[Note: The following article was originally published as Issues 41 and 42 of The Joe Sheehan Newsletter. Given the attention recently paid to Jack Morris in a recent Sports Illustrated piece, we are reprinting this article, with Joe's permission, to address some of the questions pertaining to Morris' career and his Hall of Fame worthiness.]
This whole thing started at the winter meetings. I ran into a newsletter reader who wanted to talk to me about Jack Morris. See, I'd written this in my evaluation of Hall of Fame candidates:
"Morris wasn't as good as [Tommy] John or [Jim] Kaat, and kind of
the reverse of Bert Blyleven in that he put up good records thanks to good
teammates, and on that has built a Hall of Fame case. He's also comparable to
Joe Carter in that he's a low-borderline candidate whose support statheads
will decry but who has a World Series moment that is going to carry a lot of
weight with voters.
For his career, though, all he has are the wins and some moments. His career
adjusted ERA is just 105, meaning he was five percent better than the league
over his 3,824 innings. Morris was a workhorse, and that's a valuable thing,
but durability isn't greatness and there's no need to confuse the two. Bert
Blyleven was great. Morris was good with better support."
I had a brief exchange with the reader, who argued that Morris' value was in
his innings pitched and wins, and additionally that his value went beyond the
stats. I replied that if Morris had more value than his raw numbers indicated,
we should be able to find that using Retrosheet. It was an enjoyable exchange,
one I wish we'd both had more time to flesh out.
The problem is that I left Nashville with the nagging idea that this
could be done, that it would be possible to analyze Morris' career and
see whether he really was someone whose performance record hid some additional
value. I was thinking in terms of Greg Spira's analysis of "Pitching to the Score" at the time. In doing the research, I've also borrowed from Bill James, specifically his Baseball Abstract 1988 comparison of Walt Terrell and Danny Jackson.
What I'm trying to do is determine if we've missed something. Did Morris
actually "pitch to the score" in such a way that inflated his 3.90
career ERA? Were his wins the result of his durability for good teams, and did
he get a disproportionate amount of support from his teammates? Is my
evaluation of Morris' place in history--a good pitcher with some tremendous
career highlights, but not quite a Hall of Famer--accurate?
With that goal in mind, I entered every one of Jack Morris' 527 career starts
into an Excel file, along with the starts of his teammates in that time. I
charted the innings pitched and runs allowed in each start, along with whether
the starting pitcher left with the lead, tied or losing. I charted run support
(both while the pitcher was in the game, and for the entire game) and what the
relievers in that game did. For Morris, I also tracked whether he gave up the
lead during the game (and if so, how many times) and whether he allowed the
first run of the game.
(For the years 1991-94, I only had access to boxscores, and occasionally had
to extrapolate the game situation when the starting pitcher left. I couldn't
determine it with certainty in about 10 games, total, in those four seasons.)
Run Support
Jack Morris made his first major league start on July 26, 1977. It was a
bountiful time for the Tigers, who would bring Morris, Lance Parrish, Lou Whitaker and Alan Trammell to the majors within six weeks of each other, and seven years later watch those four form the core of a championship team.
Morris was the best pitcher on that team, and he benefited from having for his teammates such good hitters at up-the-middle positions. The Tigers would be among the AL's best offensive teams throughout Morris' career, and when he left Detroit, he found himself with even better offensive teams:
Year Team Runs AL Rank Park Effect
--------------------------------------------------
1977 Tigers 714 8 +5%
1978 Tigers 714 5 +3%
1979 Tigers 770 5 +5%
1980 Tigers 830 1 +2%
1981* Tigers 427 9 +4%
1982 Tigers 729 7 0%
1983 Tigers 789 4 -3%
1984 Tigers 829 1 -1%
1985 Tigers 729 8 -1%
1986 Tigers 798 3 0%
1987 Tigers 896 1 -4%
1988 Tigers 703 8 -4%
1989 Tigers 617 13 -2%
1990 Tigers 750 2 +1%
1991 Twins 776 4 +5%
1992 Blue Jays 780 2 +5%
1993 Blue Jays 847 2 +2%
1994* Indians 679 1 0%
Note that while Tiger Stadium was perceived to be a very good hitters' park,
it was just barely better than neutral for much of this time. The deep center
field and notoriously high grass canceled the short porch in right field. The
Tigers didn't score a ton of runs because they played in a great hitters'
park; they scored a ton of runs because they could hit.
For his career, Morris pitched for nine good offensive teams, three poor
ones--just one in his last 13 seasons--and six others that were essentially
neutral. The meat of his career, where he made his reputation, was spent
pitching in a neutral environment for good offensive teams.
Just because a pitcher's team has a good offense doesn't necessarily mean that
he'll get good run support. Did Morris, over his career, get better or worse
run support than his teammates did?
Year Team Morris' Support Others' Support
------------------------------------------------------
1977 Tigers 2.79 4.51
1978 Tigers 4.09 4.69
1979 Tigers 5.48 4.68
1980 Tigers 4.79 4.95
1981* Tigers 5.23 3.71
1982 Tigers 4.50 4.63
1983 Tigers 4.76 4.98
1984 Tigers 5.39 5.44
1985 Tigers 4.60 4.51
1986 Tigers 5.72 5.08
1987 Tigers 5.27 5.98
1988 Tigers 4.11 4.65
1989 Tigers 3.49 3.79
1990 Tigers 5.00 4.85
1991 Twins 5.06 4.76
1992 Blue Jays 5.66 4.75
1993 Blue Jays 4.53 5.54
1994* Indians 6.73 5.83
1977-90 4.85 4.76
1991-94 5.44 5.19
Total 527 starts 4.94 4.85
Run Support is defined as "runs scored while the starter is the pitcher of
record," and is presented as per nine innings.
For his career, Morris did not receive significantly more run support than the
other pitchers on his teams. There usually wasn't much difference in
individual season, either, although you can see the impact of the big
run-support years in his won-loss record. Two of Morris' three 20-win seasons
came when he got at least 10% more support than his teammates (1986 and 1992)
did, and he might have won 20 in 1981--when he got almost 40% more run support
than his teammates did--if not for the players' strike.
Morris had two of his best-supported seasons in 1991 and 1992, the years in
which he joined two different teams as a free agent and helped them win
championships, going 39-18 over the two seasons despite adjusted ERAs of 124
and 102, respectively. Much of Morris' Hall of Fame case rests on these two
seasons, when his reputation as a veteran leader was cemented. His success in
those years appears to be a function of good run support, however, along with
his well-documented durability.
Relief
What about bullpen support? Morris threw a ton of complete games, 175 in all,
or one nearly every three times out. He lost almost a third of those--Morris'
reputation as a horse would often cause Sparky Anderson to leave him in when
he was behind. When Morris left a game with a lead, though, did he get the win
more often than his teammates did?
- Morris left a game as the winning pitcher of record 170 times. Of those,
he got the win 140 times, was pinned with two losses and got 28 no-decisions,
for a "saved win" percentage of .824.
- Morris' teammates left games as the winning pitcher of record 773 times.
They got the win 603 times, tagged with losses five times, and had 165
no-decisions, for a "saved win" percentage of .780.
This isn't conclusive. Morris was pitching deeper into games than his
teammates were, collectively, and as such, leaving easier games to close. He
was leaving them to better pitchers as well, but that's a result of his
ability, not luck. The small difference above--equivalent to a win-and-half in
a 35-start season--isn't enough for me to conclude that he received better
relief support than his teammates did for any reason other than he was
throwing enough innings that he left them to the better relievers.
Distributions
So Morris' run support wasn't unusual relative to his teammates, and he was
getting about the same help from his bullpens in converting his performance to
wins. Let's look at this another way. How were the runs Morris received
distributed, and how did he perform at each level of support?
(This approach is borrowed from The Bill James Baseball Abstract 1988,
p. 153)
Team runs Times % Morris Times % Teammates
----------------------------------------------------------------------
0 24 4.6% 0-24 (.000) 105 4.6% 0-98 (.000)
1 58 11.0% 6-49 (.109) 212 9.3% 21-179 (.105)
2 57 10.8% 11-37 (.229) 281 12.3% 50-180 (.217)
3 70 13.3% 24-37 (.393) 305 13.4% 93-142 (.396)
4 66 12.5% 31-18 (.633) 323 14.2% 116-103 (.530)
5 67 12.7% 40-10 (.800) 266 11.7% 111-56 (.665)
6 52 9.9% 38-6 (.864) 206 9.0% 104-17 (.860)
7 35 6.6% 20-1 (.952) 174 7.6% 93-18 (.838)
8 31 5.9% 21-1 (.955) 120 5.3% 85-5 (.944)
9+ 67 12.7% 58-1 (.983) 290 12.7% 207-7 (.967)
[Percentages, rounding, yadda, yadda, yadda...]
We might have gleaned this from the run-support table, but this chart confirms
it: Morris got essentially the same run support his teammates did. He did a
better job than they did of turning any level of run support into a W, save
the three-run level, where it was essentially a wash.
One popular argument in support of pitchers like Morris is that their
teammates can relax when they're on the mound, and that this shows up in their
run support. Without making any extrapolation to other pitchers, I think the
data refutes the argument that Morris' presence on the mound had any effect on
the lineup. He pitched for good-hitting teams who scored whether it was Morris
or Dan Petry or Steve Searcy on the mound.
Take a look at that chart again. Given five runs--Morris' team scored five runs
in nearly half his career starts-Morris was nearly a lock, going 177-19.
Ninety-nine of Morris' 254 career wins came in games in which his team scored
at least seven runs. Not to disparage that-the run support isn't his fault--but
if Random Teammate can win 65.9% of his starts with that kind of run support,
Morris winning 74.4% of his starts at that level doesn't look so impressive.
So much of Morris' Hall of Fame case rests on his win totals for good
offensive teams that I think it's important to look at how he got those wins.
It is instructive to look at the figures above in the context of the American
League during Morris' career. Here are the percentages from the above chart,
along with the percentage of times AL teams scored that many runs in the years
1977-1994.
All others,
Runs Morris' starts American League, 1977-94
-----------------------------------------------------
0 4.6% 6.1%
1 11.0% 10.7%
2 10.8% 13.7%
3 13.3% 14.5%
4 12.5% 13.4%
5 12.7% 11.5%
6 9.9% 9.1%
7 6.6% 6.9%
8 5.9% 4.8%
9+ 12.7% 9.3%
Looked at another way:
All others,
Runs Morris' starts American League, 1977-94
-------------------------------------------------------------
0 100.0% 100.0%
At least 1 95.4% 93.9%
At least 2 84.4% 83.2%
At least 3 73.6% 69.5%
At least 4 60.3% 55.0%
At least 5 47.8% 41.6%
At least 6 35.1% 30.1%
At least 7 25.2% 21.0%
At least 8 18.6% 14.1%
At least 9 12.7% 9.3%
Without oversimplifying too much, you can see that Morris got much more of his
support in the five-runs-and-above range than his AL peers, the range where a
starter often logs a win irrespective of his performance. He was more likely
to have his offense reach this level, while other AL pitchers found themselves
working with two, three and four runs, and picking up losses or no-decisions.
This chart illustrates the basic point about Morris: he won a lot of games
because he pitched in front of good offenses. He was much better supported
than the average AL pitcher of his time, and it was that run support that
enabled him to rack up wins despite his not being a top-notch run preventer.
Same concept from the other direction: how were the runs Morris allowed
distributed, and how did his record shake out?
Allowed Times % Morris Times % Teammates
---------------------------------------------------------------------
0 41 7.8% 37-0 (1.000) 174 7.6% 152-0(1.000)
1 69 13.1% 60-3 (.952) 305 13.4% 221-20 (.917)
2 97 18.4% 61-17 (.782) 400 17.6% 217-72 (.751)
3 79 15.0% 39-25 (.609) 413 18.3% 150-118 (.560)
4 82 15.6% 25-36 (.410) 414 18.2% 77-189 (.289)
5 70 13.3% 15-39 (.278) 307 13.5% 35-187 (.158)
6 43 8.2% 12-25 (.324) 172 7.5% 12-137 (.081)
7 28 5.3% 1-21 (.045) 58 2.5% 3-53 (.054)
8 17 3.2% 1-15 (.063) 27 1.2% 2-22 (.083)
9+ 2 0.4% 0-2 (.000) 9 0.03% 0-8 (.000)
This approach is designed to address the specific point, which I've heard from
a number of people, that Morris allowed a disproportionate number of runs in
"disaster starts," where the last few marginal runs allowed hurt his ERA but
didn't affect his team's chance of winning a game that was already long lost.
There may be something to the notion. If you look at the percentages above,
you see that Morris and his teammates had roughly the same percentage of
starts in which they allowed two or fewer runs: 39.2% for Morris, 38.6% for
his teammates. At the next two run levels, though, Morris' teammates have
significantly more starts, a gap Morris makes up at six runs allowed and
above.
This is the statistical representation of something I mentioned earlier. Where
other pitchers were being pulled from games at the usual point-for much of
Morris' career, the general vicinity of four runs--Morris was being allowed to
stay in to give up more runs. Sometimes this led to games in which he'd give
up six or more runs. He got some wins out of this--14 in 90 career starts in
which he gave up six runs, as opposed to 17 in 266 for his teammates-at a cost
to his rate stats.
How big a difference does it make? You can arrive at any number of answers
depending on the assumptions you make. The extra wins Morris got from staying
in games longer are as tangible as the extra points on his ERA from the
innings he threw when he wasn't pitching well.
Pitching to the Score
That brings us to perhaps the core question of the Jack Morris Project: was
Morris' career ERA of 3.90 so high because he was a 3.90 ERA pitcher, or
because he was more concerned with innings and wins, and allowed a
disproportionate number of runs when those runs wouldn't hurt his teams?
To answer that, I tracked Morris' performance at every score, or more
accurately, the score when he took the mound to start the inning. Top of the
first: tie game. Down 3-1 in the sixth: that's -2. Whatever Morris did in that
inning is categorized based on what the score was when the inning started.
If Morris was pitching to the score... actually, let's just run the chart and
then discuss it.
IP R RA
-----------------------------
More 82 2/3 38 4.14
+7 43 2/3 19 3.92
+6 76 2/3 42 4.93
+5 123 2/3 63 4.58
+4 186 2/3 83 4.00
+3 290 2/3 150 4.64
+2 377 165 3.94
+1 549 255 4.18
0 1224 1/3 540 3.97
-1 396 206 4.68
-2 216 1/3 105 4.37
-3 111 2/3 66 5.32
-4 49 29 5.33
-5 15 14 8.40
-6 3 4 12.00
-7 1 0 0.00
Less 0 0 N/A
TOT. 3746 1/3 1779 4.27
Is this the run distribution of a pitcher who was "pitching to the score"?
Since we haven't really seen that animal in captivity, we have to guess, but I
would think we'd see a broader distribution of RAs, with a valley in the +1
to -1 range, and higher RAs as the pitcher had more runs to work with. Now, if
you squint, you can see that valley, although it's more from 0 to +2 than
centered around 0. However, Morris was pretty bad when trailing by a run,
suggesting that he was good at turning small deficits into big ones. He also
didn't show a big jump in his RA when he was ahead by a lot of runs, and in
fact, his RA when he had at least a seven-run cushion is better than his
career mark. Morris wasn't turning a lot of 8-0 games into 8-4 games.
Playing with the data a little, we get:
Morris ahead: 4.24
Morris tied: 3.97
Morris behind: 4.82
This isn't surprising, as to be behind, Morris would have had to have given up
at least one run, a pretty good first indicator that he was going to give up
more. The sharply rising ERAs as Morris falls behind follow from this; if
Morris started an inning down multiple runs, he probably didn't have it that
day.
Another way of looking at it:
Difference IP R RA
----------------------------------
>5 207 103 4.48
5 138 2/3 77 5.00
4 235 2/3 112 4.28
3 402 1/3 216 4.83
2 593 1/3 270 4.10
1 945 461 4.39
0 1224 1/3 540 3.97
Does this chart reveal or conceal? By folding the data from ahead and behind
together to get a sense of whether Morris pitches to the score, we get a set
of almost random data points. In general, Morris seems to have better numbers
in close games, but this chart is hiding information we already have, that
Morris' performance in blowouts in which he's ahead isn't much different from
what he does when he's ahead by just a few runs. His RAs in games in which the
difference is five or more runs is inflated by those outings where he's being
blown out.
Looking at the year-by-year data isn't particularly informative. Morris has a
couple of seasons that look like they could be examples of "pitching to the
score," but the evidence is weak, and if you have enough pitcher-seasons,
you're bound to have some that fit whatever pattern you're looking for.
Here's 1979, for example:
IP R RA
-----------------------------
More 6 4 6.00
+7 0 0 N/A
+6 3 3 9.00
+5 9 1/3 7 6.75
+4 16 2/3 5 2.70
+3 19 1/3 8 3.72
+2 29 3 0.93
+1 19 6 2.84
0 59 1/3 27 4.10
-1 15 7 4.20
-2 12 2 1.50
-3 9 4 4.00
-4 0 0 N/A
-5 0 0 N/A
-6 0 0 N/A
-7 0 0 N/A
Less 0 0 N/A
TOT. 197 2/3 76 3.46
Morris pitched very well when protecting a small lead, and gave up runs when
he had a larger cushion. However, that 4.10 RA in tie games is well above his
seasonal mark. Morris showed a similar pattern in 1980. With a seasonal RA of
4.50, he had an RA when pitching with a one- or two-run lead of 2.98.
If these figures represented a skill, though, and Morris was actually allowing
runs based on the score of the game, then you would expect him to show it
throughout his career. He didn't. Instead, he had seasons like 1983:
IP R RA
-----------------------------
More 10 2 1.80
+7 6 3 4.50
+6 9 3 3.00
+5 20 6 2.70
+4 13 1/3 2 1.35
+3 16 4 2.25
+2 24 8 3.00
+1 37 20 4.86
0 84 39 4.18
-1 24 12 4.50
-2 20 1/3 9 3.98
-3 22 7 2.86
-4 7 2 2.57
-5 1 0 0.00
-6 0 0 N/A
-7 0 0 N/A
Less 0 0 N/A
TOT. 293 2/3 117 3.59
That theoretical "valley"? It's a peak here, as Morris was at his worst in
one-run and tied games. In fact, that looks like what you might find in a
theoretical "unclutch" pitcher, someone who chokes when the game is close.
Morris won 20 games in 1983, largely by throwing 293 2/3 innings for a team
that was fourth in the AL in runs scored.
I can find no pattern in when Jack Morris allowed runs. If he pitched to the
score-and I don't doubt that he changed his approach-the practice didn't show
up in his performance record.
I tracked a couple of extra categories for Morris, things that I thought might
fit someone "pitching to the score." Morris made 527 starts, and in 235 of
them, 44.6%, he gave up the first run of the game. He had 41 career starts in
which he allowed no runs, so unless he was perfect that day Morris was as
likely as not to put the Tigers behind.
Setting aside those games in which he allowed the first run, Morris gave up a
Tiger lead in another 109 starts. [Note: the definition of "blowing a lead" is
extremely generous. Morris had to be on the mound when the go-ahead run
scored. This excludes all leads blown by relievers, even if the runs scoring
were charged to Morris.]
Of everything I've presented here, I believe this is the one point that best
refutes the arguments for Jack Morris as a Hall of Famer. We know his raw
numbers don't stack up, and we know he has some bonus markers-a no-hitter,
Game Seven of the 1991 World Series, three other championship rings. What we
now know is that instead of "pitching to the score," as his supporters claim
he did, Morris actually put his team behind in 344 of his 527 career starts.
All told, Morris blew 136 leads in 527 starts, or about one every four times
out, and that's using a generous definition of "blown lead." Take this with a
grain of salt, but having gone through the man's career, I wish I had tracked
the number of times Morris turned a lead into a tie. He would quite often turn
2-0 into 2-2, then 4-2 into 4-4, before leaving with a 5-4 lead. If I lose my
mind at some point, perhaps I'll go back through his career and track those
occurrences.
Conclusion
As I said, I don't know what the performance record of someone who had
successfully pitched to the score would look like. I am certain, though, that
for a pitcher to build his Hall of Fame case on the notion that he did such a
thing, he couldn't have put his team behind in nearly two-thirds of his career
starts, and he couldn't have blown leads once a month throughout his career.
Jack Morris was a very good pitcher whose primary skill was durability. He
benefited from coming up with a number of good players, players who would form
the core of a good offense that scored lots of runs for him. He happened to
have a career in a down period for starting pitchers, so he stands out among
his peers more than someone with his performance record would in the 1970s or
1990s.
He's not a Hall of Famer. As much as I loved watching Game Seven in 1991, and
as much as I think the man got cheated by collusion in 1986, he's not a Hall
of Famer.
* * * * *
I couldn't have done this without Retrosheet. I
acknowledge the help of everyone who works on that project, and particularly
David Smith for his yeoman work for baseball fans and researchers everywhere.
The play-by-play information used here was obtained free of charge from and
is copyrighted by Retrosheet. Interested parties may contact Retrosheet at 20
Sunset Rd., Newark, DE 19711.
Joe Sheehan is an author of Baseball Prospectus.
You can contact Joe by clicking here or click here to see Joe's other articles.
|
Under The Knife (04/23)
Team Health Reports: Washington Nationals