I found injury data through Driveline's injury database*, and compared that with pitch count data from the 2009 season. I calculated each pitcher's pitches per appearance in 2009, and listed pitchers in order from most to leas pitches thrown per appearance. From there, I looked up pitchers one by one in the injury database. I noted how many days they spent on the DL in 2009, and how many days they spent on the DL in 2010.
*Free, to a degree. I looked up approximately 200 players before I was cut off.
I stratified the pitchers I had into groups of 25. So, the 25 pitchers with the highest average pitch counts were the first group, the next 25 pitchers were the next group, and so on, for a total of 7 groups (175 players). Here are the average pitch counts, and days spent on DL, for each group:
The data suggests that pitchers with higher pitch counts in 2009 were less likely to injure their arms. In fact, the number of days on the DL plummets with the highest pitch count group. The results are completely opposite what we would expect if higher pitch counts make pitchers more susceptible to injuries.
Now, I'm quite confident the little table above misses some really important parts of the picture. I doubt pitcher injuries would drop a ton if every hurler tossed 120 pitches every night from here on out. However, it appears that the group which throws the most has also figured out something about durability that others have not.
One thought is that this might be pitching Darwinism, to a degree. Perhaps teams are so cautious with their pitchers now that only the bodies best-suited for durability ever get a chance to accumulate big pitch counts. The high pitch counts are merely a measure of their durability. They have nothing to do with developing or maintaining durability in the first place.
Somehow, I think the Verducci effect fits in DL data somewhere. Pitchers with high pitch counts might actually be among the least likely to have a big spike in innings pitched. Since the prevailing thought in baseball is that most starters are done around 100 pitches, any starter consistently going over 100 pitches was probably limited to the century mark before breaking through. That might naturally limit inning spikes in the top group of the data table in a way that isn't guaranteed for any other group.
Another factor might be how teams use the disabled list in general. More players hit the DL, plain and simple. This is likely due in part to modern medicine's ever-increasing ability to notice and diagnose injuries.
However, let's also be honest. Some players are easier to place on the DL than others, simply based on how replaceable they are. The pitchers firing 105 pitches a game are the workhorses of a staff. In many cases, they are tremendously talented, because they have to do something special to warrant the perceived "risk" that comes with throwing over 100 pitches. Who would you think harder about placing on the DL with something like "shoulder tightness" - Felix Hernandez or Blake Beavan? Who's also more likely to talk Eric Wedge out of getting pulled because of a high pitch count?
Back to my main point though. Conspicuously absent in all these theories is the thought that pitch counts do anything to help or hurt a pitcher's chances of getting injured.
In 1988, at 41 years old, Nolan Ryan made 33 starts. He threw under 100 pitches in only 7 of them. He topped out at 141 pitches in a laborious complete game masterpiece that came rather early in the season (April 27). Not everyone is Nolan Ryan, but we'll never know who is a bit like him with the current limitations placed on pitchers. The limitations are arbitrary. If anything, they harm pitchers, given the higher injury rates.
Some baseball franchise should be bold, buck modern trends, and train all their pitchers to prepare for 120-pitch starts. I know that 120 is as arbitrary as 100, but here's my thinking: It seems that the greatest injury risks come from going farther than what the arm has trained for. Most starts will still last 100-110 pitches, before the game is over or the pitcher is ineffective.** Training for 120 pitches should make a normal game feel easier. More importantly, if a pitcher is performing well on a particular night, they wouldn't be dangerously overexerting themselves until 135 pitches or so. That's a massive amount, which should be more than enough to cover most games.
**Then again, maybe a pitcher won't get tired around 100 pitches if they have trained for more than 100 pitches. Who knows?
I am surprised that pitch counts haven't been investigated much, given how data-driven baseball has become. It's hard to argue that pitch counts do anything to improve a player's health. As long as most teams adhere to current pitching practices, there is an inefficiency for some forward-thinking team to take advantage of. Who will be the team that gets a step ahead of everyone else?