I don't see anything wrong with what Chris just said, but I'd like to add a fairly quick mathematical note. Regression "to" the mean is better read as regression "toward" the mean. It's really a rather weak term, mathematically. It simply implies that when the sample observation you draw is an outlier (i.e., bad injury luck), the next sample observation is likely to be
closer to the mean. Again, for example, if a really large number of pitchers (say, 8) experience an injury in the first quarter. All regression towards the mean says is that the number of pitchers injured in the second quarter will be closer to the mean than the outlier (in this case 8) is. It doesn't say that the next observation will be "close" to the mean! So, if the mean is 2, regression to the mean does
not state that the injuries will be 1, 2, or 3 (all of which are "close" to the mean). The number of injuries could be 7, and you would
still have regression "towards" the mean, simply because 7 is closer to the mean than is 8.
I point this out because, from my reading, it looks like there is some confusion among sports fans on this subject. Some people seem to think that regression to the mean implies that
every sample observation following an outlier will bring the sample closer to the population mean. That isn't what it says. It refers to
only the observation
immediately following the outlier (and no further observations, so it says nothing about actually getting to the mean), and it doesn't even say that
that one observation will be "close" to the mean, in any real sense. Really, regression towards the mean is a very weak concept, and I think that sports fans tend to read more into it than is actually there.
For further insight into the subject, you can go to
https://fs.blog/regression-to-the-mean/ (just one of several websites devoted to the subject).