A lot of the analysis in sports is pretty statistical these days.

Take a look at some of the SABERmetrics crowd in baseball say (multiple regression for example, is often used --- and more sophisticated things than that), or the models behind the Duckworth-Lewis system in cricket. A lot of sports science is heavily statistical (in our sense of the word) these days

On the other hand, I'd say the same about transportation.

Sure the summary stats the SABERmetricians come up with are just statistics in the more common sense, but the analysis that gets to them is often sophisticated

@Kodiologist that observation raises an interesting question: sports are a subset of games. Which is most appropriate? Or do we apply both tags?

@GeneralAbrial It might be best to get rid of [games] and have one tag for card games, one tag for video games, etc.

Good idea @Kodiologist. think that a [sports] tag would be helpful. We have [games] but it is too generic and people do not associate it with [sports] immediately (I did not when checking and the fact that extremely few of the questions tagged with [games] refer to sports -the vast majority refers to poker hands and board games- shows that most users of the tag [games] do not either.)

@gung: The scope of tag will be questions where the user is concerned with the application of Statistics in a sport/sporting environment. I think it will be a convenient conceptual tag. And as Glen_b mentions some sports metrics are not trivial at all. (Glen_b one of my favourite sports/maths application is the Perron-Frobenius theorem and its use for ranking - what can I say, I enjoy Lin. Algebra.)

@Glen_b: Yes, [transportation] is very heavy on Statistics but the potential issue is that it has very low usage.

@usεr11852 Yes, it hasn't had much use here yet

I am not at all convinced that [sports]->[games] is a good choice

@GeneralAbrial: sports is not a subset of games! how are e.g. swimming or running "games"?

@amoeba: Well... I think of a game as a situation that the actions for the optimal outcome (usually winning) are dependent on the actions of the other players (usually opponents). In running, for example you need to run faster than the others. Running a 10.0" 100m sprint would make a male teenager the faster teenager ever and would be unbelievable performance for an amateur but it would not suffice (probably) for a podium finish in Olympics because there are pro-athletes.

Clearly there is no element of cooperation in a 100m sprint but for longer distances like marathons etc. there is a clear element of cooperation in the function a "rabbit" (or more elegantly worded "pacemaker") runner.

@Glen_b what do you mean exactly by "[sports]->[games]"? Do you mean migrating a tag? In general I think that [sports] is better used the score described to gung above.

@usεr11852 It's a fine definition, but then almost everything in life is a game.

@amoeba: Isn't that a fact of life?

:-) This would make a tag rather vacuous though. We can also have a [life] tag. Or [42].

[42] would be a nice Easter egg tag I guess! :D

Games was an example used at length by the philosopher Ludwig Wittgenstein to illustrate the variety of meanings a word could have. (Apologies to anyone to whom the words 'the philosopher' or 'Ludwig' seem unnecessary.) Incidentally, I am reminded of the quip 'Statistics is not a spectator sport', meant to be read as 'Statistics is not a spectator sport'.

@NickCox: Nice quip, but I am unsure about the relevancy of it in 2016. Given some excellent articles on 'Significance' and other "public interest" journals/avenues (eg. 5-3-8) , hasn't Statistics become a bit of a spectator sport?

@amoeba ^ what they said!

@usεr11852 If we get rid of the tag sports in favor of the tag games we would merge the first tag into the second.

That's the intent of the symbolic [sports]->[games]

It represents that merging

Cool, thanks for the clarification. I do not think that merging the tags is a good idea.

@Glen_b I know this question wouldn't belong in CV, because it is software-specific, and that the answer is probably very straightforward. I wonder if you could offer some hint as to what I'm overlooking. Thank you!

@AntoniParellada The underlying issue isn't software specific though so it might survive. But it's effectively a duplicate, since this has arisen many times across different software. There's more than one (equivalent) definition of the test statistic

I could perhaps suggest a way to edit it so it would at least be likely to be seen as on-topic by the more zealous voters but it would probably end up closed as a duplicate of another question

Both Wilcoxon tests have multiple definitions that are all linear transforms of each other