We are now in the dead zone: Super Bowl’s over, March Madness isn’t here yet and Spring Training has yet to begin. Bringing up this fact is a cheap gimmick, but cut me some slack—I still have to write another one of these intros for the Links tomorrow morning.
The good news is, in addition to the first ever Hardball Times Preseason Book, which will include some cool analysis into the upcoming season that I’m dying to see myself as it comes together, you have something else to look forward to: the annual THT Five Questions season preview series. All columnists and websites need to have some gimmicks—if you don’t, the constant deadline pressure and ridiculously short news cycle will eat you alive. Seriously though, we’ll be running team previews looking at the biggest questions facing every team, and what some of the answers might be, starting mid-March, so mark down your calendars.
Now if you’ll excuse me, I need to think up another gimmick for the Links.
And onto the questions:
The Elbow’s Connected To…
I know Tommy John surgery is common knowledge now, but I don’t really understand rotator cuff and labrum problems. How good are pitchers when they come back? Do you have any examples of them?
- Eli N., Toronto
David Gassko: It’s difficult to get information on injuries—unfortunately, detailed info is not really available anywhere—which makes it difficult to study the effects of any injury. Nonetheless, we do have some anecdotal evidence as well as medical knowhow that suggest that rotator cuff and labrum injuries are much tougher to come back from.
The Ulnar Collateral Ligament (UCL) can be totally replaced in Tommy John surgery, and the result is a UCL that feels and functions like it’s brand new. On the other hand, a rotator cuff or labrum can only be repaired, and that kind of surgery can never get you back to 100%. Experience shows that pitchers who seriously injure their rotator cuffs or labrums generally are not the same pitcher after surgery, and often experience serious complications that make a successful comeback impossible (more often than Tommy John patients these days).
I found John Brattain’s recent articles on “average” hitters and pitchers fascinating. Many analytical models use the mythical average player as a baseline for comparison, and many critiques of managerial strategy base their arguments on assumptions about average lineups. But it seems to me that there are relatively few truly average performers (i.e., players who hit for average power, make contact at average rates, walk and strike out an average amount of time, run the bases with average competence and play average defense) in the league.
I spent just a few minutes looking at the 2006 statistics of everyday players and found only a few who were within 20 points of the ML averages for batting, on-base and slugging. I don’t know what the standard deviations and skews are in these categories, but I wonder how many players could be legitimately labeled as average in all three categories. And I wonder if those average players hold their jobs for long if they never excel in one or more area. Just how rare is it to be average?
Chet O., Lincoln, Nebraska
Bryan Tsao: Something that’s crucial to remember when looking at “average” stats is that those stats are compiled by a combination of starters and backups. In a typical season, a healthy regular will accumulate 500 to 600 plate appearances, with 100 to 200 going to a backup. That’s probably the main reason there are very few “average” players out there—while a few “tweeners” probably get playing time each year, the “average” will always be centered around relative extremes, as you pointed out.
However, with that caveat in mind, it’s still important to compare players to the average, as it gives us a baseline against which to compare players. It is kind of arbitrary, but you do need something. As Jeff Sackmann noted in this article, it’s just important to remember to try to compare starters against starters and backups against backups.
Dave Studeman also noted to me when discussing this question that the mean isn’t the only kind of “average.” In discussions like this, median can often be more relevant, as in most cases baseball’s playoff structure means that it doesn’t matter how much better a team is than the worst team, but rather how good it is compared to the best.
In a recent article, John Beamer mentioned that there exists a theoretical maximum R between projections and actual performances (0.77). Can you explain what the basis of this theory is, or at least link the article where Tangotiger explains it?
- Dave R., Toronto
John Beamer: Tangotiger calculated the theoretical maximum for R when comparing forecast systems on his blog. I won’t try to impress you with my math knowledge when Tango explains it more lucidly than I ever will. You can find the link here.
Here he calculates that the maximum R is 0.73, but that is predicated on 550 plate appearances. If you ratchet that up to 650 you get an R of 0.77 as demonstrated here.
In short, all you need to calculate is the expected and observed variance to determine R. The expected variance depends on the number of trials, so the more trials you have the higher the theoretical R.