***DISCLAIMER: This one is quite a long story.***
Second, I'm hoping to hear some comments one way or another on this one merely because I'm using stats to back an opinion without any true factual evidence, so I will be curious to see how many agree or disagree.
As mentioned, Phil Humber of the White Sox threw a perfect game this past Saturday against the Mariners. It was merely the 21st perfect game in recorded history at the Major League level, and only the 19th of the modern era (post 1900). First, tons of credit to Humber, no matter how you twist it, the outcome is an accomplishment regardless.
However, I believe it is only fair to ask how the game is changing so much that perfect games are occurring at a very increased rate. Of the 21 perfect games, 11 of them have occurred in the last 28 years. The first ten came over a span of 101 years.
Do pitchers today have better stuff? Can fielders cover a better range of the field? Stadium conditions? Umpiring calling a wider strike zone?
My hypothesis is that umpires are playing the greatest role in the increased frequency of the perfect game. Sadly, there is no way to really measure that unless you meticulously watch every pitch of every game, perfect game or not. What we can do is look at a few details not to determine, but to get an idea how just how an individual pitcher was for a particular game.
These numbers don't exist for all box scores, but we can look at the last ten perfect games (dating back to Tim Browning's in 1988) plus the infamous imperfect game of Armando Galarraga in 2010 in greater detail including the number of pitches, strikes, strikes by contact/swinging/looking, ground/fly balls, etc. Note for Galarraga's game, all the numbers have been adjusted as if a perfect game occurred (essentially exclude the last batter of the game). Again, I point out these don't provide true definitive evidence of anything, but I used these numbers to try some conclusions about these games.
I came up with an equation that I think pinpoints what percentage of pitches fell on the pitcher/infield and batter only. That said, there are a few assumptions being made here, so bare with me, but feel free to point out if you disagree.
I say pitcher/infield because I count every ground ball out as a successful out on the part of the pitcher. Of course, a diving stop on a ball up the middle is really the work of the infield, but again, I'm making an assumption here. On the flip side, I assume every fly ball out isn't the work of the pitcher, but rather the work of the fielders. This is based on the idea that pitchers rarely try to get a batter to fly out.
Next, I consider it a success when a pitcher throws a strike that the batter swings and misses at, or fouls off. That isn't to say a pitcher shouldn't be credited for when a batter is caught looking, but the outcome of such a pitch falls on the umpire making the call where as a swing and miss for foul ball fall on the pitcher/batter only. The resulting equation is as follows:
% Pitcher is Solely Responsible = (# of P. - # of B. - # of S.L. - # of F.B.) / # of P.
P. = Pitches
B. = Balls
S.L. = Strikes Looking
F.B. = Fly Ball Outs
It is far from perfect, but it should give a far better idea of how much of the game a pitcher was in control of his own fate. The results of our 11 games provide the following percentages:
- R. Johnson (04) - .556
- D. Martinez (91) - .490
- T. Browning (88) - .480
- D. Wells (98) - .408
- D. Cone (99) - .386
- A. Galarraga (10) - .386
- M. Buehrle (09) - .379
- D. Braden (10) - .349
- R. Halladay (10) - .330
- P. Humber (12) - .323
- K. Rogers (94) - .306
Of these 11, Randy Johnson has the highest percentage, and only one over 50%. This makes sense when you realize he only allowed 7 fly outs (less than everyone else except Dennis Martinez), and had far more swinging strikes than everyone else (26) due to his over powering fastball. That is to say, Johnson left half the game up to the umpires, outfield, and luck in determining where fly outs could land for a hit. A perfect score of 100% would require only strikeouts and ground outs for the entire game without a single strike looking or ball being called.
On the flip side, the lowest percentage belongs to Kenny Rogers. He allowed 12 fly outs compared to only 7 ground outs, and had only 10 batters swing and miss and 13 foul balls compared to 22 strikes looking. It could be argued that Rogers was just painting the corners perfectly that day, and for all I know, he was, but by doing so, he was leaving the game in the hands of the umpire.
Looking at that further, the four perfect games (including Galarraga's) between 2009-10 saw a combined 23 pitches where the batter swung and missed. I find it pretty astounding that in a game where no batter could reach base, that only 23 pitches were whiffed at. Meanwhile, those four games had a combined 97 strikes looking, and those four games rank in the top five of these eleven for strikes looking. Is it purely coincidental that the five most recent perfect (and imperfect) games had the most strikes looking of those on record? Is it far fetched to believe that previous perfect games of which there are no such records, could have had less than these five most recent ones? Note that aside from Rogers' low score, the five lowest are the five most recent games.
Assuming you're still reading this, I put together one other stat to define how much control these pitchers had in each game. In a perfect game over nine innings, a pitcher will face 27 batters. Based on the same assumptions I made above, we credit a pitcher for ground outs and strikeouts only. However, since the whole debate is whether recently pitchers are getting more calls in their favor, we have to weight strikeouts. For this, I used the following equation:
% Pitcher is Solely Responsible = (((S.M. + Foul) / (S.M.+ S.L. + Foul)) * K + G.B.) / 27
S.M. = Strikes Swinging and Missing.
S.L. = Strikes Looking
Foul = Foul Balls
K = Strikeouts
G.B. = Ground Outs
I don't believe this equation works quite as well as it puts more emphasis on the infield doing their job well, but it does promote the pitchers who kept the ball on the ground and out of the outfield which really are the best ingredients for a perfect game.
- D. Martinez (91) - .756
- R. Johnson (04) - .642
- A. Galarraga (10) - .604
- T. Browning (88) - .560
- M. Buehrle (09) - .541
- R. Halladay (10) - .516
- D. Wells (98) - .500
- K. Rogers (94) - .411
- D. Braden (10) - .382
- D. Cone (99) - .366
- P. Humber (12) - .362
The main differences between the two rankings is that Dennis Martinez jumps Randy Johnson for first, largely due to the fact that 17 of his 27 outs were ground outs. Meanwhile, Buehrle, Halladay, and especially Galarraga climb the list significantly because they created as many ground outs as fly outs if not more. The four lowest on this list all had at least 12 fly outs which puts more pressure on the outfielders to cover ground, and rely on a bit of luck that a ball did get hit into the gap or fall for a blooper.
As fans, most of us enjoy watching a perfect game, even when it isn't our team. We might root for the Dallas Branden's and Phil Humber's of the world, but at what point do the begin to devalue the perfect game? Were we as excited by Humber's a few days ago when we witnessed four others only a couple years ago, and a half dozen more in the couple decades before that? Do we really appreciate the feat that occurred ten times in a 101 year span prior to this recent outbreak? Moreover, are changes in the game not causing, but maybe assisting some of these outcomes?
I may have to try this technique on no-hitters in the future. In the mean time, congratulations to Andrea Chiaradio and Phil Humber, even if Humber was only in control of his destiny 32% of the time.
Interesting argument you present here. Statistically speaking, you may be on to something here. Personally I think the increase in perfect games over recent years has more to do with the 5-man rotation, as well as improved strength and conditioning. Having an extra day's rest (or extra day of being a gym rat for the Roy Halladay's of the world) allows for a pitcher to be better prepared for his start and in turn, have a more likely chance of being more effective and last longer for said start. Sure pitchers in the old days threw more complete games, but that was more or less expected of them, whether they had good stuff or not. Improved strength and conditioning ties in with this in that there is a much larger knowledge base by team trainer's, and individual pitcher's, as to how best to treat a pitcher during his off day's and prepare them for their next start. So my argument is that the reasoning(s) for an increased percentage of perfect games in recent years has more to do with qualitative reasons, and not quantitative. That being said, I think your research make a case for a legitimate statistical argument. my only real disagreement is the umpire hypothesis. Every umpire has their own strike zone and I don't think there has been any drastic differences through the years. If anything, the strike zone is smaller than the old days, because the high strike isn't called very often.
ReplyDeleteVery good research, I enjoyed it.
Thanks for the comment Reece! I do agree that added rest and improve strength and conditioning plays a larger role. It is certainly a number of factors. Defense has improved drastically over the past few decades, and that easily plays into half of a perfect game as noted by the numbers above.
ReplyDeleteThat said, your comment caused me to look at something else I hadn't before and that was age. Humber at age 29 and 4 months is the second youngest pitcher to throw a perfect game since Tom Browning in 1988 (Dallas Braden being the youngest). Almost everyone since has been 30 or older. Now if you look prior to Browning, only two pitchers were over 30 (Cy Young and Jim Bunning). Therefore, that certainly proves the training is playing a big role.
I did a bad job with the premise of the umpire comment. I agree the strike zone has diminished in recent decades. I am curious if an umpire who is inconsistent for a game is the driving force behind it. Maybe the strike zone is a bit larger for a particular day (inadvertently of course).
Pitch counts are generally higher now than the more efficient perfect games of the past. My guess is that umpires, even if just for an individual game play the largest role on that. Overall, unless you're Randy Johnson, the pitcher controlling his own fate I think has gone down one way or another in the last ten years.