It is entirely possible and entirely practical, in modern baseball, for a team to use a three-man starting rotation.   I realize that this is probably not going to happen, but. . this is how it could work, and this is why it would work.

            Suppose that a team used a three-man starting rotation, but limited each pitcher to 80 pitches a start or five innings.   (This actually would work with 90 pitches a start, but 80 is more conservative, so I’m going to use 80 as a working premise.)  Anyway, a starting pitcher always and absolutely comes out of the game as soon as

            1) He has pitched five innings, or

            2)  He has thrown 80 pitches.

            No exceptions.   Eighty pitches, it’s the fifth inning, you’re ahead 9-0 and you have two outs and two strikes on the hitter. ..tough luck, Sally, you should have thrown more strikes earlier in the game.   

            OK, let’s assume that.   The pitcher pitches no more than 5 innings in any start, and throws no more than 80 pitches.  How much could such a pitcher pitch, in a season?

            In theory, such a pitcher could make as many as 54 starts in a season and pitch as many as 270 innings, but in practice this would never happen.    In order to reach 54 starts the pitcher would have to stay on the three-game schedule all year, and it’s difficult to imagine that happening.    The schedule would have to run perfectly, no rainouts and double headers, no disruptions from days off and the All Star break.   The pitcher would have to be 100% healthy all year, no stitch in the side, no flu or food poisoning, no blisters.   It is hard to imagine all of that happening.

            Since the pitcher will never pitch more than five innings in a start, in order to average 5 innings a start, he has to pitch five innings in every start.   Realistically, that is never going to happen.  Most pitchers are going to have a good many starts or many starts in which they fail to get through five innings in 80 pitchers, and every pitcher is going to have at least some starts like that.   Most pitchers are going to give up 7 runs in the first three innings and get knocked out of the game once or twice a year, certainly given 50-some starts.   Realistically, a pitcher working in a three-man rotation would have an upper boundary of about 52 starts, 245 innings in a season.  

            That’s not a fantastic number, is it, 245 innings?    116 pitchers pitched 245 or more innings in a season in the 1950s, 189 in the 1960s, and 275 in the 1970s.    This went down to 134 (pitchers pitching 245 innings) in the 1980s, 46 in the 1990s, and 14 in the first decade of this century.  

            The thing is, though, that this decrease in pitchers pitching 245 or more innings has not been accompanied by any obvious decrease in starting pitcher injuries.  

            Study here. . .this is a new study; I have done ones like it before.   Not only has this decrease in innings pitched by starting pitchers not led to an obvious improvement in pitchers’ health, but the number of pitchers who are able to stay in rotation from season to season has actually decreased measurably.  

            If you look at the 100 pitchers pitching the most innings in the years 1970 to 1975, those 100 pitchers pitched an average of 297 innings each—what seems to us now like a fantastic workload.   Yet, despite pitching 297 innings each, 96 of those 100 pitchers were still in the starting rotation (pitching 150 or more innings) in the following season, 87 in the second following season, and 76 in the third following season.   68 of the 100 were still in the starting rotation four years later, 61 were still in rotation five years later, 45 were still in rotation six years later, and 37 were still in rotation seven years later.   (The six and seven-year figures are slightly depressed by the 1981 strike, and should be slightly higher, but never mind.)

            If you look at the 100 pitchers pitching the most innings in the years 2000 to 2005, those pitchers pitched an average of only 231 innings—yet the numbers of those pitchers staying in rotation one year later, two years later, three years later, etc. are actually much lower than in the parallel study for 1970-1975:

            I understand, of course, that many things have changed in this time period, other than just the workloads of the starting pitchers, but there is no question that there is an increase in mid-career injuries to good starting pitchers in this time period (Dontrelle Willis, Brandon Webb, Mark Mulder, Johan Santana, Carl Pavano) rather than a decrease.   

            The decrease in starting pitcher workloads has occurred in three stages.    First, in the years 1974 to 1985, teams changed from four-man to five-man starting rotations.   (The Cardinals were using a five-man starting rotation in the late 1960s, but this did not start the trend.  The trend started in the mid-1970s). 

            Second, pitch limits were put into place in the mid- and late 1980s, when STATS began counting the number of pitches in every game and putting this in their box scores.   There was a period about 1990 when virtually every box score published in any newspaper came directly or indirectly from STATS (perhaps literally every box score.   We were not aware of any exception.    Some newspapers were purchasing their box scores from other sources, but those sources were getting their information from STATS.  Certainly all of the wire services were getting their data from us.   Some local newspaper may have persisted in creating their own box scores.)  STATS made pitch counts available as a part of their package, and many newspapers published them.  

            Managers would be criticized whenever pitchers had very high pitch counts, and this led to the adoption of pitch limits by starting pitchers.    Craig Wright’s research on pitcher injuries was critical to this process.

            Third, from 1990 to approximately 2008, perhaps continuing to the present day, pitch limits were walked backward to lower and lower numbers.    I don’t have the book in front of me, but my memory is that Craig’s 1987 book, The Diamond Appraised, argued that pitchers younger than 25 years of age should never throw more than 130 pitches in a game.   Other researchers in that era, picking up on Craig’s work, started figuring "pitcher abuse points" for longer starts, and this, along with expanding bullpens, caused the pitch limits to shrink from 130 to essentially 100 pitches.   The outrage of fans after Grady Little allowed an obviously tired Pedro Martinez to blow a lead in the 2003 ALCS also contributed to that process.  

            So three stages:

            1)  Switch from the four-man to five-man rotation (1974 to 1985),

            2)  Imposition of starting pitcher limits (1986 to 1991),

            3)  Pitch limits walked backward (1990 to 2008). 

            All of this was done in an effort to reduce injuries to starting pitchers, and not for any other reason.    Some people will argue NOW that this may not have reduced injuries, but it may have made starting pitchers more effective, but that’s an after-the-fact argument.    Doing these things has not, in fact, reduced starting pitcher injuries—and there is no actual evidence that a pitcher could not pitch 280, 300 innings in modern baseball without any increased risk of injuries.  

            There is evidence, and there is what we choose to believe based on the evidence.   This is what I believe.   I believe that the first stage of this process, the switch from four-man to five-man starting rotations, did nothing at all to reduce injuries, and that there was no reason at all to make that switch.  Managers were reacting to a fear of injuring their pitchers by overworking them, but the first step they took to protect their pitchers had no protective value.   I believe that the second change (putting in place pitch limits) was absolutely necessary and appropriate, and I believe that the third change (reducing the pitch limits) may have been helpful in some cases.   

            And. . .what about the improvements in sports medicine and training that we hear so much about?   If sports medicine and training have improved so much, and pitchers are being handled so much more carefully than they were back in the bad old days of Tom Seaver and Nolan Ryan, why then have starting pitcher injuries not decreased?  Doesn’t it seem obvious that they should have?

            Pitcher injuries have not decreased because we’re pursuing a chimera, a shibboleth, a mirage.    The way we use starting pitchers now is NOT, in fact, an effective way to keep them healthy and in rotation.    That is what I believe.

            But I am not arguing here for a return to pitchers pitching 300 innings in a season.   I’m arguing for a change in pattern that would put the top limit at about 245 innings in a season—actually not that far from where it is now.   

            But here’s the thing:  starting pitchers would be pitching a few more innings, but in a pattern that reduces the stress load associated with those pitches.  

            It is universally accepted, I believe, that pitches thrown when a pitcher is tired are more likely to contribute to injury than pitches thrown when a pitcher is less tired.     Which exposes a pitcher to more injury risk:  throwing 600 pitches in four starts (150, 150, 150, 150) or throwing 600 pitches in six starts (100, 100, 100, 100, 100, 100).   Of course we believe that throwing 150 pitches in a start is more than proportionally stressful than throwing 100 pitches in a start, given that the number of pitches is equal.   That assumption is the foundation of the current structure, in which pitchers are taken out of the game at about the 100-pitch limit.

            Does it not follow, then, that a pitcher can pitch more innings in a season if the number of "fatigue pitches" is reduced?    Do you not believe that a pitcher has more fatigue in his arm at 100 pitches (heading into his last ten) than he would at 70 (heading into his last ten)?   If you do not believe that, why would you not believe that?  

            And if you believe that a pitcher has more fatigue at 100 pitches than he does at 70, then does it not follow that a pitcher can pitch more innings in a season if he never throws more than 70-80 pitches in a game, than if he throws 100 to 110 pitches in a game?  

            A pitcher working in a three-man starting rotation can pitch (a few) more innings in a season than pitchers now pitch, with no increase (and probably a decrease) in the risk of injury.  I know that I’m out on an island here all by myself, but this is what I believe. 

            And, I will argue, there are other benefits that would flow to anyone who was bold enough to actually do this, other reasons why it is practical to do this.   First point:  the starting pitchers would come to love it. 

            It has been talked about forever, why not do away with starting pitchers, and just use each pitcher three innings at a time; it has been talked about forever, and actually tried a couple of times.   But one reason it does not work is that starting pitchers feel dispossessed.   The rules require you to pitch five innings to be credited with a win, but you can get a loss facing one batter.   Who wants that burden every time you go to the mound?   You make 70 starts of two innings each, you can finish the season 0-27 with a 2.50 ERA.   

            But in this system, a starting pitcher would have a relatively open pathway to twenty wins—and an outside possibility of winning thirty.   In 1974, when starting pitchers pitched more innings per game, starting pitchers accounted for 75% of pitcher wins; now, it is 70%.   With this structure we have to assume that that number would drop back a little bit further, but starting pitchers would still get the win in about 67% of games.    An average team wins 81 games, so that’s 54 wins for the starting pitchers on an average team.   There are three of them, so that’s 18 wins apiece—on an average team.

            Maybe not quite, but in this system there would be a great many pitchers who would go 16-17, 17-19, and 18-17.   The good pitchers are going to go 20-13, 21-11 and 22-14—records like Warren Spahn had every year.   And a good pitcher on a great team. . .well, the sky’s the limit.  

            Another advantage to this method would be that it structures the bullpen in such a way that one or two spots could be reclaimed for the bench.   In all of baseball history up to 1990 there were only 11 pitchers who pitched 50 or more games and averaged less than one inning per game.  Now there are about 100 such pitchers per season, or three per team.

            This is not ENTIRELY necessary; a few of those pitchers could carry some actual innings.   With a three-man starting rotation, you could structure the staff this way:

            Starting Pitchers (3)

            Closer (1)

            8^th Inning guy (1)

            6^th-7^th nning pitchers (3)

            Left-handers (2)

            That’s an 11-man staff, and I think that’s enough.    If the starting pitchers pitch a little less than five innings per start that be 760 innings per season, which leaves about 700 for the bullpen.  With eight relievers that’s 88 innings per reliever.   That doesn’t strike me as an extraordinary number. 

            One other advantage that I see to this method is that a starting pitcher taking the mound knowing that he has only 80 pitches to work with and that he has to get through five innings is well positioned to understand the importance of attacking the hitter.   You can’t really afford to waste pitches—and I think that having a deeper understanding of that principle would work to the benefit of the game and of many of the individual pitchers. 

            Well. . .this is just my idea.   History is like a river; it chooses a course, and it digs into that pathway until it is so deeply entrenched that it seems impossible to move it.   I’m not expecting to move the river here; I am merely arguing that it could equally well have taken a different path.