On Matt Olson’s Strikeout Rate
This is just a little study, took me 15 minutes to do the study but will probably take me an hour to write it up. I had a question in "Hey, Bill" this morning about whether Matt Olson’s 2021 decline in strikeout rate is predictive of a lower strikeout rate in the future. The answer, not to keep you in suspense, is that it definitely, definitely, definitely IS. In a study of other players in history who have had large declines in strikeout rate in season X, more than 90% had lower strikeout rates after season X than before season X. We just don’t see numbers like that; "more than 90%" just doesn’t happen. But it did.
OK, here’s what I did. I took every player’s strikeout rate through the end of the previous season (Season X – 1), and his at bats in the focus season (Season X). Matt Olson through 2020 had struck out 442 times in 1,483 at bats, or 29.8% of at bats. He was almost a .300 striker-outer, the historic standard of a quality striker-outer. Given 565 at bats in 2021, he could have been expected to strike out 168 times. He actually struck out only 113 times, or 55 times less than expected.
This is a fairly historic number; it is about the 8th highest number of all time. It’s in the top 10. Anyway, having figured this number for all players, I eliminated from the study any player who (a) had played less than 350 games prior to Season X, (b) played in the 19th century or the years in the early 20th century when strikeout data is missing, or (c) played in 2019 or later, since that would not give us a true read on their "subsequent" strikeout level.
The LARGEST decline in strikeouts vs. previous expectation of all time was 65, by Ed Brinkman in 1969. Brinkman prior to 1969 had 2,644 at bats and had struck out 490 times, so, given his 576 at bats in 1969, he could have been expected to strike out 107 times. He actually struck out only 42 times, or 65 times fewer than expected. This is the largest number of all time.
And you know who had the SECOND-largest decline ever?
Ed Brinkman in 1970. Brinkman’s low strikeout rate in 1969 reduced his "prior strikeout rate" number from .1853 to .1652, but even so, Brinkman had 625 at bats in 1970, so he could have been expected to strike out 103 times. He struck out only 41 times, or 62 fewer than expected.
We could safely say, then, that Brinkman DID follow through on his lower strikeout rate from 1970. Having previously struck out 105 times (1966), plus 99 times in just 447 at bats (1964) and 86 and 82 times in other seasons, he never again struck out more than 79 times, despite generally increased playing time.
Many of you will immediately remember the unusual circumstances of Brinkman in 1969. The mound was lowered in 1969, so strikeouts were down all around the American League, and Brinkman’s team that year hired one Ted Williams to be their manager, giving Brinkman access to an unusual resource for hitters. Some hitters benefitted from that, some didn’t. Brinkman did.
But before you pooh-pooh Brinkman’s achievement in cutting his strikeouts by 65 in 1969, consider this: Brinkman HIMSELF accounted for about 10% of the major league decline in strikeouts per at bat in 1969. The decline in strikeouts in 1969 was 60 per team in the American League, or about 720 for the league. Brinkman declined by 65, or about 9% of the league total, and Brinkman’s decline was larger than the average decline PER TEAM. And in the National League in 1969, strikeouts per team actually INCREASED. They increased by 19 per team, from 950 to 969. So. . .the individual elements of that are quite a lot larger than the league-wide elements.
The largest decline in strikeouts by someone not named "Ed Brinkman" was by Larry Hisle in 1976. Hisle also completely followed through on his decline in strikeouts. Having previously struck out 162, 139, 128 and 112 times in a season, Hisle had only one more season over 100 strikeouts, that being 106, and Hisle actually had his best seasons not in 1976, when his large drop in strikeouts occurred, but in 1977 and 1978. His big decline in strikeouts PREDICTED or Led or Anticipated his improvement in performance.
Other players with very large declines in strikeouts in a season include Julio Lugo (2005), George Scott (1969), Jeff Conine (1999), Matt Williams (1993), Cecil Cooper (1980), Carlos Delgado (2000) and The Big Hurt, Frank Thomas, in 1993. Sometimes a decline in strikeouts seems to come ahead of subsequent improvement in performance, and sometimes an older player, realizing that he can no longer reach the seats, cuts down his swing. Butch Hobson had a big decline in strikeouts in 1979, and his average did increase, but that didn’t foretell great things for him. Dave Kingman in 1985 and Mike Schmidt in 1986 had big declines in strikeouts as they neared the end of their careers.
But in the aggregate. . . there were 126 players in my study who cut their strikeouts by 35 or more, based on their previous strikeout rates. Of those 126 players, 116 had lower strikeout rates AFTER year X than before year X. We’re not including the focus year in the "rest of career" number; the focus year, Year X, is a thing in itself, not part of the "before" or "after" strikeout rate. The "after" strikeout rate is almost always lower.
The average strikeout rate was .211 BEFORE the focus season, .132 during the focus year, and .170 after the focus season.
So I learned three things of value by doing this study, (1) that large one-year declines in strikeout rates are certainly predictive, (2) that Ed Brinkman in 1969-1970 is truly a singular event in baseball history, and (3) that the National League strikeout rate actually increased in 1969 although the mound was lowered. I did not know that, and I am very surprised to learn that.
I am unable to get on my own site as a reader to post comments.. . .software issue. . .so I will respond to a couple of questions in this space. First, one reader stated that some of the players in the study only had small declines in strikeouts of 10 or so. This is COMPLETELY false; all of the players in the study struck out at least 35 times fewer than than would have been expected based on previous career data, as is stated in the article. Also, if you think about it. . .if there were players in the study who had declines in strikeouts of only 10, we obviously would not get a persistance rate of 90+%.
In regard to the request for additional data from Tom Tango, I regret to tell you that I dumped the data after running the study. It was just a real quick study, just took me a few minutes as mentioned above, and I never created a unique data file for it; rather, I just used one of my existing data files, and then I didn't save the data because I didn't want to mess up the existing data file. Stupid: I should have saved it. I did figure the average at bats in the focus season, which was 570 or 580, something like that, and in the previous career, which I think was about 2300, but I'm sorry to say that I didn't actually save it.