WAR Rates by 2014 Salary: A First Pass |

Suppose we sorted by salary all the 2014 position players. What salary lines would you guess might split them in two equal groups, by (1) total plate appearances, (2) total WAR, and (3) total salary?

Those answers, and more, after the jump.

Plate appearances: $2.1 million
WAR: $3.6 million
Salary: $10.3 million

I derived these figures with data from Baseball-Reference and a spreadsheet with formulas to find various totals and averages for any given salary range.

Those splits are interesting, but they’re cluttered by a lot of guys who didn’t play much, especially September rookies. For the rest of this post, I’ve trimmed the pool to players with at least 200 PAs, or 60 games, or a $2 million salary. (Anyone paid $2 million was expected to play somewhat regularly.) This cuts the pool down from 645 to 401 players. For what it’s worth, the excluded players totaled 8% of all PAs, 7% of salary, and -4.8 WAR (but 3.5 defensive WAR).

The new 50/50 split points:

PAs: $2.6 million (176 players at or above that salary; 225 below)
WAR: $3.5 million (154; 247)
Salary: $11 million (57; 344)

The top one-seventh by salary earned as much as all the rest. But this isn’t a class-warfare post.

Here are tables for the 50/50 split points listed above. “No.” is the number of players in that salary range. “PA”, “Salary”, “WAR” and “WAA” are group averages per player, while columns with an asterisk are rates per 650 PAs. The pool total is repeated at the end of each table.

50/50 Split by Total PAs:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$2.6 M and up

176

470

$9.09 M

1.77

$5.14 M

0.21

$12.6 M

2.44

0.30

2.60

-0.30

30.6

89%

50%

52%

Less than $2.6 M

225

362

$0.88 M

1.26

$0.70 M

0.07

$1.6 M

2.27

0.12

2.07

0.31

26.5

11%

50%

48%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

More on column headings: “WAA” is Wins Above Average. “Off.” and “Def.” are offensive and defensive WAR. (Here’s B-R’s “WAR Explained” page.) The “%” columns are the group’s share of the pool total in that category. It’s a lot of data for one table, but I couldn’t decide what to leave out.

50/50 Split by Total WAR:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$3.5 M and up

154

493

$9.97 M

1.94

$5.15 M

0.31

$13.1 M

2.55

0.41

2.72

-0.32

30.7

85%

46%

50%

Less than $3.5 M

247

357

$1.06 M

1.20

$0.88 M

0.02

$1.9 M

2.19

0.04

2.01

0.28

26.8

15%

54%

50%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

100%

50/50 Split by Total Salary:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$11 M and up

532

$15.98 M

2.33

$6.85 M

0.58

$19.5 M

2.85

0.71

2.99

-0.39

32.3

51%

18%

22%

Less than $11 M

344

389

$2.58 M

1.34

$1.92 M

0.06

$4.3 M

2.25

0.10

2.19

0.09

27.7

49%

82%

78%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

In each case, the higher-priced players scored more WAR and significantly more WAA, both in raw terms and per 650 PAs. But that came at a steep price.

Those split points were mainly curiosities. Here are some more tables with the same format. I’m focusing on WAR and WAA per 650 PAs, and the cost per WAR:

50/50 Split by WAR per 650 PAs:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$800K and up

251

439

$6.85 M

1.59

$4.30 M

0.14

$10.1 M

2.36

0.21

2.46

-0.22

30.1

96%

67%

Less than $800K

150

359

$0.52 M

1.30

$0.40 M

0.11

$0.9 M

2.35

0.20

2.08

0.45

25.5

33%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

Wow. Per time played, regulars earning less than $800,000 were just as productive as those earning more, at one-tenth the price.

We know the outline of this even without the data. The timing of arbitration and free agency dictates that salaries peak long after performance, on average:

The top 50 in WAR averaged 28 years old, $6.3 million and 5.4 WAR.
The top 50 in salary averaged 32 years old, $16.5 million and 2.4 WAR.

Also … In all these tables, the cheaper guys scored more defensive WAR. Defense peaks younger than offense; younger players are cheaper; and offense gets paid more than defense — just ask Ryan Howard or Prince Fielder.

Another 50/50 split for WAR rate comes at $2 million, and makes equal-sized groups:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$2 M and up

201

454

$8.23 M

1.65

$5.00 M

0.14

$11.8 M

2.35

0.20

2.50

-0.27

30.5

92%

56%

Less than $2 M

200

364

$0.71 M

1.32

$0.54 M

0.12

$1.3 M

2.36

0.22

2.14

0.34

26.2

44%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

But something quite interesting happens if we split the higher salaries into two equal-sized groups:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$7 M and up

518

$12.89 M

2.28

$5.67 M

0.56

$16.2 M

2.86

0.70

3.01

-0.34

31.5

70%

31%

37%

$2 M to < $7 M

103

394

$3.80 M

1.05

$3.63 M

-0.25

$6.3 M

1.72

-0.42

1.86

-0.18

29.2

22%

25%

18%

Less than $2 M

200

364

$0.71 M

1.32

$0.54 M

0.12

$1.3 M

2.36

0.22

2.14

0.34

26.2

44%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

Finally, a three-way split of the $2-million-plus guys:

Salary Range

No.

Salary

WAR

$/WAR

WAA

Salary*

WAR*

WAA*

Off.*

Def.*

Age*

Sal.%

PA%

WAR%

$12 M and up

535

$16.45 M

2.40

$6.84 M

0.63

$20.0 M

2.92

0.77

3.04

-0.43

32.3

48%

17%

21%

$6 M to < $12 M

494

$8.38 M

2.10

$3.99 M

0.47

$11.0 M

2.76

0.62

2.93

-0.19

30.3

27%

17%

20%

$2 M to < $6 M

384

$3.49 M

0.93

$3.74 M

-0.34

$5.9 M

1.58

-0.57

1.72

-0.20

29.2

18%

22%

14%

Less than $2 M

200

364

$0.71 M

1.32

$0.54 M

0.12

$1.3 M

2.36

0.22

2.14

0.34

26.2

44%

Total pool

401

409

$4.48 M

1.48

$3.02 M

0.13

$7.1 M

2.36

0.21

2.34

0.00

28.6

100%

That’s a stark underachievement by the $2- to $6-million group. They used 22% of PAs and 18% of salary, but produced just 14% of WAR.

Could this be a product of selection bias? (See also note (1) below.) Twelve of these 92 players missed the thresholds for PA and games, and were included only because they earned at least $2 million. However, those 12 didn’t significantly affect the group rates. Excluding them would raise WAR/650 from 1.58 to 1.67, still way below average.

This intermediate salary range has a broad mix of ages. Of the 80 that met the playing-time threshold, more than half were age 30 or older. And it’s they who sink the group average:

Age 30+ — 1.0 WAR and -1.2 WAA per 650 PAs
Under 30 — 2.3 WAR and 0.2 WAA per 650 PAs

Eighteen of 42 in their 30s scored zero or less in raw WAR, while just eight of 38 in their 20s were below half a WAR.

A few of these older players got courtesy salaries, for lack of a better term: Paul Konerko got $2.5 million last year because he’s a franchise icon, not because there was any real chance of a return to productivity. Some were on tail-ends of longer deals. Some were signed with cockeyed optimism after a good season that defied their age. Some signed with teams that don’t fret about money. A lot of them weren’t worth the roster spot.

One more thing stands out in these tables:

$12-million-plus had the highest WAR/650 of any substantial group I’ve seen in this study (50+ players). The rate’s even higher at $16-million-plus (25 players), despite flops by the top two in salary. In general, premium-priced players did produce at the highest rate in 2014.

But the marginal cost is extreme. The $12-million club produced an extra 0.7 WAR per 650 PAs, compared to all those making less — at an added cost of $15.5 million per 650 PAs. That’s roughly $22 million per marginal WAR.

__________

Now that I’ve built the spreadsheet, I could do more with it — age studies, splits by method of acquisition, etc. Are you interested? I know I should have used multi-year data, but I’d gone too far before it occurred to me. And I have to do pitchers, of course. Next time.

I welcome your comments and suggestions. You might also point me to sites already doing this sort of thing. I know there’s a lot on price-per-WAR by Dave Cameron and others, but … well … I just don’t browse as much as I might.

__________

Notes:

(1) The playing-time thresholds cause selection bias. Younger, cheaper players who play poorly are more likely to get benched before reaching my cutoff points. Plus, I included everyone who earned $2 million, regardless of playing time. In my defense, I’m trying to look at both performance and value for investment. Prince Fielder, Dan Uggla, Shane Victorino and Carlos Quentin divvied up $50 million while contributing basically nothing; I think ignoring that would be worse than the alternative.

(2) Forty-five players in the pool had no salary listed on B-R, so were assigned the minimum salary, per B-R’s advice. At a glance, I don’t think any of them would have earned more than $800,000, the lowest of the dividing lines used here. This group played at roughly the level of those with published minimum salaries.

(3) I had no list of those who missed the whole year, so I excluded all of them. Anyway, that’s rare for position players.

0 0 votes

Article Rating

11 Comments

Oldest

Newest Most Voted

Inline Feedbacks

View all comments

Monday: Pulling for Punto | The Frosty Mug - All Things Milwaukee Brewers

9 years ago

[…] worth noting, however, that getting a little better is typically not easy or cheap around baseball. John Autin of High Heat Stats broke down players by salary and discovered that the financial cost of adding one win above […]

mosc

This entire analysis is distorted by including pre-arb players in the first place. They have no business being judged on performance per dollar basis. Also multi-year deals should be averaged like the league office does for computing team salary thresholds.

Author

John Autin

Reply to mosc

mosc, I get your point — lumping all player together blurs the big differences in salary according to service time. But to the extent that that I offered any analysis of the data, I don’t think it was distorted. This was a broad overview, and a first pass. Different data sets address different questions. If we want a price per buyable WAR, we’d have to look at buyable players. But there are broader questions that need a complete data set. As to using average salary of long-term deals: I would like to have that data, but I don’t know where… Read more »

Doug

Your split showing equal WAR and WAA production rates above and below the $2M salary mark would indicate that small market clubs can compete (see Oakland and KC). However, when you split it three ways, showing a big WAR production rate deficit in the mid-tier salary range, that indicates there is still an advantage for the big market clubs. Shrewdness and discipline can improve chances of developing top-producing young talent, but there is still the element of chance involved. Thus, those clubs who can get the same production from top tier (and top salary) free agents will have the upper… Read more »

Reply to Doug

Well this is distorted by the assumption that you have replacement level players who can contribute from your farm system. Particularly for starting pitching, that’s not the case. A qualifying starting pitcher at replacement level is of huge value. You can say he has a negative WAA and paying him $10m a year is absurd but it’s not a choice teams have. The replacement options will be much lower and although cheaper will not provide replacement level innings, will stretch the bullpen, and wreak havoc on your 40 man roster as they get injured. Again, the players that sign contracts… Read more »

mosc, I agree with everything you’re saying here — except the part about my assuming the availability of replacement-level players. I guess there’s an implicit assumption in computing a marginal cost per WAR for those making $12 million-plus versus those making less. But I don’t mean to imply that cheap talent is just there for the taking — nor, on the flip side, that every team can sign a Robinson Cano. Certainly, there can be no meaningful conversation about player costs and production without considering the fixed structures of arbitration and free agency. Part of my reason for compiling this… Read more »

Reply to John Autin

Baseball protects seniority. The current system avoids the basketball and football problem of overpaying unproven players based on potential all but ensuring the only overpay is to far more seasoned vets. Yeah, Mike Trout could have gotten hit by a bus after 2013 and gotten paid only a couple million bucks for two historically incredible seasons and that’s not fair. That said, there aren’t many guys who succeed in pre-arb years that don’t eventually get paid. Trout’s already got 9 figures attached to his name, for example. Harper could turn out to be a bumb during his arbitration years and… Read more »

mosc, those are good points about MLB’s salary structure compared to other major sports. But my concern here isn’t fairness — just the complexity of building and sustaining a team in the framework of three vastly disparate salary tiers.

I agree that a baseball GM’s job is very interesting. But I think the 3-tier structure much reduces the number of trades, and leads to some Frankenstein trades that don’t fit on-field needs as much as they should because so many contract factors need to be balanced.

Mike L

Mosc and John A, the arbitration system, interlocked with service time and free agency, is a way to reconcile the competing interests. It’s definitely not perfect, and it doesn’t fairly value on field performance, but that’s not its intent, until players reach free agency. In the past, MLB owners have argued that they need to be compensated in team control and tiered limits because of the expenses of player development/minor leagues. They have a point, when compared to football, which has the gigantic college football complex acting as the minor leagues. The answer to those six years of control is… Read more »

Hartvig

A good example of this would be comparing the Royals signing of Kendrys Morales to the Pirates signing of Corey Hart. Morales has a career slash line of .271/.324/.460, is a DH and will turn 32 next year. The Royals signed him to a 2 year/ $20 million contract (the second year is for $11 million with a $1.5 million buyout) Hart has a career slash line of .271/.330/.480, play the outfield and will turn 33 next year. He signed a $2.5 million dollar contract with another $2.5 million in incentives. They both sucked last year. Yes there are differences.… Read more »

Reply to Hartvig

Bizarrely, part of the reason for overpaying Morales is to convince the home fans that the Royals aren’t going to take Billy Butler’s departure lying down and have ponied up to replace his bat.

But, I doubt many Royal fans will be convinced that Morales will make up for Butler. Silliness, really, and insulting to your fan base, to overpay just to make a point (and make it very unconvincingly).