All You Never Wanted To Know: Quality Starts

I have an editor friend who is astonishingly forgiving. If you show up an hour late for lunch, she will pleasantly lie that she just got there herself. If you bump into her car, she will say that dent isn’t even noticeable. And so on. But if you use the word “quality” to mean “general excellence,” she will unfailingly turn schoolmarm.

“You mean HIGH quality,” she will snap. Dictionaries allow for “quality” to mean “superiority in kind” (Merriam Webster) or “general excellence” (Oxford) but my friend will tell you that this is because dictionaries have become wimpy and won’t stand up for anything (“Dictionaries are of low quality these days,” she might say). In her world “Quality” will always mean a DEGREE of excellence, meaning the word will always need a modifier to make it say anything. Something can be of high quality, low quality, barely acceptable quality. But to say something like “This is a quality sandwich,” or “She is a quality human being,” will make her insane.

And so, naturally, she cannot stand the whole concept of the “quality start.” “It’s doesn’t MEAN anything,” she will say. But the funny thing is: I think even by her rigid definition, the term “quality start” works. Because a quality start isn’t necessarily a “high quality” start. It can be. It isn’t necessarily a “inferior quality” start. But it can be that too. The quality start standard is very simple: A pitcher must throw at least six innings and allow three earned runs or fewer. But that’s a big range.

In July 2000, Mark Mulder went 6 2/3 innings, gave up 15 hits and nine runs — but only two were earned and so that was a classified as a quality start.

In June 1997, Randy Johnson struck out 19 in a complete game but allowed 4 runs. That was not a quality start.

In July 1982, Mike Scott allowed seven hits and walked five in six inning, didn’t strike out anybody, gave up seven runs, but only three of those were earned. Quality start.

In April 1974, Gaylord Perry went 15 innings and allowed four runs. Not a quality start.

So just as the word “quality” can be purposely vague — kind of like “Nilla Wafer” seems purposely vague* — the quality start is a blurry, indistinct thing. And yet, it’s also a fixed thing. We can count them. And while it doesn’t give us a complete picture of a pitcher, well, it gives us a little something we can look at, discuss and so on.

*As comedian Gary Gulman asks: Do they mean “VA-Nilla? Is that the point? And is it really a wafer? I don’t see it as a wafer.

Kyle Davies now has the worst quality start percentage for any pitcher with 100 or more starts. He earned that on Saturday night with his decidedly low quality three inning, eight hit, seven run (though only four earned) outing. It was his first outing back from injured reserve, which would normally be a “hey, let’s let the guy get healthy before making any judgments” kind of moment. Unfortunately, this is pretty much how Kyle Davies pitches. The performance only lifted his ERA a little bit — from 7.46 to 7.77 — and as mentioned he passed John Halama for lowest quality start percentage for a career.

Lowest quality start percentages (since 1950)
1. Kyle Davies (46 of 141, 32.6%)
2. John Halama (39 of 119, 32.8%)
3. Kevin Jarvis (40 of 118, 33.9%)
4. Casey Fossum (42 of 120, 35%)
5. Brian Meadows (44 of 122, 36.1%)

It takes skill to get this many big league starts with that low a quality start percentage. To give you an idea: There have been about 750 pitchers who have gotten 100 starts since 1950. They throw a quality start about 55% of the time. These are the most established guys, understand, and so you EXPECT them to throw a quality start more often than not. Only 26 of the 750 or so have a quality start percentage lower than 40%.*

*And I now give you the stat of the day: Two of the 26 currently are in the Kansas City Royals rotation: Davies and Bruce Chen. And Luke Hochevar is also less than 40% but he only has 83 starts. No wonder they are going to a six-man rotation — when you have that many options …

In Davies case, people act all shocked about why the Royals keep letting him pitch — I have received numerous “he must have pictures of Royals GM Dayton Moore” emails — but it’s really not a mystery. Davies has good stuff. He’s a likable guy. He was up in the big leagues at 21 as a big prospect and has flashed occasional signs of fulfilling his potential. For a Royals team desperate for a break, they see their best hope as giving him chances to blossom. It’s not really a mystery. It’s self-defeating. It’s pointless. It’s soul-crushing for fans who want the Royals to actually improve. But that’s different from a mystery.

Let’s switch gears, though, and talk for a moment about the HIGHEST quality start percentages.

Highest quality start percentages (since 1950)
1. Tim Lincecum (101 of 139, 72.7%)
2. Tom Seaver (454 of 647, 70.2%)
3. Adam Wainwright (83 of 119, 69.7%)
4. Mel Stottlemyre (247 of 356, 69.4%)
5. Roy Oswalt (216 of 316, 68.4%)
6. Josh Johnson (77 of 113, 68.1%)
7. Bob Gibson (328 of 482, 68%)
8. Roy Halladay (226 of 338, 66.9%)
9. Felix Hernandez (127 of 190, 66.8%)
10. Randy Johnson (403 of 603, 66.8%)

Greg Maddux, in case you were wondering, is at 64.9% — but from 1992 to 1998 it was an amazing 81% (183 of 226). Koufax for his career was at 65.3% — but again, at his height, he threw a quality start pretty much every time out. From 1963-66, he threw quality starts 126 out of 150 times. That’s 84%.

Here’s one for you: Only two pitchers have had a quality start percentage of 90% or better in a season. You should be able to guess it, I think. I like giving trivial questions where people can actually get the answer. So what do you think? Two pitchers with 90% quality start percentages in a season. Go ahead: Start the Jeopardy music. Think about it. I’ll wait for you. Don’t look ahead.*

*Here … I’ll even put a Posterisk in here so you won’t look ahead. Have you picked up your Sports Illustrated iPad app yet? You’ve got to do it. This week. we have a cool surprise cover, we have video from the movie Major League and soccer commercials with Dwight from The Office, we have samples of the Top 60 sports songs ever (though, I must admit to disagreeing with a few of them). It’s pretty great. And, it’s free. As in: No money. As in … oh, you ready for the answer? OK.

Bob Gibson in 1968 is one of the pitchers, of course. That year, Gibson made 34 starts — 32 of them were quality starts.

The other? Yep: Dwight Gooden in 1985. What a year. He had quality starts in 33 of his 35 games.

Other amazing quality start years:

– Koufax in 1966 (36 of 41)
– Seaver in 1968 (31 of 35)
– Seaver again in 1971 (31 of 35)
– King Felix in 2010 (30 of 34)
– Wilbur Wood in 1971 (37 of 42)
– Roger Clemens in 1990 (27 of 31)
– Don Sutton in 1980 (27 of 31)

Fun stuff. But, of course, once I got started playing around with these numbers, I could not leave it there. I wondered about what I’ve heard baseball people call the TQS — TRUE Quality Start. Opinions vary about the TQS, of course, but generally I think we might be able to get a consensus that a True Quality Start is a pitcher going at least SEVEN innings and allowing at most TWO earned runs. Obviously, the True Quality Start has some of the same problems as the Quality Start, but for the most part any pitcher who throw seven innings and allows two earned runs has done a commendable job.

Before I tell you who has the worst TQS percentage (hint: It’s actually NOT Kyle Davies), here’s a painstakingly compiled list of the Top 10:

Highest True Quality Start Percentage (since 1950)
1. Bob Gibson (245 of 482, 50.8%)
2. Jim Palmer (262 of 521, 50.3%)
3. Mel Stottlemyre (177 of 356, 49.7%)
4. Sandy Koufax (155 of 314, 49.4%)
5. Tom Seaver (318 of 647, 49.1%)
6. Bob Veale (125 of 255, 49%)
7. Andy Messersmith (143 of 295, 48.5%)
8. Tim Lincecum (66 of 139, 47.5%)
9. Roger Clemens (333 of 707, 47.1%)
10. Whitey Ford (206 of 438, 47%)

The average True Quality Percentage, by the way, for starter who made 100-plus starts: 34.8%. That means you would expect an an established starter to throw a TQS about one third of the time, maybe slightly more. This seems a good time to sing the praises of Mel Stottlemyre. He had the misfortune of playing his first year with the Yankees in 1965 — the first year since the 1920s that the Yankees were no good. And they were no good his entire career — the one decade that the Yankees were terrible — and he blew out his rotator cuff at 32 just before they got good again.

But he won 164 games with a career 2.97 ERA. And you can see he is high on the list in both Quality Start Percentage and True Quality Start percentage. A large part of this was the pitcher’s era: His 2.97 career ERA added up to a nice 112 ERA+. Carl Hubbell’s 2.98 ERA, though, gave him a 130 ERA+. There just weren’t many runs being scored in Stottlemyre’s time, and so his percentage of well-pitched games is a bit misleading. But he did lead the league in complete games twice, place in the Top 10 in ERA five times, and his sinkerball yielded few home runs. And while it was the era, it’s still true that he threw more True Quality Starts than Koufax. He was a damned good pitcher.

Lowest True Quality Start Percentage (in reverse order)
10. Kevin Jarvis (17 of 118, 14.4%)
9. Elmer Dessens (20 of 140, 14.3%)
8. Scott Elarton (23 of 170, 13.5%)
7. Mark Hendrickson (22 of 166, 13.3%)
6. John Halama (14 of 119, 11.8%)
5. Claudio Vargas (13 of 114, 11.4%)
4. Tim Redding (14 of 144, 9.7%)
3. Kevin Correia (11 of 122, 9%)
2. Kyle Davies (11 of 141, 7.8%)
1. Casey Fossum (9 of 120, 7.5%)

So Casey Fossum edges out Davies as least likely to give you seven innings and allow two earned runs. But, at last check Kyle Davies was still in the Royals rotation. So the battle’s not over yet.

21 thoughts on “All You Never Wanted To Know: Quality Starts

  1. Mark Coale

    I love the SI app (especially now that it’s free to subscribers).

    however, I couldn’t believe that “Joltin’ Joe DiMaggio” didn’t make the sports songs list. And this from a Yankee hater.

    You know who had a bizarre quality start last night? Barry Zito, who actually started a game, sat through a 3 hour rain delay (with 20 degree temp drop) and then came out and pitched 3 more innings.

    Reply
  2. Ed

    You forgot about Pedro going 26 of 29 in 2000 — I didn’t bother counting up all the seasons (because baseball-reference doesn’t list quality starts as a stat, which means I had to manually count them using the game logs), but I think it’s possible he had a stretch even better than Maddux’s from, say, 1997-2003.

    Reply
  3. Ed

    I was mentioning Pedro’s 2000 season just as a potential addition to other amazing quality start years — but actually, if you want to round up, he hit 90% that year. 26 of 29 is 89.6%….technically not at the 90% threshold though.

    Reply
  4. Dave V.

    It looks like going into this season, Kevin Correia must have been worst of all time, coming in with a 6.7% TQS percentage (7 out of 105). However, he has 4 such starts this season already.

    Reply
  5. ChopperToChipper

    Joe, thank you very much as always for the high quality of your writing. I hope that the move to North Carolina is treating you well. I spent three years in the state and found it to be a wonderful place.

    I think you just missed another addition to the 90% list when it comes to Maddux’s brilliance. As a kid growing up going to games at Atlanta-Fulton County Stadium I remember the strike year being Maddux’s best, and it turns out that he was 24/25 (96%) when it came to quality starts that year. It’s also possible you’re simply excluding the year that the Expos were going to win the World Series from consideration.

    Reply
  6. Vidor

    Stottlemyre’s first year with the Yankees was not 1965. He was a midseason callup in 1964 and, as David Halberstam wrote about in “October 1964″, he was a big reason that the Yankees won the pennant. He started three times in the 1964 World Series, with one W, one ND, and an L in Game 7.

    Reply
  7. Biggus Rickus

    Pedro did have a better peak than Maddux. Given his era, he may have had a better peak than anyone ever. There are a lot of amazing stats from his brilliant 2000 season, but the one that always amazes me was his .737 WHIP. I also pulled up Maddux’s ’95 season to see what his QS numbers were. He was 23/28, but he probably should have been 26/28. He allowed one run in each of his first three starts, but was pulled after 5, 5.2 and 5 innings for some reason.

    Reply
  8. jjjjj

    A person who objects to the use of “quality” to mean “high quality” and thinks dictionaries have *become* lax is not too well-informed. The linguistic practice of taking a word that can be good or bad and using it to mean one or the other is thousands of years old. In Latin “formosus” literally means “having a form” but it came to mean “having a good form,” “beautiful,” like Spanish hermoso. “Casus” means “thing that happens,” but comes to mean “disaster.”

    Reply
  9. gwowen

    @jjjjj Quite right. The OED has cites for “quality” used as synonymous for “excellence” going back to 1874, and arguably, further still. I don’t mind language prescriptives, when they’ve taken sufficient care to actually be correct in their prescriptions.

    Reply
  10. KCDC

    Davies needs 6 more starts without a TQS to sink below 7.5%. If I know the Royals and Davies, he’ll get there in August.

    Reply
  11. bannister19

    To be fair to Bruce Chen, he’s had a career filled with nothing by injuries. He has by far been the best Royals starter over the last two years.

    As for Kyle Davies, well he just sucks.

    As for the statistic itself, I guess it obviously depends what you define as quality, as you mentioned.

    If I had to make it a statistic, I would say:
    6IP, 2ER(Or Less) (3.00 in ERA terms)
    7IP, 3ER(Or Less) (3.85 in ERA terms)
    8IP, 3ER(Or Less) (3.38)
    9IP, 4ER(Or Less) (4.00)

    Plus one more parameter, your WHIP for that game could not be higher than 1.30. Meaning, if you’re going to go 6IP, you can not allow more than 7 hits/walks combined.
    If you’re going to go 7, you can allow 9 runners.
    If you’re going to go 8, you can allow 10.
    If you’re going to go 9, you can allow 12.

    It obviously has flaws. I’ve seen games where a pitcher has allowed 6 hits and 5 of them were infield hits and the other was a little blooper. Anything will have flaws.

    Finally, for the current QS, I don’t mind it too much because basically, any time a starter goes 6ip and allowed no more than 3ER, it keeps the baseball team in the game. Teams score an avg of around 4.5 runs per game per team per season. If you allow 3, you still give your bullpen some leniency and the numbers are in your favor to win.

    Now, the major flaw is in my opinion, the 9IP 4ER one.

    Reply
  12. 7692cc2e-a6b9-11e0-b8b2-000bcdcb2996

    8. Scott Elarton (23 of 170, 13.5%)
    6. John Halama (14 of 119, 11.8%)
    4. Tim Redding (14 of 144, 9.7%)

    all Astros prized pitching prospects… wonder why they’re in the cellar now…

    (well there’s plenty of reasons but just another prime symptom of the overall disease… and it isn’t lupus, even if he wasn’t allowed to play in the astrodome.)

    Reply
  13. deadrody

    And for the lowest percentage of “true quality starts” in baseball history, you too can have a HOFer to augment your championship seeking Red Sox.

    I give you Curt Schilling virtually straight up from AZ to Boston for Casey Fossum. Nice job, AZ.

    Reply
  14. allan

    We all have heard announcers make fun of the Quality Start because it = a 4.50 ERA, which is more league average than quality.

    But some time ago, someone figured out how many Quality Starts were actually simply the nice but meh 6 IP/3 ER minimum. If I recall, it was a clear minority, so most Quality Starts actually *are* a quality start.

    Anyone remember that?

    Reply
  15. Pingback: Winning Sucks | Bill DiFilippo

Comment: