfection.
No one is being harsh (yet ;_; )! I like the discussion that's being had and I feel like more people are agreeing with the information I'm presenting than are disagreeing! I know that this article is by no means perfect, and there are things I can learn from for next time! I don't have a math background so I appreciate the feedback and I hope my responses make sense. I think that the data I collected is valid but I always knew some people would disagree! And that's so ok! TL is a very intelligent community and I'm more than happy to be sparking some discussion if I can't convince everyone!
|
Cyan, umm... You go "I am biased", "My analysis is crappy" and at the same time you say that you are "really comfortable" about your comments and "numbers don't lie".
This thing with what the data actually shows, and what your opinion is are two different things. Some things here are facts from the analysis, which we can disagree on because some do the analysis in a flawed way, some add in personal assumptions that others don't agree on, but ultimately there is an objective truth about what the data says, and which assumptions are needed to reach what conclusions. When it comes to statements like "the data shows this", you can't guard yourself with "this is my opinion, disagree if you want"; you can be plain, objectively wrong. Opinions don't enter in that at all. Bias together with lacking analytical skills (both of which you have happily admitted) will skew your answer, but you are still wrong.
To guard yourself from bias as much as possible, it is important that you lay out your argument clearly and slowly for how you go from the data, through the analysis to the conclusion you claim. I don't feel that you did that at all regarding the statement first in your conclusions: "When it comes to eSports in general, this data suggests that the most important thing to stimulate growth is to ensure that the professional scene is locally involved." If you want to make that statement, you need to lay out every logical step from the data leading to it, which you haven't done whatsoever. You just made some random plots on thread views, and then make this huuuuge statement about how eSports works out of nowhere, and claim that is based on the data.
Again, I am not necessarily disagreeing with the statement itself, only your claim that it is based on the data.
|
Well first off I NEVER said my analysis is crappy. I think my analysis is spot on in fact . My graphs could have been better in terms of how pretty they look, but that wouldn't change the numbers on said graphs.
My bias comes into play when I start talking about WHY the foreign news stories are more interesting to the community. Its bias because there is no possible way that I can measure how important it was for fans in 2010 to identify with local pros, but the data leads me to that conclusion. Likewise, I have no idea what made the LCS teams decide against opening SC2 rosters, I wasn't invited to those meetings. But I can use what I know about the history of SC2, what I've learned about the fans from this project, and what I know about these organizations, and I can make very educated conclusions about what PROBABLY happened. Those spaces are where my bias comes in. I don't think I'm wrong at all, and frankly I think that looking at these statistics and drawing any conclusion other that 'yeah wow the foreign scene is actually more important to fans than KR' is incorrect. But I respect that opinion and I welcome any discussion on it. I arrived to my conclusion by gathering data on the most popular news stories on the biggest SC esports site, categorizing them based on what scene they were pertaining to, and simply counting where most of these popular stories were coming from. A majority of them come from outside of Korea. I have to disagree with the idea that I took a leap at this point when I said 'When it comes to eSports in general, this data suggests that the most important thing to stimulate growth is to ensure that the professional scene is locally involved'. I think that's exactly what the data is showing.
I don't think that my numbers are incorrect at all. I think that the data I gathered is valid, shows the trends we would all expect to see, and I think that I've gone about analyzing it in the right way.
I think this is the last post I'll make to this thread for today at least. I feel like I'm bumping it the most and I don't want to just self bump all the time! I'm really happy to respond to questions and comments! If I wasn't able to defend my methodology or results then it wouldn't be a very good piece at all right? :D
|
On August 18 2015 03:29 CyanEsports wrote:
Before starting in I want to mention that this is by no means a comprehensive history of all the important events that occurred in StarCraft eSports’ history. There were many many stories that I didn’t see at all. Huge events like Idra being dropped from EG don’t show up in TL’s news section. I assume that this is because Team Liquid eventually became a major StarCraft team itself and didn’t report on major happenings from its competition.
I'm not sure how you found your news stories but we definitely report on happenings in the scene regardless of whether it's from "competition" or not. The thread you mentioned was pretty big and can be found at http://www.teamliquid.net/forum/starcraft-2/411840-evil-geniuses-releases-greg-idra-fields
|
On August 18 2015 23:33 R1CH wrote:Show nested quote +On August 18 2015 03:29 CyanEsports wrote:
Before starting in I want to mention that this is by no means a comprehensive history of all the important events that occurred in StarCraft eSports’ history. There were many many stories that I didn’t see at all. Huge events like Idra being dropped from EG don’t show up in TL’s news section. I assume that this is because Team Liquid eventually became a major StarCraft team itself and didn’t report on major happenings from its competition. I'm not sure how you found your news stories but we definitely report on happenings in the scene regardless of whether it's from "competition" or not. The thread you mentioned was pretty big and can be found at http://www.teamliquid.net/forum/starcraft-2/411840-evil-geniuses-releases-greg-idra-fields
Woah shit hi there!
I never meant to imply that these stories didn't exist on TL.net, but they don't appear in the official 'news' section. Unfortunately, the way forum posts work and the way the Wayback Machine works means that I wasn't able to look at the forum section of the site at all. It moves too fast for the Wayback Machine to log it effectively and the numbers would be enormously skewed if I went back to look for these stories in the state they are now. I wanted to see the community impression while the stories were as topical as possible.
I certainly didn't mean to imply that TL is shutting out topics on other esports teams and I'm sorry if that's how it reads. I meant that some of the big stories didn't appear under the 'news' section on the site.
|
Looking at the 'Idra released from EG' thread, I'm actually really surprised to see that the last comment is only nine days after the thread was started. I expected way more necro bumps and such! Another issue with trying to go back through the forum to collect 10 years worth of data is that there's just way too much to sort through. There are 706 pages of Broodwar threads alone. With necro bumping in mind, I'd have to open each individual thread to find the original post date. Sadly, unless someone wants to pay me to do this, collecting data from the forum sections of TL.net is just not possible.7
There might be something to an idea for going back and purposefully looking for some of the larger stories that have happened in the past four years though. We'll see.
|
|
|
|
|
|
EPT
