It must be frustrating to be a Toronto Blue Jays fan. In a division with two big spenders and last year’s breakthrough team, the Blue Jays often find themselves overlooked. They’ve quietly managed to average more than 85 wins the past three seasons without even sniffing a playoff berth.
After four straight seasons averaging more than four wins above replacement (WAR), including a 5.7 WAR season in 2006, Vernon Wells has turned in two consecutive seasons that hover around the 1 WAR level. While that in itself is somewhat surprising, the more interesting fact is how he achieved that one win per season.
In 2007, Wells had a simply awful year at the plate, no matter how you measure it. An 85 OPS+, a .306 Weighted On Base Average (wOBA), -14.5 batting runs below average. Not a great season from a player expected to carry the offensive load. Wells did manage to maintain some defensive value in 2007 though, coming in as a slightly above average center fielder (at least according to some systems).
Last season was a complete reversal for Wells. His offensive performance rebounded considerably, all the way back to a 121 OPS+. That equated to about 10 batting runs above average. It would have been more, but Wells missed a fair amount of time with knee and back injuries. Those injuries may have contributed to an extremely poor year in the field, where he was over 12 runs below average in his limited time.
So, Vernon Wells over the past two seasons has been an enigma, wrapped in a riddle, wrapped in an $18 million contract.
Okay, that’s a little unfair. Because of the way his contract is structured, his base salary is only $1.5 million this season (plus an $8.5 million payment toward a signing bonus). But the contract escalates to $20 million next season if you include the signing bonus, and stays above the $20 million mark until 2014.
With those contract figures in mind, and considering his extremely uneven performance in 2007 and 2008, J.P. Ricciardi and Toronto fans must be wondering what they can expect from their high-priced center fielder.
Using similarity scores
Rather than take the typical approach to projecting Wells’ performance, with its weighted averages and regression (not that there’s anything wrong with that), I’m going to do something that I hope is a little more fun.
Most of you are probably familiar with the concept of similarity scores. Bill James invented them and they’re one of the more popular features of Sean Forman’s baseball-reference.com. Essentially, they show how comparable a player is—not necessarily in value—but in the shape of a career. Power hitters tend to be grouped with other power hitters, speedsters with other speedsters and so on.
There are some major shortcomings with the similarity scores as conceived by James, most notably the lack of park and era adjustments. Because of that, most players tend to find their comparables among others in the same era.
I’ve corrected for those issues by normalizing each player season based on the expected rate for each batting event. If this sounds familiar, it’s because it’s quite similar to what Chris Jaffe wrote about earlier this week.
Jaffe has his own methodology for calculating similarity scores, but I stuck with the formula that James created. It’s not necessarily the best approach, since it’s largely based on counting stats, and stats that are often disparaged by sabermetricians. But it’s familiar and allows for comparisons against the comparables found on baseball-reference.com.
Now that we have determined the players most similar to a given player, we can use those comparables to loosely project how the players will perform in the future.
It’s far from scientific, but sometimes you find out interesting things.
Projecting Vernon Wells
Let’s apply this approach to Vernon Wells.
Keep in mind that the similarity scores are based only on offensive stats, so that tends to be the limit of our prognostication. Defense is obviously extremely important in determining value, but since I don’t have defense similarity scores (yet), we’ll have to ignore it for now.
Wells completed his age 29 season in 2008, so we’ll find his 10 most similar players at the age of 29. Those players are:
Player SimScore Dan Ford 975 Mel Hall 973 Dusty Baker 970 Gus Bell 970 Gary Matthews 966 Rip Repulski 965 Candy Maldonado 965 Paul O'Neill 962 Bruce Campbell 962 Ben Ogilvie 962
The quality of players on the list is worrisome at first glance. Many are good players, but they’re not superstar quality like you’d hope for when you’re paying $20 million a year. Of course, the comparisons are slightly misleading, because Wells is a center fielder and most of his comparable players are corner outfielders.
How did these players perform offensively at age 30 and beyond? To answer that question we turn to Sean Smith’s wonderful database of historical WAR values. Since the database goes back only to 1955, we have to eliminate Bruce Campbell from our comparison, but we’re still left with nine players to compare.
Let’s look at the aggregate performance by age for this list of players. All numbers are runs above average:
Age Players PA Batting Base running GIDP Off RAA RAA/PA 30 9 462 6.78 0.22 -0.11 6.89 0.015 31 9 283 14.89 -0.11 -0.11 14.67 0.052 32 8 393 4.25 -0.50 -0.38 3.38 0.009 33 8 344 8.75 -0.25 0.50 9.00 0.026 34 6 414 11.67 -1.17 -0.50 10.00 0.024 35 6 316 8.67 0.00 -1.17 7.50 0.024 36 4 309 6.50 -0.75 -0.75 5.00 0.016 37 3 309 -4.67 -0.67 -0.33 -5.67 -0.018
The aggregate player has a very nice bump at age 31. The nearly 15 runs above average are especially surprising considering the measure accounts for playing time and our aggregate player missed a lot of time. Most of the players on the list had outstanding seasons at the plate at age 31, led by Paul O’Neill, who was almost 50 runs above average.
Unfortunately age 32 saw a major drop-off, with 32-36 averaging about eight batting runs above average. The base running and double play numbers are slightly below average, but not enough to affect things too much.
Age 37 sees a cratering in value for the three players who are still active, to the point where our aggregate player is below average. Whether it was because of poor performance, or for other reasons, all our comparables hung up their spikes before they turned 38.
A graphical representation of the same data illustrates the nice peak at age 31 and the corresponding declines.
Putting a price on performance
The price for a single WAR in 2008 was somewhere between $4 and $4.5 million on the free agent market. Ignoring any inflation, for Wells to earn the $117 million left on his contract, he’ll need to total at least 26 wins above replacement in the next six seasons.
Let’s be generous and assume that Wells will continue to play center field for the length of his contract, and that he won’t miss substantial time to injury. Those facts alone should contribute to somewhere around 14.5 wins above replacement. That means Wells will need to contribute 11.5 WAR between his offense and defense before his age 35 season.
Our aggregate player earned 51 runs above average in that time, or approximately five wins. If we give Wells credit for the same offensive performance, he’s still 6.5 wins short of the break-even point.
That many wins on defense would be hard to achieve for any fielder in his mid-30s, let alone one whose best season rated just under one win—and who more typically clocks in at average or just below.
Of course, this isn’t really a precise projection. We’re extrapolating from nine data points that exhibit some level of similarity to Wells. It’s entirely possible he rebounds from his last two seasons and contributes much more than his salary. It seems more likely, however, that Wells will fail to perform up to the expectations set by his contract.
If that’s the case, Wells could be a substantial drag on the Blue Jays’ payroll well into the next decade, which makes their chances of successfully hurdling to the top of the AL East even slimmer than they already are.
References & Resources
Current WAR information courtesy of Fangraphs.
Historical WAR information courtesy of Baseball Projection.com.
For an explanation of how similarity scores are calculated, see the details at baseball-reference.com.
Contract information courtesy of Cot’s Baseball Contracts.