Related to our recent post on the absurdity of open access fees, somebody recently informed me that a group of statisticians that work in sports decided they were sick and tired of this with the Journal of Quantitative Analysis in Sports and launched a new open access journal, the Journal of Statistics and Data Science in Sports:
JSDSS was founded on three core principles.
First, our commitment to open access is more than just lip service. At JSDSS, open access means free to read and free to publish—no exceptions. JSDSS is a Diamond Open Access journal. Research supported by academic institutions, public funding, or personal effort should not require a payment to reach the audience it deserves.
Second, reproducibility is essential and not an afterthought. The credibility of sports analytics research depends on our ability to verify, replicate, and build upon earlier work. JSDSS will actively incentivize transparency in data, code, and methodology, and work that meets our reproducibility standards will receive a special designation recognizing this commitment.
Third, we believe sport is a rich and underutilized laboratory for statistical and data science innovation. From player evaluation and in-game strategy to league design and fan engagement, sports data present compelling, real-world problems that demand rigorous and creative analytical thinking. We intend for JSDSS to be the definitive venue for this work.
Cool! I should send them something. We think about statistics and data science in sports a lot around here.
My only concern is that I get the impression that the cutting-edge work on sports analytics is happening outside of academia. So I hope that this new journal can get useful contributions from people who are working in sports analytics who have material they can share without compromising their competitive advantage.
What contributions can academic statisticians make to sports analytics?
Or we could flip it around and ask, What contributions can academic statisticians (like me!) make to sports analytics? Here are a few things:
– Developing general methods that can then be used in sports analytics (as here);
– Writing textbooks explaining general methods that can then be used in sports analytics (as here);
– Teaching students who can then work in sports analytics, or consulting on sports analytics projects, which can be thought of as a form of intense teaching;
– Doing work in sports analytics which, although it is not cutting edge, can still give valuable insights (as here);
– Evaluation and criticisms of published work in sports-related topics (as here);
– Contributing to the sports analytics community (as here, and indeed as in the present post);
– Collaboration on sports strategy, or sports medicine, or the sociology of sports, or various other places where sports links up with academic research;
– Clearing up confusion on topics related to statistics and sports (as here), along with social-sciency thinking about how these misconceptions persist;
– Sports-related research where it can be helpful to have an outside perspective, something available to academics who aren’t on a deadline (as here).
There are probably some more things I didn’t think to include on this list.
Another area where statisticians can help: determining the effects of various sports on longevity. I hear that is publishable research, perhaps even open to public funding.
The journal sounds terrific. Thanks for flagging it.
I was a little surprised to see Miller’s old Hot Hard research cited as ‘clearing up confusion.’ I was under the impression that research by Lantis and Nesson, using the same data, had provided an important correction to Miller’s analysis. They concluded “any hot hand in basketball is only present in extremely similar shooting situations and likely not in the run-of-play.” That is, Miller’s limited “hot hand” was evident only in mechanically repetitive exercises (like Gilovich’s experiment), but we seem to still lack evidence that it’s a meaningful factor in actual sports. Has Lantis/Nesson been effectively rebutted? (I confess I don’t stay abreast of this research.)
https://www.nber.org/system/files/working_papers/w29468/w29468.pdf
https://www.nber.org/system/files/working_papers/w26510/revisions/w26510.rev0.pdf
Guy:
Miller and Sanjurjo’s hot hand research, as well as related work by others, did clear up a lot of confusion, because they demonstrated the subtle errors of the original Gilovich et al. work. Going forward there’s room for further research in basketball and other sports. I think the best next step (which is probably being done already by various teams in research that has not been made public) will be to go beyond binary hit/miss to consider the difficulty and accuracy of the shot. Value will come from including more data. But Miller and Sanjurjo’s work is conceptually very important in that it resolved some apparent paradoxes. I don’t think there’s any need to “rebut” the papers you cite; researchers can just study the problem more directly using relevant data.
Andrew, on behalf of the other founding editors, thanks for the shoutout to *JSDSS*.
We would certainly welcome a submission or three to JSDSS.
I think there is plenty of cutting edge sports work that is being done in academia. The program for the upcoming Cascadia Symposium on Statistics in Sports (CaSSiS), https://www.cascadiasports.com/, is one example. Another example is the interest from professional teams in hockey, basketball and baseball in several of the sports talks at JSM this year.