I got a couple of questions to the effect of "Doesn’t it seem like there are more pitcher injuries already this year than there have ever been?", to which I replied inside my own head "No, what it actually seems like to me is that every season since I was 13 years old, people talked about there being more injuries this year than there ever were before."    There have always been injuries, but for some reason people don’t learn to expect them. . .or so it seems to me; I don’t know.   When you get old you kind of lose your internal registry on certain things; you can’t tell whether other things are moving faster or you’re moving slower.

                Anyway, it occurred to me that we could track this in a simple way by tracking the number of starts during the season which aren’t by one of a team’s first five starters.  If pitchers get injured, they have to use "other" starters, other than their first five.

Starts by "S6" and greater aren’t exactly the same as "pitcher injuries", because sometimes pitchers are lifted from the rotation for ineffectiveness, rather than injury, and sometimes other things happen—scheduling quirks, trades, minor injuries that cause a rotation pitcher not to make his first start until May 15, etc.    In general, we can say that when injuries increase, games started by pitchers not in the starting rotation at the start of the season increase, and when starts by pitchers not originally in the starting rotation increase, injuries have increased.

                Let’s establish the base lines.    I studied all teams since 1970, except for the strike-shortened 1981, 1994 and 1995 seasons; I left 1972 in the study.  I identified each pitcher’s "Opening Season Starting Rotation" as their first five starting pitchers of the season, and anyone else who started for them during the season as a 6^th starter.

                In the first ten games of the season, Sixth starters average .07 starts per team, or 1% of all starts.

                In the first twenty games of the season, Sixth starters average .67 starts per team, or 3% of all starts.

                By the end of the season, Sixth starters average 26% of all starts:

                Redundancy for the sake of clarity, at the risk of speaking to you all as if you were a classroom full if idjits, we can use this data, during the season, to check whether injuries to starting pitchers are or are not running ahead of schedule.

                This season—to answer the essential question of the project—the number of 6^th starter starts is in fact clearly running high.     I’m not sure my count is right, because it is difficult to chase down disorganized data, but I have 35 6^th starter starts in the majors through 20 games, or 1.17 per team, which is close to twice the normal number.

                Are the numbers different now than they were in 1970?   Yeah, they are, but not all that much different.   From 1970 to 1979, 6^th starters made 24% of all starts.   In the last ten years it’s 27%.

                This chart tracks how the percentage of games started by pitchers who weren’t in the starting rotation at the beginning of the season changes as the season goes on:

                So at this part of the season, early May, normally 10 to 15% of the games are being started by 6^th starters or lower.

                The year in which 6^th starters made the most starts was 2009; 6^th starters made 1,561 starts in 2009, or 52 per team.   This dropped to 39 per team in 2010, 38 per team in 2011.  

                The 1973 Texas Rangers and the 2006 Kansas City Royals got only 49 starts total out of their first five starting pitchers, leaving 113 games (for each team) to be started by 6^th starters.   As you might guess, this is not a good sign; both teams lost 100 games.   These are the teams with the most starts from 6^th starters, and their won-lost records:

                There was exactly one team in the data which only used five starters.  That was the 2003 Mariners, and they finished 93-69.  The 2012 Reds (97-65) only had one start from a sixth starter, and several teams only had two.

                The winning percentage of 6^th starters is actually not all that bad.   In the data there were 46,939 games started by 6^th starters, with a won-lost record (for their teams) of 22,263-24,676, a winning percentage of .474.   That’s not that bad, but the thing is, the more starts you have that are out the rotation, the further down you’re getting into the organizational depth.   You’re not really talking about 6^th starters; you’re talking about 7^th starters, 8^th, 9^th, 10^th, and on down.    Those guys who started the season as the #3 guy in Double-A aren’t going to come up and pitch well all that often.    I would guess you’re talking about a .485 winning percentage from true 6^th starters, but .350 from the back-end guys, the guys who are pushed into major league roles when they shouldn’t be there.   The true sixth starter is very often the pitcher who started the season as the organization’s #1 minor league pitching prospect.