I initially created Baseball Graphs to develop new and better ways of presenting baseball statistics. The 2003 season was my first full-season attempt, and it turned out pretty well. I now spend a lot of my time creating and formatting the baseball stats at The Hardball Times.
Along the way, I've also picked up the writing bug. I spend most of my energy writing for THT, but sometimes I like to write about more geeky stuff, baseball research and things like that. That's what this blog is for.
I'm very interested in graphics and other information display, so I'll probably post links to creative stuff like that. I also plan to write historical articles that complement the historical graphs. And I have inexplicably developed an interest in baseball win-related stats, such as Bill James's Win Shares and something called Win Probability Added. So you can expect a few posts about that stuff too.
If you're interested, I've created a complex spreadsheet full of WPA logic, which you can download at ftp.baseballgraphs.com/wpa. If you'd like to read more about Win Shares, try these links:
Win Shares Intro
A number of folks have e-mailed me asking for some value analysis, as in “which player delivered the most Win Shares for the money? John Konstantino analyzed the data in Pete’s spreadsheet (available on the home page) and sent me a table of best and worst value in the majors last year.
Before we get to that, however, let’s look at the following graph of salary vs. Win Shares:
Despite its looks, there is a positive correlation betwen salary and Win Shares (correlation coefficient of .24). But as you can clearly see on the graph, there was a pretty wide spread. Graphically, it’s clear that the best value in the majors last year was Albert Pujols, and the worst value was Mo Vaughn.
Value is typically defined as Quality divided by Cost (or Performance divided by Cost, depending on your industry). Average major league value last year was .3 Win Shares for every hundred thousand dollars.
Per John’s table, as you can imagine, all the value leaders belonged to the $300,000 club:
Name WS Salary WS/$100K
M Giles 28 $316 8.8
S Podsednik 22 $300 7.3
J Guillen 20 $300 6.7
A Huff 21 $325 6.5
A Pierzynski 22 $365 6.0
O Hudson 18 $303 5.9
M Ellis 18 $308 5.9
M Bradley 18 $314 5.7
B Wilkerson 18 $315 5.7
B Webb 17 $300 5.7
H Blalock 17 $303 5.6
And here’s a list of all players who earned at least $3 million this year and contributed no Win Shares.
C Park $12.9
D Neagle $9.0
A Ashby $8.5
D Palmer $8.5
P Astacio $7.0
J Mesa $5.2
G Rusch $4.3
R Gutierrez $3.9
R Dempster $3.3
J Powell $3.3
O Daal $3.0
S Estes $3.0
K Jarvis $3.0
A lot of pitchers on that list. One last note: according to John’s analysis, the worst free agent signing of the offseason was Tom Glavine.
I’ve added a comment below, with additional information and a link to a full page of data. I also want to display this graph of the data:
I think this graph does a nice job of showing Nate Silver’s point (which I hope I’m representing correctly). Notice how the relative WS Value of a batter declines with his absolute salary, no matter how good a batter he is (also, note that Bonds actually was a very good value, relative to the trend line). The value of a marginal run, to use Tango’s term, rises at an increasing rate.
One final point: you can see how Pujols stands out from the band of all other players. He really did provide extraordinary value in 2003. Ah! The power of the visual display!
Since the “value” list is so heavily tilted toward the minimum wagers, maybe a better metric would be to calculate, in $, each player’s contribution and then subtract their salary from that. For example, Pujols contributed $13.67 mil (41/.3 * 100000). Subtract his $900K salary and his value becomes $12.76 mil, tops in the league. The rest of the top five are Giles, Soriano, Wells and Gagne.
Greg would that mean that Barry Bonds’ value is $-2 million because he has 39 Win Shares and a salary of $15 million. Are you saying then that having Bonds instead of a few cheaper players actually hurts the giants. I find that kind of hard to believe.
Microsoft is a billion$ company, but that doesn’t mean it’s not overpriced at 80$ / share, right? You may be better off buying its competitors, if they’ll give you a better return on your money.
If we expect 48 win shares if we spend 16 million$, then we didn’t get a good enough return on 42 win shares, right?
(Personally, I would do marginal win shares / marginal $, though I’m not sure it’ll have much of an effect.)
Great suggestions and comments everyone. Thanks.
I went back to the data to follow Greg’s suggestion, and found that I had to clean it up a bit more. I may be missing a couple of ballplayers and salaries, but not many. The final numbers: 7,269 Win Shares and $2.2 billion in salaries. $2.2 billion! I had never added up all the salaries before. That’s 0.335 Win Shares for every $100,000—or 1.1 wins for each million in salary. Or $300,000 per Win Share.
Nate Silver had a great article about salary and value on the Prospectus site earlier this year. Unfortunately, BPro’s search engine seems to be one of the worst in the business, and I can’t find it.
If I remember correctly, one of his points is that there is an exponential relationship between value and salary, not a linear one. As a player’s value goes up, his salary does, and should, go up at an increasing rate. This is partly due to marquee value and marketing value. But it also has to do with the capacity issue: 40 Win Shares out of one position is worth more than 20 out of two positions, because you can backfill the second position with replacement value. So I think that Joe makes a good point.
Anyway, I followed Greg’s advice and computed net Win Share value. I’ve posted the entire results at this page:
http://www.baseballgraphs.com/winshares/wsvalue.html
As already mentioned, Pujols and Vaughn were the best and worst values. At the same time, you can see Nate’s point: players such as Delgado and ARod were all at the bottom of the list.
I think this is the BP article you’re looking for: http://www.baseballprospectus.com/news/20030103silver.shtml
Thanks, Greg. That is it.
Upon review of the article, Nate offered several rationales for the increase in salary paid per marginal value unit, but not the one I mentioned. That must have been something I thought at the time.
<I>The value of a marginal run, to use Tango’s term, rises at an increasing rate.
</I>
To be clear, that should read: The COST of a….
The value remains fairly constant, be it at 1 million$/run, or 2 million$/run, or whatever your model says. The extra value to put your team into the playoffs is pretty low, which is why a linear model is sufficient for our purposes.
| <<Previous Article: Baseball Primer thread | Next Article: Is WSAA Valid?>> |