Plotting all of Greg Maddux’s career Game Scores
Posted by Andy on July 5, 2010
Here's a plot of the game scores for all 740 of Greg Maddux's career starts:
Let me first tell you how this plot works:
- Each gray circle is an individual game start for Maddux. They are ordered sequentially from his first career game start in 1986 to his last career game start in 2008. So the main scale at the bottom ranging from 1 to 740 are his 740 career starts in date order.
- I added some lines and boxes separating his starts into years. The first few starts were in 1986 (not labeled) and each subsequent year is labeled.
- The red line is a moving average of 25 of Maddux's starts. For example, where the red line crosses the 200th start line, it's an average of his game scores in his 188th through 212th career starts.
- The blue line shows the average game score for all National League starts in each given year.
So what does this plot tell us?
- The average game score tends to be right around 50. So in a gross sense, we see that Maddux started out a little below average in 1986-1987, climbed above average by the end of 1990, was well above average for the majority of 1992-2002, and gradually fell just below average at the end of his career.
- His highest peak was in the period 1992-1995. Incidentally, he won the Cy Young award each of those seasons...
- The behavior in 1998 might be deceiving. It looks like he took a major nosedive that year. However the numbers don't match up with that. He had one of his best years that season, including a league-leading ERA and WHIP as well as 18 wins. I think the real issue is his 1999 performance, which (by virtue of the averaging) tends to drag down the second half of his 1998 numbers. 1999 was a below-average year for Maddux (what most any other pitcher would call a great year), with his lowest ERA+ since 1991. Note too that league game scores plummeted to well below 50, so the baseline dropped in 1999, meaning that Maddux was still more above average in 1999 than you might think based solely on the red line.
- Note too that you can pick out the highest and lowest game scores across Maddux's career. For example, other than one start in mid-1996 with a game score of 10, he didn't throw any games below 20 from mid-1990 until early 1999, a span of nearly 300 starts. That's incredible.
July 5th, 2010 at 8:21 am
Very cool. I'm not sure why you choose to do the moving average from the midpoint. That's what generates the deceiving drop in the red line before it actually happens in the individual games. I'd use a backward looking MA
July 5th, 2010 at 10:39 am
That's very cool. Maddux's '92-95 is kind of underrated if that's possible. His '94-95 don't get the full attention they deserve because of the strike. He had a 271 & 262 era+ during those years which are among the top 4 in baseball history. From 1992-1998, Maddux had a 191era+, which is just insane when you think about it.
July 5th, 2010 at 10:45 am
Maddux's 1999 season was the one that caught Voros McCracken's eye and led to his seminal paper on BABIP and a pitcher's lack of control of it which led to DIPS theory. After allowing a BABIP in the mid to upper .200's through the 90's, Maddux's BABIP suddenly and inexplicably jumped to .331 in 1999 after which it went right back down to the upper .200's.
IOW 1999 was just a major fluke year for Maddux and it shows in the trend line here.
July 5th, 2010 at 11:46 am
What is interesting to me is looking at the worst games he pitched in any given year. At first in his career it was pretty normal, he'd get shelled (game scores of 30 and below) every now and then, 21 total game scores below 30 in his first 5 seasons. Then from '92 to '98, a period of seven seasons, he only had 8 times when his game score was even below 40! Plus now he's throwing more games in the 80+ range too. That's a crazy combination of consistency and brilliance.
Also, this is awesome, I'd love to see it for more pitchers.
July 5th, 2010 at 11:52 am
Err, by 8 times under 40 I mean 11, but my lack of counting skills doesn't really affect my point I think. That 10 game score in 1996 is also interesting, since it's a freak outlier for that period. That game was Atl vs Col, June 7, 1996, where he gave up 7 runs on 11 hits in 3.1 innings of work, without any HR allowed incidentally, he gave up 9 singles and 2 doubles.
http://www.baseball-reference.com/boxes/COL/COL199606070.shtml
July 5th, 2010 at 6:36 pm
Maddux got screwed by the Cy Young voters in 1998. He had 206.2 innings and a 1.65 ERA through mid-August but cooled off a bit the final six weeks of the season (which if you notice is just as good as his performances during the strike shortened 1994 and 1995 seasons). But still, he was so dominant during the first 4 and a half months of the 1998 season, that a slump during late August and September didn't do a whole lot of damage to his statistics. He still finished with a 2.22 ERA, 187 ERA+, 251 IP, and 204 K's.
But the idiot voters chose Glavine because of the win totals and nothing else.
July 5th, 2010 at 6:41 pm
1994: 1.56 ERA, 202 IP
1995: 1.63 ERA, 209.2 IP
1998: (through August 18th start): 1.65 ERA, 206.2 IP
July 5th, 2010 at 7:06 pm
Love the graph!
As a sidebar, looking at Maddux' similarity scores, 8 of the top 10 similar pitchers are in the HoF (and the other should be). No surprise there. But 5 of the top 10 simlir BATTERS are in the HoF, and another one (Glavine) may be. That seems a tad bizarre.
July 5th, 2010 at 8:45 pm
Without looking, it's not that bizarre because most pitchers who last as long as Maddux have had good careers at the very least. So anyone who hit as badly as Maddux in as many PA must be a pitcher who lasted a long time.
July 6th, 2010 at 8:56 am
I love this! I'm looking at doing some grunt work on this for a couple of pitchers, but how did you find the average game score for all of the starts in a league for a particular year? Did you use the play index?
July 6th, 2010 at 1:42 pm
Redsauce:
I got the numbers here:
http://www.baseball-reference.com/leagues/NL/2009-starter-pitching.shtml
and I changed the year and picked off the relevant number on each page.
July 9th, 2010 at 5:29 pm
[...] In case you were still unaware, Greg Maddux was awesome. [...]
July 11th, 2010 at 10:55 pm
If Maddux's career had been ten years shorter, the impact that the 1994-1995 strike had on his two greatest seasons might well have been considered the most significant negative career impact resulting from the strike (I remember that Bill James, in deciding in the updated Historical Baseball Abstract that Clemens beat out Maddux for peak & career value, ignored the impact of the strike when comparing innings numbers for their best seasons--since Clemens was hurt in 1994 and subpar in 1995, this hurt Maddux far more than Clemens), as it might have meant the difference between Maddux getting into the HOF and not doing so. As things went in reality, it was just a speed bump, and he will go in on the first ballot. Who said life isn't fair? 🙂