I found a straightforward way to redo the normalized strikeout rate study. Check out my previous post on the subject for the rationale.
For those who care, here is the updated method:
- Instead of taking a single league-average K/9 rate, I took 5 numbers for each player’s career–the league average from his first season, his final season, as well as the points 1/4, 1/2, and 3/4 through his career. I then averaged those numbers and used that as the league average for that given player. This is of course not entirely precise, but is really a very good approximation.
- I made a lookup table in Excel, which made this exercise a lot easier. I then just listed one set of numbers for the league-average K/9 rate in each season, and then for each player did 5 lookups to get his 5 numbers.
- By using this method, it was easy to put the top 200 pitchers (by innings pitched) into the table. I could easily do more, as well.
Click through for the results.
Here are the top 25 pitchers all time in normalized K/9 (among the top 200 in innings pitched):
Rank Player K/9 Avg K/9 Normalized K/9 1 Rube Waddell 7.04 3.22 2.19 2 Dazzy Vance 6.20 3.22 1.93 3 Nolan Ryan 9.55 5.58 1.71 4 Randy Johnson 10.61 6.28 1.69 5 Lefty Grove 5.18 3.22 1.61 6 Walter Johnson 5.34 3.34 1.60 7 Bob Feller 6.07 3.88 1.56 8 Pedro Martinez 10.04 6.44 1.56 9 Amos Rusie 4.64 2.98 1.56 10 Tommy Bridges 5.33 3.52 1.51 11 Chief Bender 5.10 3.44 1.48 12 Hal Newhouser 5.40 3.76 1.44 13 Bill Donovan 4.71 3.30 1.43 14 Ed Walsh 5.27 3.74 1.41 15 Bobo Newsom 4.98 3.54 1.41 16 Roger Clemens 8.55 6.16 1.39 17 Curt Schilling 8.60 6.22 1.38 18 Jack Stivetts 3.81 2.76 1.38 19 Tim Keefe 4.57 3.34 1.37 20 Mark Baldwin 4.33 3.18 1.36 21 David Cone 8.28 6.16 1.34 22 Tom Hughes 4.66 3.50 1.33 23 Christy Mathewson 4.71 3.54 1.33 24 Red Ames 4.79 3.62 1.32 25 Bob Gibson 7.22 5.46 1.32
There aren’t a lot of surprises here, except maybe that Rube Waddell and Dazzy Vance are way out in the lead.
There are only 10 pitchers who were at least 50% above league average for their careers–and as strikeouts get more frequent, that’s getting harder and harder to do. It seems unlikely we’ll see another pitcher join the upper ranks of this list any time soon.
Now here are the bottom 25 guys–remember, though, that these pitchers would good enough to make the top 200 list for most innings pitched, so they couldn’t have been horrible–they just got it done a little bit differently.
Rank Player K/9 Avg K/9 Normalized K/9 1 Lew Burdette 3.15 4.92 0.64 2 Frank Dwyer 1.80 2.80 0.64 3 Tom Zachary 2.07 3.02 0.69 4 Ted Lyons 2.32 3.38 0.69 5 Bob Forsch 3.65 5.18 0.70 6 Bob Caruthers 2.86 4.02 0.71 7 Al Spalding 0.77 1.04 0.74 8 Slim Sallee 2.67 3.60 0.74 9 Vern Law 3.68 4.92 0.75 10 George Bradley 2.05 2.72 0.75 11 Freddie Fitzsimmons 2.43 3.22 0.75 12 Mel Stottlemyre 4.25 5.58 0.76 13 Jim Slaton 3.99 5.22 0.76 14 Clark Griffith 2.54 3.30 0.77 15 Doyle Alexander 4.08 5.24 0.78 16 Mike Torrez 4.15 5.32 0.78 17 Claude Osteen 4.19 5.34 0.78 18 Mike Morgan 4.55 5.78 0.79 19 Jim Perry 4.32 5.46 0.79 20 Tommy John 4.29 5.40 0.79 21 Joe Niekro 4.39 5.48 0.80 22 Danny MacFayden 2.65 3.30 0.80 23 Al Orth 2.54 3.10 0.82 24 Bert Cunningham 2.36 2.84 0.83 25 Eppa Rixey 2.70 3.24 0.83
This group is also mostly pitchers from long ago, which surprised me a bit since, as I mentioned above, it’s easier to be below average now since the average is so much higher. (Compare, for example, Bob Forsch and Al Spaulding, who have nearly identical normalized K rates despite a 5-fold difference in their raw K rates.)
I guess the main issue is that there’s a smaller spread in talent these days. 100 years ago, pitchers used a lot of different styles. These days, just about all pitchers have a power arm to some degree and the strikeout is an essential part of successful pitching. It could also be that teams these days select for pitchers with good K-rates, knowing that a guy who can’t consistently strike out batters doesn’t have a chance at long-term success. This may result in a smaller standard deviation of K rates around the average today.
Anyway, as reader kds suggested, I will re-run this analysis using strikeouts per plate appearance to see how much difference that makes.
If anybody wants the numbers for a specific pitcher, let me know and I’ll pull them off my spreadsheet.
I guess the thing I would really like to see is how many standard deviations above/below the mean these pitchers were for their time. I think that would, in many ways, be more informative – simply because Rube Waddell is #1 on this list, but his K/9 is barely above the league average in Randy Johnson’s time. While he was over 100% above the mean, I don’t think we can directly say, “Had he played in Randy Johnson’s time, he would have had 13.7 K/9.” While standard deviations wouldn’t allow us that kind of certainty, either, I think they would come closer. But, they’re way more work, so it’s probably not worth it. It’s just a frivolity, after all.
I wonder what the highest individual season is for normalized K rate? Waddell in 1902 had a 6.8 K/9 in a league context of 2.5 K/9, which is an astounding normalized rate of 2.72.
This is a very interesting study. I appreciate the work that goes into something like this. Like Dr. Doom said, I’d also like to see the standard deviations. I believe pitchers from earlier eras have a bit of an advantage without them. That probably is true for a lot of historical comparisons. I think there used to be a wider distribution of talent in all areas in the early days.
Standard deviation is not too difficult to determine, I think, if I look at a specific set of years (such as 1900, 1910, 1920, etc). The mere fact of more players nowadays will tend to push SD up but we can account for that. I’ll look into doing this study.
Actually Andy, it’s possible that more players will tend to push SD down, as Lawrence sort of hints at below (I actually teach a stats class at university and use that Steven Jay Gould clip about the .400 hitter in the standard deviation class). More spread of quality will push SD up, as it was 100 years ago, because the deviations from highest to lowest quality player are big with a less robust farm/development system (hence .400 hitting, or maybe much higher than average k/9 rates). But more players, in the modern day, means more competition, and training methods have become so good that the spread of talent isn’t as big as it was 100 years ago. With not only integration but also big money incentive, and recruitment from central America, etc, I’d say the worst baseball players are alot closer to average today than 100 years ago, even though there are more of them in the majors. And the best baseball players are also closer to the average today than 100 years ago, which is why you can’t just say Babe Ruth would hit 200 homers in a season if it were today, or that Rogers Hornsby would hit .424. The level of pitching that Hornsby faced was, on average, worse, which allowed him to hit more consistently. Today he’d probably hit .350 or something.
My point is that the average level of play has gone up, for both hitters and pitchers, meaning that it’s harder to get a strikeout, and it’s harder to get tons of strikeouts because the average competition you’re facing is higher (and on the flipside, it’s harder to get a hit). Randy Johnson probably achieved a bigger feat being 1.69 times better than average in his career time than Rube Waddell did being 2.19 times better.
(of course, since K/9 rates have gone up in general, the standard deviation might go up numerically just by virtue of the fact that you’re dealing with bigger numbers, making things a bit more complicated. But I’d be willing to bet Randy Johnson is more SD above the mean than Rube.)
Steven Jay Gould, the renowned paleontologist, evolutionary biologist, and big baseball fan, theorized that as a particular species matures, evolution reduces the difference of the extremes between a particular characteristic in a species. In the same way that we don’t have animals the size of dinosaurs roaming the land anymore, we no longer have .400 hitters and pitchers throwing 400 innings a year.
I realize that this is a rather, well, strained analogy. Let me express it another way: as the average level of talent goes up, it is more difficult for the very best baseball players to stand out from the rest, as the greats did 75/100/120 years ago. For instance, if Wade Boggs or Tony Gwynn or Rod Carew had their careers before WWII, they would’ve likely to hit .400 in a season, probably more than once. Likewise, Clemens and Greg Maddux and Randy Johnson likely would have won 30+ games.
Baseball is not a one-dimensional sport like track and field, so comparisons can only be made on how players perform against their competition, not on some absolute unchanging scale like the time of a mile run.
I see what you’re getting at. The average talent level does go up, as year-round training and better nutrition lead to superior athleticism, which is then focused on a sport at a younger age, organized and coached more efficiently. You read stories of some older ballplayers discovered on a coal mine team or on some barnstorming tour through the south in winter. Now, promising talent is identified so early, and information about them is so readily available, that such stories are increasingly improbable. The fabled “open tryout” has gone from a romantic, rags-to-riches opportunity to a sign of complete desperation.
I think the whole “Linsanity” deal in basketball stems in part from this trend. This is a guy who played Division I basketball and was in the league last year. It’s not like he just walked in off the streets and started dunking on anything in short pants. Yet this is what passes for “great unknown emerges as star player” now. And the open tryout is now reserved for reality television contests. The odds that a 20-year old bystander who never played organized ball, ever, could improbably find himself noticed by a major league club, tried out, and found to be a quality player, are longer than ever.
“dunking on anything in short pants” just made me LOL and wake up my wife. Thanks a lot for getting me in trouble.
nightfly,
To support your last sentence, Bill James in one of his Historical Abstracts pointed out the difference between MLB at the turn of the twentieth century and now with this anecdote:
Between 1900 and 1905, there were at least three documented cases of a spectator purchasing a ticket for an MLB game, and ending up playing in that game. That would never happen nowadays.
Well, it _couldn’t_ , since their MLB contract would first have to be approved by the Commissioner’s Office, but the larger point that you made is that potential baseball talent is much more efficiently funneled up through the minor league system before it reaches the bigs. 100+ years ago, it really wasn’t unusual for players to be signed directly from sandlots or college teams to MBL rosters and play immediately.
BTW, here’s a Hardball Times article from a while back which covers some similar material: http://www.hardballtimes.com/main/article/strike-zone-dominance-in-context-dazzy-and-pedro/
Might be interesting to compare their Top 10 with the one above.
It also attempts to answer my question above, and rates Vance’s 1924 as the best K season ever in context, right ahead of the aforementioned Waddell 1902 season (min. 200 IP).
Wow. That was a really great article at THT. Very interesting to incorporate BBs, as well. Pretty close to what Andy’s done here, though.
I don’t want to say “I told you so” about Dazzy Vance’s 1924 season being amongst the best strikeout seasons ever.
OK, I admit it, I _do_ want to say “I told you so”: see my comment previously in the post:
Normalized strikeout rates of the all-time greats, comment #21
Where is Jamie Moyer on the list? He’s got a reputation as a low strikeout guy in a high-K era so I’m surprised he’s not on the list for lowest rates
Moyer is actually 29th on the “lowest” list, so good call there. His actual K/9 is 5.38 while his era number is 6.36, so his normalized rate is 0.85. It’s not quite as low as you might think since he played way back into the 80s when K rates were a little lower, and he was not always the soft-tosser that he is now. He also has a wicked changeup that fools a lot of hitters.
I would like to know where on the list of the top 200 one would find the NK9 of 1.00.
There are 2 pitchers with a normalized K/9 of exactly 1: Burt Hooton and Jesse Tannehill. They rank 108/109 in terms of highest normalized K/9. Thus, it would seem that more than half of the top 200 pitchers are above-average, which makes sense since they are the longest-tenured pitchers.
You have the wrong Tom Hughes linked in the article.
I’m guessing you are intending this Tom Hughes.
http://www.baseball-reference.com/players/h/hugheto01.shtml
Weird…Bert Blyleven, Steve Carlton, Tom Seaver, and Fergie Jenkins were ahead of Clemens and Mathewson on the earlier list with 4,000 IP, but they didn’t make it on the top 25 on this list. Could their pitching in both a low strikeout and a high-strikeout era affect their normalized rates that much with the new methodology?
Wait a minute. Jim Slaton is in the top 200 pitchers of all time, ranked by IP? I would not have guessed that.
Great work Andy. Love seeing the old guys and new guys mixed together. May not be exact, but it sure feels right to me. You never hear about Christy as a strikeout guy, but there he is at number 23.
Just noticed that “admin” wrote this post by accident…it was me, and now I can’t change it because I left the editor window open at the office…I’ll fix it tomorrow.
Where does Sandy Koufax rank on here?
Koufax is not among the 200 pitchers with most career IP, actually. I manually checked him, though, and his NK/9 is 1.74, way up high.
Thanks, Andy. Is it in a spreadsheet you could link to? THT article used K/PA but only had data through 2004, so the numbers should be the same for guys who retired before 2005.
OK I re-ran this analysis using K/PA instead of K/9. The results are largely similar, with the vast majority of players sliding up or down just a few spots.
However, there are a few extreme players.
Here are the guys who moved up the list at least 10 spots, meaning guys who didn’t allow a lot of baserunners and therefore had fewer chances to strike batters out (and thus were hurt in the original rankings):
Cy Young, Old Hoss Radbourn, Greg Maddux, Tommy Bond, John Smoltz, Charlie Buffinton, Jim Whitney, Robin Roberts, Babe Adams.
Here are the guys who moved down the list by at least 10 spots, meaning the opposite–these guys allowed a lot of baserunners and had slightly inflated K totals as a result:
Chick Fraser, George Mullin, Rick Sutcliffe, Earl Whitehill, Howard Ehmke, Burleigh Grimes, Bump Hadley
Andy, are you going to do a similar study for relievers? I just want my boy Billy Wagner to have his moment in the sun and let everyone (who doesn’t already) know how historically nasty he really was.
Raphy has actually done a really thorough study–let’s convince him to post it!
Definitely! I was thinking the same thing. I would be really curious to see Wagner’s numbers, in particular.
B-ref has “OPS Against” stats going back to 1950. The best OPS Against career numbers among all pitchers with at least 500 career IPs:
Mariano Rivera .552
Billy Wagner .558
Sandy Koufax .594
Hoyt Wilhelm .595
Frankie Rodriguez .597
J.R. Richard .600
Rivera and Wagner are in a class by themselves here (also in both ERA and ERA+).
Unfortunately, birtelcom, Wagner fell ~100 IP short of qualifying for the all-time ERA+ title.(1000 IP, he had I believe 902).
Wags for the HOF!!! Wait, the poor guy sadly has no chance….****snuffs out Billy Wagner altar candle****.
Not sure if Wags would have had a shot in any case, but his postseason struggles probably sealed his fate.
Yeah, that and retiring at age 38 with a full tank of gas still left. he gave some odd reason like, “I want to see my kids more.” Some people, geez.:-)
Personally, I think the biggest Problem for Billy Wagner is that people seem to measure relief pitchers pretty much strictly by their number of saves. Wagner didn’t accumulate enough, so he’s out. He was behind Hoffman and Rivera in his own time, so there’s no way he’s getting in based on that. It’s a shame, really, because he was such an effective pitcher.
Exactly. No chance. But that’s what I meant about retiring with a full tank of gas—-he hadn’t lost anything, with a 13.5 K/9 and a career-low 1.43 ERA at age 38. Had he wanted, it seems he easily could have pitched as long as L Smith, T Hoffman, etc. and pushed that save total past 500 at least.
I was at Citizens Bank ballpark on 9/7/05 when Biggio went yard off Wagner with a three-run bomb that sealed the game in the top of the ninth for the Astros. I half-jokingly said to my friends at the moment, “there goes the wild card”….sure enough, Astros take the wild card by 1 game and make it all the way to a WS sweep by ChiSox.
But, yeah, Wagner was great and fun to watch. And an absolute talent with plenty left in the tank up to retirement
Mel Stottlemyre: One of the few consistently good Yankees during their 1965-1975 pennant drought. It’s a shame he couldn’t have pitched a little longer.
I wonder what this is going to mean for the future?
I suspect that there is an upper limit to how hard an unassisted human can throw. At least do it over the course of 200 innings per season. And ML batters can catch-up to the hardest throwers these days.
In order to get more Ks, are the next-gen aces going to go down the Pedro mix of large pace variation (“What the heck is he going to throw now?”)? Or are they going to go the Mariano route of one pitch, really well thrown (“Here comes that cutter again!” Or maybe we’ll see someone go back to the past and start throwing junk with a 95mph fastball for variation?
Certain for sure is that pitcher are going to get rated on K/9. Just how are they going to get the Ks when everyone throws pretty much the same thing?
I think SO rates are more dependent on the hitters than the pitchers. Certainly, there’s a much wider range of SO rates among hitters — last year, Drew Stubbs whiffed more than 5 times as often as Juan Pierre, while the range among qualified pitchers was less than 3 times.
SO rates are high now mainly because the majority of hitters try to work deep counts and try to hit for power. If SO rates decline, it will likely be mostly because of a change in that dominant approach.
I’m stunned by this comment because once again you have pointed out something extremely revealing that should have been obvious, i.e. the smaller standard deviation in strikeouts among hitters than pitchers. You’re right that this is strong evidence that the rise in strikeouts has more to do with hitters than pitchers. Time for me to research a post on this.
Bill James made this point years ago. While we tend to think that the outcome of an at-bat is controlled by the pitcher, it’s actually the hitter who has more control. I think the example he used was homeruns. There are players who hit 40 home runs and players who hit 0 homeruns. But there aren’t pitchers who do the same.
This goes also to the point of the style of hitting has changed, and particularly how it has changed since Ruth. Intuitively, the more the batter chokes up, the harder it is to strike him out. The longer you wait on an individual pitch, the more critical velocity and movement are to the outcome, and the more likely the outliers (pitchers with exceptional stuff/speed) are likely to separate themselves.
FWIW, a spread of 40 HRs between qualifying pitchers is not uncommon. Last year, Bronson Arroyo allowed 46, while Charlie Morton yielded just 6 in 172 IP. Arroyo’s HR/9 was about 7 times as high as Morton’s.
Random junk: The lowest qualifying HR/9 in the expansion era is .085 by Dave Righetti in 1981, 1 HR in 105.1 IP.
The HR was hit by Gorman Thomas, who also had 2 doubles in the same game off Righetti. Those were the only XBH Thomas ever got off Rags in 25 ABs.
John – There are only 23 instances in which a pitcher has given up 40 or more homeruns in a season. Not sure how many times that resulted in a spread of 40 HRs between qualifying pitchers but I’m guessing very, very few. Far fewer than the reverse (i.e., a spread of 40 homeruns between qualifying hitters).
I concede the point. I thought there were more pitcher 40-HR seasons.
Blyleven did have a 40+ spread over several qualifiers in 1986, though.
This seems like a very good system for normalization, especially for something so easy to calculate. But the one possible flaw i see is that because you are only taking numbers from 5 years an exceptional year for a pitcher or for the league can really skew the numbers one way or another. I can’t think of a way to prevent this off the top of my head but otherwise it seems like a great system.
Raphy’s going to post an even better method that avoids that problem and covers a lot more pitchers.
FWIW, I now have a spreadsheet giving normalized K rates based on the entirety of the pitcher’s career — i.e., it divides his career K rate by the average of the MLB K rate during his career.
I have this for all pitchers with 1,000 IP (roughly 1,000 pitchers).
Still not perfect, since it’s not weighted — a high MLB K rate for a given year has the same impact on a given pitcher regardless of whether he threw 1 inning or 300 IP in that season.
It took me a while to make Excel 2007 do my bidding. The “AVERAGEIFS” function was the key for me, but as usual, I had to find someone outside the Microsoft universe to show how to make it do exactly what I wanted.
The top 20 in my list (min. 1,000 IP):
Player, Factor
Rube Waddell, 2.24
Dazzy Vance, 2.06
Cy Seymour, 1.91
Sandy Koufax, 1.83
Nolan Ryan, 1.78
Smoky Joe Wood, 1.77
Randy Johnson, 1.76
J.R. Richard, 1.69
Amos Rusie, 1.69
Pedro Martinez, 1.67
Walter Johnson, 1.67
Lefty Grove, 1.66
Bob Feller, 1.65
Dizzy Dean, 1.62
Doc McJames, 1.59
Van Mungo, 1.58
Kerry Wood, 1.58
Lefty Gomez, 1.58
Sam McDowell, 1.58
Lee Smith, 1.57
Tommy Bridges, 1.57
I forgot to mention that my data is based on SO per batter faced, not per inning.
Intuitively, this list looks and “feels” better than the other career lists I have seen here. By lowering the minimum to 1,000 IP, you include quite a few of the legendary hard throwers, of all eras and career lengths.
The only names I go “huh?” to are Cy Seymour and Doc McJames. Well, I guess no matter what the minimum is, someone’s going to just barely meet it doesn’t belong. Doc McJames is a sad story, he died at age 27 after being thrown off a horse in 1901.
I want to point out a statement made in the Sunday New York Times this past week (Feb 19th). In an article about culture and baseball in 1968, Tim Wendel states that Bob Gibson was saddened by the death of MLK (April 4th).
“Perhaps as a result, Gibson did not start the season well. But when Kennedy was gunned down after winning the California primary, Gibson responded by pitching his first shutout and went on to put up a season for the ages…”
From a human perspective, to speculate that Hoot pitched poorly or well as a response to political-cultural events is kinda disrespectful to Gibson. Perhaps as a journalist he could phone Gibson and actually ask him how he felt.
But from a statistical perspective it just shows that “old school” mentality is still getting printed in the old black and white.
On June 6th Gibson clearly was having a bad season, his record was 4-5.
and…
1.66 era
.498 ops against
All losses came while pitching at least 8 innings.
All losses while giving up 3 ER or less.
Wonder what Wendel would say about these seasons (just to pick a few at random).
Ned Garvin, 1904 BRO-NYY, 5-16, 159 ERA+
Hal Newhouser, 1942 DET, 8-14, 162 ERA+
Darold Knowles, 1970 WSA, 2-14, 174 ERA+
Or, maybe Ted Lyons retired in 1946 because his 1-4 record showed that, at 45, he just couldn’t get it done anymore (actually, Lyons completed every game he started that year, and had a 174 ERA+ – he retired after being named the White Sox manager).
Wait, is Wendel implying that Gibson was so thrilled by Kennedy being shot that he pitched incredibly well? That’s… kind of a horrible thing to say.
I think what he’s saying is that Gibson brought renewed focus after Kennedy’s assassination. It’s good writing if not especially accurate
Here’s the article:
(And, to be fair, the article is pimping a book soon to come out. Perhaps the book digs a little deeper)
http://www.nytimes.com/2012/02/19/sports/in-1968-sports-helped-to-temper-a-year-of-rage-and-upheaval.html
Not a bad article; for those of you not around in 1968, the Tigers great season really did help the city of Detroit (somewhat), by giving people something to rally around.
Taken out of context, it’s easy for us to ridicule his point about Bob Gibson starting “poorly” because of MLK’s assassination I’m going to guess that he was just looking for an easily understood stat to move his narrative along, without digging deeper. Still, Voomo Zanzibar is right in that if the writer is going speculate about the cause/affect of the start of Gibson’s season, he should have actually contacted Gibson himself to ask him.
Bill Freehan advanced this same theory (about Detroit, not Gibson) in his 1969 diary, “Behind the Mask.”
Which is a great read, by the way, especially if you’re a Tigers fan. Not a “Ball Four”-style expose, it’s just a diary of their ’69 season (the year after the championship) with reflections on his own career and the state of the game.
Andy:
I would love to see how you did the spreadsheet. Any chance you could Email it to me?
The thing that jumped out to me was Bob Forsch, with a 3.65 k/9 rate, and a normalized rate that is 5th lowest of all-time for the top 200 pitchers, yet he managed to throw 2 no-hitters. His 1983 no-no he must have been fireballing, as he racked up 6 k’s, but his 1978 gem only yielded 3.
Seems pretty lucky for a guy with that rate to have 2 no-hitters.
Why do I not remember David Cone as a high strikeout pitcher? As a Jays’ fan, my memories of him must be dominated by his two tours of duty in Toronto and his stint in KC in between, and he wasn’t a great strikeout artist during those seasons.
Cone averaged 7.3 SO/9 for his two Toronto stints, which was below his career average but still well above the AL average for that era. His high SO game with Toronto was 11:
http://www.baseball-reference.com/boxes/BOS/BOS199506280.shtml
I like the helpful information you supply to
your articles. I’ll bookmark your blog and test once more right here frequently. I am somewhat sure I will be informed plenty of new stuff right right here! Best of luck for the next!