To what extent is "Power" a function of youth?

            To some extent, obviously, but not to a great extent.   I looked at 16 pitchers who had careers of 350 or more starts, and for each of those I counted how many "power starts" he had over the first half of his career, and how many he had over the second half of his career.  

            Three of the 16 pitchers actually had more power starts over the second halves of their careers than over the first halves; those three were Steve Carlton, Bret Saberhagen and Roger Clemens.  

Two pitchers had dramatically fewer power starts over the second halves of their careers; those two were Tommy John and Catfish Hunter.    John had 98 power starts over the first half of his career, 37 over the second half; Hunter was at 95 and 43.  

The other eleven pitchers had moderate declines in the number of power starts in the second halves of their careers.   Overall, the pitchers had 82% as many power starts in the second halves of their careers as in the first halves.  

            A "power start" was defined as any start with a Power Score of 50 or higher.   For pitchers who had odd numbers of career starts, I counted the middle start as being part of both halves, the first half and the second.    Gibson’s totals don’t match his career starts because a few of his early starts weren’t included in the Retrosheet data that I was using.

To what extent is "Power" predictive of future performance?

            In this study as in many previous studies, Power tendency is certainly a predictor of future performance.

            I took all pitchers in the study, and figured for each his career starts at the end of the season.   From that data, I sorted out all pitchers who had 100 to 124 career starts at the end of a season, and who had 25 or more starts in the season in question.    This gave us a group of 424 starting pitchers.  From this group I eliminated those whose base season was 2001 or later. 

            From the remaining pitchers I sorted out the 100 pitchers who had the highest percentages of Power Starts, and the 100 who had the lowest percentages.

            There is little quality difference between the two groups in the base season or in their careers up to that point; actually, the pitchers with fewer Power Starts early in their careers did slightly better, early in their careers, than the pitchers with more Power Starts.  However, in the remainder of their careers, the "Power" pitchers made an average of 199 more starts.  The "Control" pitchers made an average of 131 more starts.

Of what actual use in the Power Game Score?

            Well. . ..I don’t have a great record of understanding what the potential of my ideas might be.   For years, I used to project next-season stats for every major league hitter—and never even printed it, because I didn’t think the process had any value.    Once John Dewan talked me into printing the projections numerous other parties also stepped forward with their own projections, and projecting player stats entered the sabermetric mainstream.   In retrospect, I don’t know what I was thinking.    I would assume that every major league team now routinely attempts to project the next-season performance of their players, because if you don’t, how do you know what you have?

            I’ve had other ideas that I thought were tremendously valuable that no one ever paid any attention to, and which led to nothing.   I’m not the best judge—nor, for that matter, is anybody else.  The Roman poet Lucretius anticipated atomic theory, predicted the existence of germs and outlined the theory of evolution—and no one paid any attention to him whatsoever for fifteen centuries after his death.

            Anyway, I don’t see Power Scores for games as being a major tool; I see it more as kind of a fun thing to play around with.   Power Scores might help us, for example, to evaluate young pitchers more quickly.   Since we know that extreme power pitchers have Power Scores of 50 or above in 80% and 80%+ of their starts, we might be able to move more quickly in "typing" young pitchers.

            It is possible that Power pitchers are more effective against certain types of teams, control pitchers more effective against others.   We have here a tool to draw lines between them.

            If pitchers make transitions from Power to Control in mid-career, this method might help us to spot those transitions more quickly and more reliably, thus might help us in projecting those pitchers.

            I have made several not-very-successful efforts to outline how we can form "families" of pitchers, on the theory that members of a family of pitchers might have similar career tracks.   Power Game Scores could be useful in identifying to which family a pitcher belongs, should a more successful family-typing system be developed.

            It could be that changes in power type are predictive of other changes in a pitcher’s career.

            But basically, I just kind of like the system, because I like knowing what pigeon hole to put things in and what peg to hang them on.  Knowing that Nolan Ryan had Power Games in 85% of his career starts, Lary Sorensen in only 7%. . ..this seems to me to be worth knowing in and of itself.

            One more thing before we go.   Let’s choose the major league Power Pitcher of the Year for each year from 1955 to the present:

            Remember; it’s not the best power pitcher; it is the "powerest" pitcher—the pitcher with the highest percentage of Power Starts.   And, to balance the scales, the control pitcher of the year: