As you may know from this blog I love digging into topics from the big picture point of view.
As an experienced fantasy basketball player I wanted to explore multiple questions in that area but I hit a wall of complete lack of data. Why?
Because on one side, even if you played in 30-50 leagues in your fantasy career it’s not only a small sample size but it’s also biased – you were in every one of them. So that’s not a good start for big picture analysis.
On the other side, simple search on Yahoo and ESPN revealed there were at least half of million fantasy basketball leagues last year alone. So they were sitting on that kind of data and as far as public is concerned they did almost nothing with it in terms of analysis. For some reason that bothered me very much.
I tried to contact those major providers to ask whenever they did something with it or even just to suggest what they could do but not surprisingly that didn’t work so in February of 2013 I took matters into my own hands.
I immediately stumbled into a problem – most leagues were private so as an outsider I had no access to them. Nothing I could do about that but I knew there were also some publicly visible ones… but how to find them?
Unfortunately during the season there wasn’t a comprehensive list of all public fantasy leagues… at best there were only lists of various Top100 but by definition those were extreme outliers so I wasn’t interested.
So I stopped thinking about the efficient solution and focused on a way which would bring me the most leagues possible – searching using brute force of scripts. As I was the most familiar with Yahoo leagues and taking into consideration their hourly and daily connection limits I searched automatically by ID for publicly visible Yahoo leagues’ settings. One by one starting at #1. Due to aforementioned limits and availability of free time it took me over a month to get to 30.000th ID… and I had to stop because I needed more data than just the league settings. But after such search I had a nice long list of all publicly visible basketball leagues so it was way easier to gather information about the drafts and final standings.
That’s how I collected data from 6788 Fantasy Basketball Leagues from 2012/13 NBA season.
And the first lesson here was… only around 22.6% of all leagues were public!
Overall this project wasn’t pretty or efficient but I accomplished what I hoped for. Mostly. It turned out that some of the leagues existed only on paper so either they didn’t draft at all or they didn’t start after the draft but the most annoying cases where with leagues which… stopped being public! Who does change that option 4 months into the season? I don’t know but thankfully all those examples above where the exceptions which wasted only a small minority of the data collected.
What’s more, while 30 000 leagues checked and over 6000 collected sound like a lot for one person to have… in a grand scheme of things it was basically a tip of an iceberg because judging by IDs Yahoo alone had over 200 000 basketball leagues last year. Maybe I’ll start earlier next year to expand this project but we’ll see, maybe it won’t be even necessary.
What do I plan to do with all those leagues?
I’d like to start with some obvious topics and questions like…
What are the most popular settings in fantasy basketball?
Does seeding even matter in head-to-head fantasy basketball?
Do actual results confirm or deny a theory of unfair snake draft?
Comparing the effect of playoffs in head-to-head leagues to roto leagues.
Is it true that punting in roto leagues is not a winning strategy?
What were the average amount of stats needed to win each category in roto leagues?
and probably many more… I’m guessing that during research I’ll stumble into more topics to explore and hopefully at least one reader will add at least one interesting question to answer.