August 21, 2007
Non-Contact Part IV: Take, Jive, and Flail
There was a lot of good email from last week's column, so let's just dig in. Reader D.D. had some comments about the note on which I finished up last week's talk of strikeout rates:
I'm sure you'll address this, and have already thought of it, but when you mention 'it's logical to think that if a strikeout is the best a pitcher can do, it's the worst thing a batter can do, right? But if you look at the lists of high-strikeout batters above, does the strikeout being the worst possible result show up here?'
I must be telegraphing my punches, because D.D.'s absolutely right. The PA and career strikeout minimums we required to make the strikeout kings lists coherent also virtually guarantee that many of the high-strikeout mavens would not suck-since guys who whiff a lot and also don't hit are unlikely to stick around to collect 500 plate appearances in a season, or strike out 500 times for their career. It's a little like losing 20 games-you usually have to be a pretty good pitcher to have the opportunity to lose that much.
To shed some light on this topic, let's look at some analogous lists, this time with pitchers:
Career Walk Rate, 1959-Present (min. 650 career walks) Name BB PA BB Rate EqERA Bobby Witt 1375 11003 12.50% 4.81 Sam McDowell 1312 10589 12.39% 3.85 Nolan Ryan 2795 22575 12.38% 3.90 Blue Moon Odom 788 6531 12.07% 5.15 Russ Ortiz 805 6875 11.71% 4.73 Tom Griffin 768 6567 11.69% 5.00 Shawn Estes 840 7205 11.66% 4.87 J.R. Richard 769 6669 11.53% 4.00 Al Leiter 1163 10334 11.25% 4.01 Jamey Wright 753 6761 11.14% 4.88
Now, this is our first peek at another non-contact result that some consider the opposite of the strikeout-the walk. Walks aren't even arguably the worst thing that can happen to a pitcher on the mound-the home run is almost as difficult to defend, and its impact is far more immediate-but as the second coming of Rick Ankiel reminds us, few things can derail a pitching career faster than a loss of control. Alongside their walk rates, I've given you these guys' normalized career ERAs or 'EqERAs,' taken from the translated stats section of the DT Player Cards. The translated stats are set up so that the average pitcher's EqERA is 4.50, and as you can see, a number of these high-walk pitchers still managed to pitch pretty well.
Highest Unintentional Walk Rates, 1959-Present (min. 150 IP) Year Name IP PA BB UBB Rate K Rate EqERA 1. 1986 Bobby Witt 157.2 741 143 19.03% 23.48% 5.51 2. 1991 Randy Johnson 201.1 889 152 17.10% 25.65% 4.10 3. 1959 Herb Score 160.2 712 115 16.15% 20.65% 5.67 4. 1971 Nolan Ryan 152.0 705 116 15.89% 19.43% 5.30 5. 1992 Randy Johnson 210.1 922 144 15.51% 26.14% 4.17 6. 1977 Nolan Ryan 299.0 1272 204 15.49% 26.81% 3.08 7. 1991 Jose DeJesus 181.2 801 128 15.48% 14.73% 3.98 8. 1975 Nolan Ryan 198.0 864 132 15.28% 21.53% 4.08 9. 1975 J.R. Richard 203.0 905 138 15.25% 19.45% 5.30 10. 1976 Nolan Ryan 284.1 1195 183 15.15% 27.36% 3.98
This search is refined slightly, so that we're looking at the unintentional walk rates of the pitchers in our database. This eliminates the element of managerial interference, and gives us a more direct view of the pitcher's control, or lack thereof. As it stands, we get a Hall of Famer (Nolan Ryan) another guy who's a lock to join him (Randy Johnson), plus two fellows (J.R. Richard and Herb Score) who had some nice years before their careers were cut short by devastating injuries, three if you count Jose DeJesus blowing out his elbow after his breakout season in Philly. All of the pitchers except DeJesus maintained well above-average strikeout rates. As we can see by his prominent placement on both lists, Bobby Witt was the somewhat more successful pitching counterpart of Dave Nicholson-a man beset upon by people who thought they could fix his inability to stay within the strike zone. I guess every pitching coach wants to be the guy who puts the glasses on Rick Vaughn.
Getting back to strikeouts again, reader M.P. chimed in with another observation:
One thing that seems noteworthy is that the top two rates for career percentage K rate are actually higher than the all-time season tops. I cannot think of a single other stat where that is the case-nor even many where it is feasible. What a strange stat!
I think this is another effect of the selection bias attendant to the minimums we set last time, and the fact that, for hitters, strikeouts are a negative stat; even if you're in the camp that says that a strikeout is just another out, a high-strikeout hitter is someone who's adept at making that certain kind of out. So you're likely to see careers where that player's playing time at the major league level comes in smaller pieces than, say, a hitter who's extremely adept at hitting home runs.
For example, setting the leaderboard's minimums to 400 plate appearances, rather than 500, yields an entirely new set of leaders:
Season Strikeout Rate, 1959-Present (min. 400 PA) Year Player Pos G PA K K Rate 1. 1997 Melvin Nieves RF 114 405 157 38.77% 2. 1987 Bo Jackson LF 116 434 158 36.41% 3. 2002 Russell Branyan LF 128 435 151 34.71% 4. 2002 Jared Sandberg 3B 101 401 139 34.66% 5. 1963 Dave Nicholson LF 125 520 175 33.65% 6. 1987 Rob Deer LF 134 566 186 32.86% 7. 1986 Rob Deer RF 134 546 179 32.78% 8. 1988 Pete Incaviglia LF 115 467 153 32.76% 9. 1996 Melvin Nieves RF 120 484 158 32.65% 10. 1991 Rob Deer RF 133 539 175 32.47%
As you can see, lowering the threshold drops the protagonist of last week's column, Dave Nicholson, into fifth place. In the interest of saving space, I'm linking the sortable stat reports where lower minimums are set. (WARNING: These are authors-only "all years" searches, that are, for most intents and purposes, "read only" reports; sadly, you won't be able to perform any "all years" searches of your own, you'll only be able to look at the ones that I've already set up. Sorry for the inconvenience.) At 300 plate appearances, Nicholson's 1963 season drops to tenth place, and lower the threshhold down to 200 plate appearances, seventeenth place.
If you'll indulge me one more reader mail, we'll move this discussion of batter's strikeouts along. Frank Greenberg writes in:
I know that you people in general are proponents of players who go deep into pitch counts. While no one can doubt some of its effectiveness (building up pitch counts in starting pitchers) it is also taking the aggressiveness out of the game. I am 62 years old, as a kid we were always told with two strikes on you protect the plate, if it's close, swing or at least try to foul it off. Not any more-players now take pitches with two strikes and pray that the umpire calls it a ball. Players seem to be just as happy to walk and, let the next guy worry about driving in the run.
Leaving a few things aside (such as the fact that Youkilis's strikeout rate this year, 15.13%, is below the league average of 16.78%) what is the deal with stathead types actually favoring guys that whiff a lot? If the strikeout is the pitcher's most powerful weapon, and a high strikeout rate is an important signpost of his future success, how can it also be that a high strikeout rate is a good thing for hitters, too?
The idea is that high strikeout rates bring with them things that you want out of a batter-more walks and more power. But in the Prospectus Basics article on strikeouts I referenced last week, the correlation between strikeouts and walks is tenuous (many studies suggest there's no connection at all) and the correlation between strikeouts and isolated power, while positive, isn't terribly strong.
The question might boil down to an issue of style. Not all whiff-prone batters are the same-some strike out a lot because they go deep into counts looking for good pitches to drive, others strike out a lot because they go up to the plate hacking away like a guy trying to break open a piņata full of diamonds. There's an ongoing nature/nurture debate over these two types of hitters, with nurture types (like Mr. Greenberg seems to be) alternately accusing the hackers of not being taught discipline, and the "take and rakers" of not being taught to hit aggressively.
The nature argument posits that these two types of hitters exist because of skill sets each naturally possesses, and that it's like trying to beat a square peg into a round hole if you insist that hacker work the count, or that a take-and-rake guy expand the strike zone. If you believe that, the key thing is to identify the best of each world, and let them be themselves-don't try to force David Ortiz to hit like Ichiro Suzuki, or vice versa.
The answer is probably somewhere in between. I'll leave you with one more list, of strikeout rates for swinging and called strikeouts, which we'll discuss next time.
Season Called/Swinging Strikeout Rates, 1999-Present (min. 500 PA) Year Name Called K% Year Name Swinging K% 2006 Pat Burrell 11.11% 2002 Jose Hernandez 25.77% 2002 Ben Grieve 11.05% 2001 Jose Hernandez 24.66% 2004 Mark Bellhorn 10.81% 2003 Jose Hernandez 24.52% 2001 Luis Castillo 10.29% 1999 Preston Wilson 21.55% 2001 Ben Grieve 10.17% 2004 Jose Valentin 21.23% 2005 Pat Burrell 10.16% 2000 Preston Wilson 21.07% 2002 Mike Cameron 10.00% 2001 Jim Thome 20.81% 2001 Pat Burrell 9.71% 2001 Richie Sexson 20.69% 2004 Adam Dunn 9.69% 2003 Sammy Sosa 20.20% 2000 Jim Edmonds 9.49% 2000 Richie Sexson 20.10%
Tom Tango, "Strikeouts-Just Another Out?": A study examining the effect on player performance of a substantial increase or decrease in strikeout rate.
Russel Carleton, "Is Walk the Opposite of Strikeout?" in By the Numbers, Feb. 2007: A more academic approach, looking at the relationship between walks and strikeouts.
Nate Silver, Lies, Damn Lies, "Strikeouts and Hitter Projections": Nate discusses the role of strikeouts as a positive value in projecting hitter performance, and the distinction between early-count and late-count hitters.
William Burke and Jason Pare contributed indispensable research to this article.