Measuring team efficiency 
By Tom Tippett
December 5, 2002
It goes without saying that wins and losses are the most important things
to consider when judging a team's performance. They are, after all, what
the game is all about and what determines who gets to keep playing until
there's only one winner left.
The next most important things are runs scored and runs allowed. You
win games by outscoring your opponents, so the connection between runs
and wins is very strong. It's not perfect, though, and every season produces
a few teams that win more or less than you'd expect given their run differential.
If runs are one step removed from wins, then the baseball events that
produce runs are two steps removed from wins. You score runs by putting
together singles and walks and doubles and steals and homers, and you
prevent runs by holding the other team to a minimum of those things.
In most cases, there's a very direct relationship between wins and runs
and the underlying events that produce runs. But that's not always the
case, and in this review of the 2002 season, we'll identify teams where
those relationships didn't hold up. If the past is any guide, this will
give us some very strong hints about what is likely to happen with those
teams in the future.
To explore the relationship between runs and wins, we'll use the pythagorean
method that was developed by Bill James. To explore the relationship between
offensive events and runs, I'll introduce a new statistic that I'll call
the run efficiency average. This number will tell us which teams
were unusually good at turning offensive events into runs and unusually
good at keeping the other team from doing the same.
We'll end up with three measures for each team -- one for offensive efficiency,
one for defensive efficiency, and one for pythagorean efficiency -- that
will tell us which teams squeezed more wins out of the hits and walks
and homers and other events that occurred during their games. And which
teams squandered their output to the greatest degree.
And we'll take a look at some history. We'll see that teams that are
unusually efficient (or ineffecient) have exhibited a very strong tendency
to revert back to the norm the next year. In other words, if your team
was especially inefficient in 2002, there is every reason to believe things
will be better next year. And the opposite is true, too. If your team
was very efficient this year, don't count on a repeat performance next
year.
That's good news for the Cubs, Brewers, Devil Rays and Tigers. And bad
news for the Angels, Braves, and Twins. It's way too early to start predicting
what's going to happen in 2003, and all thirty teams are quite capable
of improving or regressing based on their off-season moves and the development
of their younger players and prospects. But we can say that these seven
teams (and a few others to a lesser degree) go into the offseason in better
or worse condition than it might seem based solely on their 2002 win-loss
records.
Converting runs into wins
Others, notably Rob Neyer and the Baseball Prospectus crew, have written
extensively on ESPN.com about the Bill James pythagorean method, a well-established
formula that says that a team's winning percentage is tightly coupled
with runs scored and runs allowed. The expanded standings on ESPN.com
include run margins and expected win-loss records that are derived using
this formula, and Rob's home page showed pythagorean standings every day.
So I'm not going to go over that ground again.
I will, however, try to put the 2002 results into historical context.
For instance, the Red Sox and Cubs won 8 fewer games than their run margin
would normally produce, while three teams (Minnesota +7, Oakland +6, and
Detroit +6) won at least six more than expected. How unusual is this?
And what tends to happen to teams that stray from their expected win totals?
I started by computing the expected and actual win totals for every team
since 1962, the first year the 162-game schedule was used in both leagues.
The more games you play, the larger the differences between expected and
actual wins, so I didn't want to mix seasons with different schedules.
For that reason, I left out the strike-shortened 1972, 1981, 1994, and
1995 campaigns, leaving a total of 72 league-seasons.
In those 72 league-seasons, the average team "missed" its pythagorean
projection by only 3.2 games, indicating that there is indeed a very strong
relationship between runs and wins. How does 2002 compare? This year,
AL teams were off by an average of 4.1 games while their NL counterparts
missed by an average of 3.0 games. Overall, the 30 teams had an average
difference of 3.5 games, slightly higher than the historical average but
well within the normal year-to-year fluctuations.
The wackiest season, in the pythagorean sense, was the 1978 National
League, whose teams missed their projections by an average of 5.3 games.
In the NL West that year, the Dodgers led the league in both scoring and
fewest runs allowed, outscoring their opponents by 154 runs, while the
Reds were only +22 on run differential. But Cincinnati won nine more games
than expected and the Dodgers five fewer, turning what could have been
a runaway win by the Dodgers into a close battle that saw LA win by 2-1/2
games. In the NL East, the story was much the same, as the Phillies (-6)
edged the Pirates (+2) by two games in a race that was much closer than
it could have been. That same year, the Expos matched the Reds with a
run margin of +22, but Cincinnati (+9) won 92 games and Montreal (-8)
only 76. (The Expos run margin foreshadowed their improvement; they went
on to win 95 games in 1979.)
In contrast, the 1991 American League was closest to pythagorean form,
with an average difference of only 1.8 wins.
What tends to happen to teams with large pythagorean differences? Here's
a list of the 22 teams that have exceeded their projected win total by
at least 8 games, along with their differences in the next year:
Team Diff Next
--------------- ---- ----
1974 Padres +12 + 7
1984 Mets +12 + 1
1970 Reds +11 - 3
1963 Astros +10 + 4
1997 Giants +10 - 3
1998 Royals +10 -11
1970 Phillies + 9 + 3
1977 Orioles + 9 + 6
1978 Reds + 9 - 1
1978 Athletics + 9 + 4
2001 Mets + 9 - 4
1971 Braves + 8 + 5
1973 Tigers + 8 + 8
1974 Tigers + 8 + 2
1976 Cubs + 8 + 5
1977 Mariners + 8 0
1979 Astros + 8 + 5
1982 Giants + 8 - 1
1987 Expos + 8 - 5
1989 Astros + 8 + 5
1992 Astros + 8 - 6
1997 Reds + 8 - 3
As you can see, only a few teams came close to matching their pythagorean
differences in the next season. In fact, these 22 teams were collectively
18 wins above their projection the year after, an average of less than
one win per team.
(Just to be clear, these next-year numbers don't represent the change
in actual win-loss record from the year before, so they don't measure
the team's tendency to get better or worse. They represent the difference
between actual and pythagorean wins the next season. In other words, they
measure the tendency to consistently win more or fewer games than the
run margin suggests, not the tendency to produce a better or worse run
margin in the first place.)
On the flip side, here are the teams with the biggest negative differences
since 1962:
Team Diff Next
--------------- ---- ----
1962 Mets - 8 + 3
1968 Pirates - 8 - 2
1974 Angels - 8 + 3
1975 Dodgers - 8 + 2
1978 Expos - 8 0
1983 Cubs - 8 + 4
1983 Rangers - 8 - 5
1985 Indians - 8 + 4
1986 Giants - 8 - 4
2000 Astros - 8 + 4
1962 Cardinals - 9 - 2
1963 Twins - 9 - 9
1964 Twins - 9 + 1
1966 Yankees - 9 + 5
1974 Athletics - 9 - 1
1975 Yankees - 9 - 1
1980 Brewers - 9 + 4
1984 Astros - 9 0
1985 Red Sox - 9 + 4
1990 Mets - 9 - 3
1980 Cardinals -10 + 3 (103 games in 1981)
1993 Padres -10 - 5
1997 Astros -10 - 6
2001 Rockies -10 + 4
1970 Cubs -10 + 3
1975 Astros -11 + 3
1999 Royals -11 + 1
1967 Orioles -12 - 2
1984 Pirates -13 - 6
1986 Pirates -13 + 1
1993 Mets -14 + 1
These 31 teams were collectively 4 wins above their projection the year
after, about as close to zero wins per team as you can get.
These extreme teams do leave us with a few unanswered questions. Why
are there more NL teams than AL teams on these lists? Why do the Astros
show up as often as they do? How miserable must the mid-1980s Pirates
fans have been when their team posted a three-year pythagorean difference
of -32 wins from 1984 to 1986? These answers, if they exist, will have
to wait for another day.
I'm not going to suggest that I have proven this beyond a reasonable
doubt, but I believe luck plays a large part. If you wanted to argue that
the over-achievers had big pluses because their manager was especially
astute or their roster was full of clutch players, it would be a tough
case to make based on the next-year records of these teams. And looking
at the under-achievers, it would be even tougher to argue that their manager
and players are fundamentally flawed based on their next-year results.
But teams change from year to year, and the under-achievers are much
more likely to fire their managers and turn over half their rosters. Perhaps
those changes were responsible for bringing them back to pythagorean normalcy.
Even though I don't believe this argument would hold up under closer examination,
it muddies the water a little.
Still, if managerial skill and clutch performance were the biggest piece
of this puzzle, why wouldn't the over-achievers, the teams that would
not be making many changes from year to year, be able to maintain
their performance to a much greater extent?
Converting offensive events into runs
In the previous section, we took one step back from wins and losses to
examine runs. In this section, we'll take another step back and look at
the offensive events -- the hits and walks that lead to the runs that
generate the wins -- that were produced and allowed by each team.
Just as there is a strong relationship between runs and wins, it's almost
always true that the more hits and walks you produce, the more runs you'll
score. Sometimes a productive team comes up short on the scoreboard because
they didn't hit in the clutch or just because they happened to hit line
drives right at people in key situations. Or the opposite could be true.
But this relationship holds up most of the time.
To shed some light on this relationship, we need a way to take batting
stats and turn them into a measure of overall offensive production. There
are several good options here, including Runs Created (Bill James), Batting
Runs (Pete Palmer), Equivalent Average (Clay Davenport), and OPS (on-base
average plus slugging average). But many of them require a computer, and
although we do computer analysis all the time, we also like to use simpler
measures that anyone can use whenever they have a page of stats in front
of them. The best of these simple methods give up very little accuracy
in return for a big gain in usability.
For this exercise, I'll use the sum of total bases and walks, or TBW
for short. TBW is not a perfect measure, but it does have a few things
going for it. It captures the most important things a team does to produce
runs -- singles, extra-base hits, and walks. It's easy to figure without
a computer. In the past, I've used both TBW and OPS for this type of analysis,
and the results are almost exactly the same, so the accuracy is more than
acceptable.
And sometimes it just seems to tell a story more clearly. For instance,
the 2002 Yankees had a team OPS of .809 compared to the .769 mark of the
Mariners. Even though I've been working with OPS figures for a number
of years, I still need to stop and think about what a 40-point advantage
means. But if you tell me that the Yankees produced 224 more total bases
and walks than the Mariners, that's something I can grasp right away.
The following table shows the offensive and defensive TBW figures for
the American League, along with the difference between these two figures
and each team's league rank based on those differences. It also shows
runs for and against, the run differential, and the rankings based on
run differential. Finally, because we're trying to trace a path from TBW
to runs to wins, it also lists the team's win total for the year.
---------- TBW ---------- ------- Runs --------
AL Off Def Diff Rank Off Def Diff Rank W
NY 3187 2629 +558 1 897 697 +200 2 103
Bos 3050 2532 +518 2 859 665 +194 3 93
Tor 2921 2995 - 74 8 813 828 - 15 8 78
Bal 2665 2981 -316 11 667 773 -106 11 67
Tam 2640 3219 -579 13 673 918 -245 13 55
Min 2911 2789 +122 7 768 712 + 56 7 94
Chi 3028 2839 +189 6 856 798 + 58 6 81
Cle 2774 2930 -156 10 739 837 - 98 10 74
KC 2728 3189 -461 12 737 891 -154 12 62
Det 2414 3006 -592 14 575 864 -289 14 55
Oak 3009 2595 +414 3 800 654 +146 4 103
Ana 2918 2647 +271 4 851 644 +207 1 99
Sea 2963 2710 +253 5 814 699 +115 5 93
Tex 3112 3205 - 93 9 843 882 - 39 9 72
As you can see, the team rankings using TBW and those using run differentials
are very similar. In fact, they're identical except for Anaheim's move
from fourth in TBW to first in run margin. The Angels were very efficient
on both sides of the ball, finishing 4th in scoring (and only 8 runs out
of second) despite trailing six other teams in TBW, and leading the league
in fewest runs allowed even though three other teams gave up fewer TBW.
(That efficiency didn't carry over to the relationship of runs to wins,
however, as they led the league in run margin but were only third in wins.)
In terms of raw production, the Red Sox nearly matched the Yankees, but
still managed to come up ten short in the win column. (The same thing
happened in 2001.) This comes as no surprise to the long-suffering Boston
fans or the incredibly smug New Yorkers who just knew the Sox would find
a way to lose despite all their talent.
It's interesting to note that the White Sox were a match for the Twins
in production even though Minnesota ran away with the division. For all
the talk about the Twins superior pitching and defense and the problems
the White Sox had in those areas, Chicago gave up only 50 more TBW, roughly
one base every three games.
And we see yet another example of how strong the AL West was this year,
with three teams in the league's top five in TBW and run differentials
and the Rangers only a little below the league average. Oakland was a
clear winner in TBW but trailed the amazing Angels in run margin, and
it took an excellent 32-14 record in one-run games to keep the A's in
first place.
Let's take a quick look at the National League before pausing to put
these TBW numbers in historical context.
---------- TBW ---------- ------- Runs --------
NL Off Def Diff Rank Off Def Diff Rank W
Atl 2808 2529 +279 3 708 565 +143 3 101
Mon 2865 2857 + 8 9 735 718 + 17 7 83
Phi 2970 2769 +201 5 710 724 - 14 9 80
Flo 2810 2927 -117 12 699 763 - 64 11 79
NY 2657 2758 -101 10 690 703 - 13 8 75
StL 2879 2630 +249 4 787 648 +139 4 97
Hou 2886 2793 + 93 6 749 695 + 54 6 84
Cin 2815 2928 -113 11 709 774 - 65 12 78
Pit 2566 2860 -294 13 641 730 - 89 13 72
Chi 2853 2814 + 39 8 706 759 - 53 10 67
Mil 2613 3086 -473 16 627 821 -194 16 56
Ari 2974 2608 +366 2 819 674 +145 2 98
SF 3045 2524 +521 1 783 616 +167 1 95
LA 2701 2653 + 48 7 713 643 + 70 5 92
Col 2826 3195 -369 14 778 898 -120 14 73
SD 2649 3040 -391 15 662 815 -153 15 66
While it's clear that Atlanta was the division's top team, their TBW
differential wasn't much better than that of the Phillies, who somehow
managed to turn a big edge in raw production into a negative run differential
and a losing season. Most of the problem was on offense, where the Phils
were 3rd in TBW but only 8th in runs scored. (Before the season, our computer
simulations had the Phillies finishing a close second behind the Braves.
In the real season, they were a very close second statistically,
but that didn't translate into the things that really matter, runs and
wins.)
The biggest surprise in the Central division was the Cubs. In fact, most
of what I just wrote about Philly applies here, too. Our preseason simulations
put Chicago third with a .500 record, and the real Cubs put up TBW numbers
that were entirely consistent with being a .500 team. But they ranked
a few places lower in runs than in TBW on both sides of the ball and they
couldn't win the close games (18-36 in contests decided by one run). By
the way, the 2001 Cubs were the division's best team statistically (+175
TBW) but failed to win the pennant; with two straight seasons like this,
it's no surprise that a managerial change was made, regardless of whether
the manager was to blame.
In the West, San Francisco outproduced Arizona but came up a little short
in the standings. Both teams were very strong across the board, however,
and the Giants showed during the second season that they really were the
best team in the league. Statistically speaking, Los Angeles was much
closer to a .500 team than their 92-70 record suggests. In fact, the Dodgers
were the anti-Phillies, turning a 12th-place ranking in offensive TBW
into a 7th-place finish in scoring. (Warning to LA fans: the Padres had
the most efficient offense in baseball in 2001 -- 13th in OPS, 6th in
runs -- and look what happened to them in 2002.)
A little TBW history
I've been putting these tables together for a few years now, and I can
tell you that TBW differentials are usually in the plus or minus 300 range.
With eight teams more than 400 from the midpoint this year, I wondered
how these figures stacked up against other teams from the past. Thanks
to Retrosheet's database of play-by-play accounts, I ran the numbers for
all seasons back to 1974. (It would be nice to go back further, but the
official stats don't include doubles and triples allowed by pitchers.)
Here are the top twenty teams from that 29-year period:
Team Net TBW Comment
---------------- ------- --------------------------------
1998 Braves + 664
1998 Yankees + 662 Won WS
2001 Mariners + 603
1997 Braves + 568
2002 Yankees + 558
1976 Reds + 550 Won WS
1995 Indians + 536 144-game schedule (+603 per 162)
2001 Athletics + 534
1984 Tigers + 530 Won WS
1988 Mets + 524
2002 Giants + 521
2002 Red Sox + 518 Failed to qualify for postseason
1974 Dodgers + 509
1978 Brewers + 493
1996 Indians + 471
1986 Mets + 466 Won WS
1999 Yankees + 454 Won WS
1996 Braves + 449
1977 Dodgers + 448
1979 Orioles + 448
Notes: Only one of these strong teams, the 2002 Red Sox, failed
to make the postseason ... only five won the World Series, a reminder
that surviving the expanded postseason format is very tough ... could
have been a great 1998 World Series if the Braves hadn't lost to the Padres
... the 1978 Brewers also had the best run margin in the AL that year,
so this could have been one of the great three-way pennant races in history.
And the bottom twenty teams:
Team Net TBW Comment
---------------- ------- --------------------------------
1996 Tigers - 727
1979 Athletics - 683
1998 Marlins - 676 Defending WS champs
1974 Padres - 630 Expansion
2002 Tigers - 592
1978 Blue Jays - 590 Expansion
1979 Blue Jays - 586 Expansion
2002 Devil Rays - 579 Expansion
1977 Braves - 539
1977 Mariners - 528 Expansion
1980 Mariners - 522 Expansion
1977 Blue Jays - 507 Expansion
1993 Rockies - 487 Expansion
1978 Mariners - 486 Expansion
1999 Twins - 484
2001 Devil Rays - 475 Expansion
2002 Brewers - 473
2002 Royals - 461
1989 Tigers - 447
1982 Athletics - 443
Notes: Eleven of these twenty teams were expansion franchises
in the first seven years of their existence ... four 2002 teams made this
list ... maybe two rounds of expansion since 1993 is the reason so many
recent teams made both lists ... Billy Martin took over as manager of
the A's after their disastrous 1979 season, led them to a winning record
in both 1980 and 1981, then made this list again in 1982 before being
fired ... the Twins have come a long way since 1999 ... in 1982, Bill
James wrote that the Blue Jays might be the worst expansion team in history,
but they got better in a hurry after that, so maybe there's reason for
Devil Rays fans to have some hope as their young prospects move up.
Run Efficiency Average
Earlier in this article, when discussing the relationship between runs
and wins, we saw that teams sometimes win quite a few more or less games
than their run margin would normally produce. And that those differences
don't tend to repeat the next year. It's very rare for a team to over-achieve
(or fall short) two years in a row, and there's a very strong tendency
to revert to a normal runs-to-wins relationship. Is this also true of
TBW and runs?
To identify teams with particularly efficient or inefficient offenses,
ones that produce more or less than the expected number of runs given
the TBW they produced, I divided runs by TBW to get something I'll call
the run efficiency average (REA). As you can see in the following
chart, which plots TBW versus runs scored for every full team season since
1974, there's a very strong straight-line relationship between TBW and
runs. In other words, we can predict runs scored from TBW with a high
degree of accuracy.
It turns out that run efficiency averages look an awful lot like team
batting averages. From 1974 to 2002, team batting averages ranged from
a low of .229 to a high of .294 with a midpoint of .261. Baseball fans
know from experience that a team batting average of .280 or higher is
very good, and one below the .245 mark is woeful.
In this time period, run effeciency averages have ranged from .225 to
.305 with a midpoint of .264. The midpoint and the spread are slightly
higher than for team batting averages, but the benchmarks are basically
the same. Anything over .280 indicates a very efficient offense, while
anything under .245 indicates a team that squandered a lot of its chances.
Like team batting averages, run efficiency averages tend to be higher
in the American League (because pitchers don't bat) and rise and fall
by a few points from season to season. (They also appear to be higher
in good hitters parks, but I'm leaving park effects out of the equation
for the time being. I'll be looking at both offense and defense, so the
park effects should cancel out when we subtract one from the other.)
So the best way to evaluate teams is to compare their run efficiency
averages to the norm for their league that season and to rank them based
on those differences. Here are the offenses that were most efficient in
the 1974 to 2002 period, relative to their leagues, and what they did
the following year:
Team REA Diff Next
------------------ ---- ------ ------
2000 Rockies .305 + .035 + .019
1996 Rockies .303 + .032 + .020
1977 Twins .298 + .032 - .010
1987 Cardinals .293 + .030 - .008
1975 Reds .287 + .029 + .019
1985 Cardinals .281 + .026 + .002
1982 Brewers .288 + .026 + .005
1992 Brewers .287 + .025 + .008
1976 Phillies .282 + .025 + .017
2000 White Sox .301 + .025 - .003
2001 Mariners .295 + .024 + .005
1999 Indians .299 + .024 + .006
1998 Yankees .297 + .023 + .003
1974 Dodgers .283 + .023 - .008
1993 Rangers .292 + .022 + .003
1991 Brewers .287 + .022 + .025
2000 Royals .299 + .022 - .002
1981 Brewers .280 + .022 + .026
1981 Phillies .273 + .021 + .001
2002 Angels .292 + .021 ???
A majority of these teams were above average again the next year, but
all but one made a move back toward the middle of the pack. On average,
they lost 19 points relative to the league. On a base of 2800 TBW, that's
a loss of 53 runs, enough to cost a team five to six wins.
Here are the least efficient offenses of this period:
Team REA Diff Next
------------------ ---- ------ ------
1988 Orioles .226 - .038 + .003
1998 Devil Rays .238 - .036 - .003
2002 Tigers .238 - .032 ???
1983 Mariners .234 - .032 - .005
1982 Reds .228 - .031 - .007
1981 Blue Jays .227 - .031 - .006
1974 Padres .230 - .029 - .020
1978 Athletics .231 - .028 - .026
1981 Mets .225 - .027 - .004
1996 Phillies .244 - .027 - .010
1978 Athletics .246 - .026 - .005
1985 Rangers .240 - .026 + .002
1993 Marlins .243 - .025 - .019
1976 Expos .233 - .025 - .019
1980 Blue Jays .242 - .024 - .031
2001 Mets .242 - .023 + .003
1996 Angels .256 - .023 + .011
1980 White Sox .243 - .023 + .019
1995 Blue Jays .252 - .023 - .013
1989 Indians .242 - .022 + .021
Again, all but one team moved up the next year, with an average improvement
of 22 points. It's abundantly clear that extreme REA values don't repeat
themselves; no matter what the environment, and no matter how good or
bad the team, the REA tends to make a big move toward the norm the next
year.
This is good news for teams that were the most inefficient this year.
Detroit, Baltimore, Philadelphia, Milwaukee, and Tampa Bay can expect
to improve their efficiency in 2003. Of course, it's bad news for this
year's over-achievers, namely Anaheim, Arizona, Colorado, St. Louis, and
the White Sox.
I won't take the space to show top-20 lists for pitching efficiency,
but I can tell you that the same pattern held on the other side of the
ball. The twenty most efficient pitching staffs moved an average of 21
points toward the norm the next year, while the least efficient improved
by 22 points. Not a single team on either list moved further away.
The five most efficient pitching staffs this year, and the five most
likely to struggle to match that performance, were Atlanta (which had
the lowest run efficiency average in this 29-year period), Anaheim, Oakland,
Minnesota, and Los Angeles. On the other hand, improvement is bound to
be in store for Colorado, Detroit, Cleveland, Tampa Bay, and the Cubs.
By the way, Detroit was very inefficient on both offense and defense
this year, and while their park might have something to do with that,
I don't think it's a major factor. If the Tigers move 20 points toward
the norm on both sides of the ball, they're looking at a favorable swing
of 108 runs, or about 11 wins, even if nothing else changes. (Of course,
moving from 54 wins to 65 wins isn't anything to write home about. They
need to do even better than that.)
In contrast, Anaheim was on both top-five lists for 2002, and they stand
to move back toward the pack offensively and defensively in 2003. That
could take a 99-win team and bring them back to the high 80s.
Converting efficiency into wins
Let's try to wrap all of this up into one neat package. We started by
showing that runs scored and runs allowed are an accurate predictor of
wins and losses. Teams that deviate from this prediction usually revert
to form the next year.
Then we showed that offensive production (as measured by total bases
plus walks) is an accurate predictor of runs scored. Likewise for defensive
production and runs allowed. For both offense and defense, teams that
deviate from the predicted number of runs tend to move significantly toward
the norm the next year.
In other words, these three forms of efficiency -- which I'll call pythagorean
efficiency (turning runs into wins), offensive efficiency (turning TBW
into runs scored), and defensive efficiency (limiting runs allowed per
TBW allowed) -- can have a major impact on the standings in any one season.
But that effect isn't likely to carry over to the next year.
Pythagorean efficiency is already expressed in wins and losses. I'll
translate offensive and defensive efficiency into wins by taking the surplus
or deficit in runs and dividing by nine. Why nine? According to the pythagorean
method, that's the number of runs it takes to add one win in a league
where the average team scores about 750 runs. By converting all three
types of efficiency to wins, we can add them up to see which teams gained
or lost the most due to efficiency in 2002. Here are the figures for all
thirty teams:
Actual ----- Efficiency ----- Adj
Wins Pyth Off Def Tot Win
New York 103 +3 +4 +1 + 8 95
Boston 93 -8 +4 +2 - 2 95
Toronto 78 -2 +3 -2 - 1 79
Baltimore 66 -2 -6 +4 - 4 70
Tampa Bay 55 -1 -4 -5 -10 65
Minnesota 94 +7 -2 +5 +10 84
Chicago 81 -6 +4 -3 - 5 86
Cleveland 74 +3 -1 -5 - 3 77
Kansas City 62 -4 0 -3 - 7 69
Detroit 55 +6 -9 -6 - 9 64
Oakland 103 +6 -2 +5 + 9 94
Anaheim 99 -4 +7 +8 +11 88
Seattle 93 0 +2 +4 + 6 87
Texas 72 -5 0 -2 - 7 79
Atlanta 101 +3 -1 +9 +11 90
Montreal 83 0 0 +2 + 2 81
Philadelphia 80 +1 -6 -2 - 7 87
Florida 79 +5 -2 -1 + 2 77
New York 75 -4 +1 0 - 3 78
St. Louis 97 0 +5 +3 + 8 89
Houston 84 -3 +1 +2 0 84
Cincinnati 78 +4 -1 -3 0 78
Pittsburgh 72 +2 -2 0 0 72
Chicago 67 -8 -3 -4 -15 82
Milwaukee 56 -4 -5 -3 -12 68
Arizona 98 +1 +6 -1 + 6 92
San Francisco 95 -4 0 +3 - 1 96
Los Angeles 92 +3 +2 +4 + 9 83
Colorado 73 +4 +6 -9 + 1 72
San Diego 66 +2 -2 -4 - 4 70
Let's work through a few examples to make sure it's clear what we're
trying to say with this table:
- Oakland won 103 games ... six more games than the pythagorean
method says is normal for a team that scored 800 runs and allowed 654,
mainly by posting an extraordinary 32-14 record on one-run games ...
offensive inefficiency (REA of .266 versus a league average of .270)
cost them two wins ... efficient pitching and defense (REA of .252)
added five wins ... overall efficiency adds up to nine wins ... a team
with Oakland's offensive and defensive stats with average efficiency
would therefore be expected to win only 94 games.
- Anaheim won 99 games ... with the best run differential in
the AL, they should have won more games than anyone, but they fell four
games short of their pythagorean projection ... but they had the most
efficient offense in the majors, picking up seven extra wins (63 runs)
because their offensive REA was .292, twenty-two points above the norm
for the league ... they also had the majors second most efficient pitching/defense
(REA of .243, saving 71 runs), good for another eight wins ... overall,
their efficiency on offense and defense overtook their pythagorean inefficiency
for a net gain of 11 wins ... an 88-win season would have been more
in line with their offensive and defensive stats.
- Philadelphia won 80 games ... one more than normal for a team
that was outscored by 14 runs ... the offense was the third least efficient
in the majors (REA of .239 versus a league average of .256), costing
them 51 runs and 6 wins ... defensive efficiency was also a problem
(REA of .261), robbing them of two more wins ... total impact was a
loss of seven wins for a team that was the league's fifth-best statistically
... with normal efficiency, they should have won 87 games.
- the Cubs won 67 games ... eight fewer than normal for a team
that was outscored by only 53 runs ... offensive inefficiency (REA of
.247 versus a league average of .256) cost them another three games
... defensive efficiency (REA of .270) lowered their win total by another
four wins ... this triple whammy of inefficiency cost them 15 wins ...
with average efficiency in all three areas, they win 82 games and finish
a respectable third in the division.
Before leaving this topic, I want to emphasize that I'm not trying to
diminish what the Angels accomplished this year by pointing out that their
offensive and defensive stats are more consistent with those of an 88-win
team. They did win 99 games in a very tough division by doing all the
little things that count: putting the ball in play so even their outs
were able to move runners over, hitting in the clutch, playing great defense,
getting key outs when they needed them, and so on. They did all that again
in the post season, when time after time they got themselves into a hole
against very good teams and found a way to get the job done when it mattered
most. It was a great run by a team that was awfully fun to watch.
The Angels remind me a lot of the New England Patriots. Both were expected
to do very little before the start of the season. Both got off to slow
starts and reached the playoffs by putting together winning streaks late
in the year. Both were more impressive on the scoreboard than in the statistical
leaderboards. Both were intelligent, fundamentally sound teams that had
to scrap for everything they got and came up with big play after big play
when things looked bleak. And because of all that, both teams were a lot
of fun to watch and served as great examples of why championships are
decided on the field, not on paper.
Summing up
I could have used a more sophisticated statistic like Runs Created to
measure the efficiency of each team's offense and defense, thereby factoring
in things like stolen bases, hit batsmen, and a few other stats that contribute
to success. But I'm partial to simpler measures like TBW that are easy
to figure, easy to interpret, and tell essentially the same story as the
more complicated stats. I especially like the fact that runs divided by
TBW, what I'm calling the run efficiency average, produces a figure that
looks a lot like a batting average, a happy coincidence that makes it
easier to get a feel for what's good, what's normal, and what's bad.
It was also very interesting to discover the strong tendency of teams
that are highly efficient or inefficient in these three areas to move
significantly toward the norm the following season. It's very rare for
teams to excel (or fall short) in this way two years in a row. That's
a good thing for team executives to know as they plan for next season.
I recall being very impressed with the Houston Astros, who refused to
panic after a disappointing 2000 season that saw them fall 8 games short
of their pythagorean projection. Many teams would have fired the manager
and turned over half the roster in a futile attempt to blame someone for
their poor showing. Instead, they chalked it up to one of those years
when things just didn't go right and were rewarded with a tie for the
division title in 2001. (Of course, after the 2001 season, they fired
the manager for failing to win in the postseason, but that's a topic for
another day.)
A number of this year's most inefficient teams have changed managers
in recent weeks, and some of those managers are going to look like geniuses
when their clubs make big gains in the win column next year. I wouldn't
mind being Dusty Baker right now, assuming the front office doesn't destroy
the team with ill-advised personnel moves this winter. The Cubs are the
team most likely to get a large efficiency-related bounce, and with one
of baseball's best-regarded farm systems, they are poised for a strong
run in the NL Central.
|
