Even after his Detroit Pistons had won 50 games and the Central Division in 2001-02, Joe Dumars knew his team was far from a finished product. With the Pistons over the NBA's salary cap, Dumars entered free agency armed with just the mid-level salary cap exception to look for an upgrade on limited incumbent point guards Chucky Atkins and Damon Jones.
The Pistons offered a six-year deal to Chauncey Billups, who had shown promise as a fill-in starter in Minnesota but figured to be headed back to the bench if Terrell Brandon returned from the knee injury that had given Billups a chance to move into the lineup. Billups, 26 when he began his first season in Detroit, quickly blossomed into one of the league's top point guards and was at the helm as the Pistons won the 2004 NBA Championship. He has since led Detroit to six straight appearances in the Eastern Conference Finals and made three straight All-Star appearances along the way.
Billups is not only the most dramatic success story amongst players signed for the mid-level exception, his deal is one of the few that teams did not ultimately come to regret.
The mid-level is a product of the post-lockout Collective Bargaining Agreement. In exchange for the introduction of the luxury tax on team payrolls and limits on individual player salaries, the NBA Players' Association won increased minimum salaries and the mid-level exception, designed to give the league's middle class more options in free agency.
In theory, the mid-level exception also helped teams over the salary cap by giving them an opportunity to add quality talent in free agency. In practice, it has been a disaster. Take, for example, the class of nine players signed during the summer of 2001, the first year the mid-level exception was set at the previous season's average salary. Nine teams signed players to full mid-level deals (or their equivalent) that summer. None of the nine finished their deals with their new team; four were out of the league before the end of the contract, and a fifth was waived when the league offered luxury-tax amnesty during the summer of 2005. That success rate is not atypical.
Through last summer, 49 mid-level-type deals had been signed (including three players--Jerome James, Nazr Mohammed and Joe Smith--who have twice been signed using the exception). As a whole, these players were predictably average the year before they hit free agency, with average ratings of a .505 winning percentage and 3.6 Wins Above Replacement Player.
Actually, because of their minutes played, the group was really more valuable than average before becoming free agents. If salaries and performance were perfectly distributed, a player making the mid-level salary could be expected to add about 2.5 WARP per season. How has the mid-level group done compared to that standard? Not well at all.
Combined, the mid-level free agents have played 160 seasons on their contracts. Of those, 52--less than a third--have been rated as worth at least 2.5 WARP. Performance over the life of the deal is even more striking. Of the 49 players signed using the mid-level exception, just 13 have averaged more than 2.5 WARP per season during the contract. Nearly as many (10) have rated as below replacement level over the course of the deal.
There might be a tendency to believe this is some kind of statistical mumbo-jumbo, but a more simplistic look at the issue offers a similar perspective. Of the 23 players who have completed mid-level contracts, just five have played out the entire deal with the same team. Eight players have either retired or been waived (most with buyouts) and five more have been traded in what I'd consider salary dumps.
If we can agree that mid-level deals have worked out very poorly for teams, the next question is why this is the case. Is there something specific to these contracts that explains their low success rate?
To investigate, I went back to a method I used two summers ago to look at the inefficiency of the free-agent market in general. As I did then, I divided players into several categories based on how their contracts were signed--rookie scale contracts for first-round picks, as second-round picks, extensions of rookie contracts, veteran extensions, as restricted free agents and as unrestricted free agents. I took the two groups of free agents and separated out the players who signed mid-level deals. Here's how they compare in terms of cost per WARP.
Type No. Salary WARP $/WARP
UFA MLE 19 100,641,965 25.0 4,029,223
UFA non-MLE 158 621,010,086 164.3 3,779,939
RFA MLE 7 39,988,759 9.6 4,158,714
RFA non-MLE 54 289,252,617 84.5 3,424,387
(Technical note: That figure is actually salary above replacement, subtracting out the NBA minimum for each player.)
While the sample sizes are small, particularly for the restricted mid-level players, in both cases teams had to pay more for each win for mid-level players than for free agents in general. However, the issue here is we're not exactly comparing apples to apples. The other contracts range all the way from the max deals for players like Kobe Bryant and Tim Duncan to undrafted rookies at low end. It's better to look at the players grouped by their salary range. This time, because the sample sizes are lower, I've lumped together both restricted and unrestricted free agents.
Type No. Salary WARP $/WARP
max-type 28 373,344,194 179.1 2,084,619
above-MLE 36 263,589,370 63.7 4,136,792
MLE 26 143,899,407 20.9 6,893,776
below-MLE 74 223,042,000 14.6 15,239,172
BAE 15 19,556,055 9.2 2,116,107
minimum 53 30,016,379 -4.2 -7,123,900
Our Baseball Prospectus colleagues are fond of quoting the Bill Veeck axiom that, "It isn't really the stars that are expensive. It's the high cost of mediocrity." These numbers tend to bear that out. Wins become more expensive as salaries go down until you reach the surprisingly-effective group of players making between $1.2 and $2.0 million, approximately the value of the bi-annual salary cap exception.
In part, the group of players making less than the mid-level exception but more than the bi-annual exception is hurt by the fact that it contains two of the league's most notably misrated players, defensive specialists Raja Bell and Bruce Bowen. Bell and Bowen combined for -7.5 WARP last season because their defensive efforts are not accurately valued. Even if you account for that, however, it's clear that these players are even less productive for their teams than those signed for the full mid-level exception. So it's hardly a given that teams that split the mid-level exception or use less than its full value are getting a better deal. (The champion Celtics did seem to work the system last year, using the mid-level to add key reserves James Posey and Eddie House.)
I'd long suspected that one reason the mid-level exception has been such a loser for teams is because there is no ability to compete in terms of salary. That is, if multiple teams are looking to sign the same player and all are over the cap, they are all limited to the value of the mid-level, meaning the length of the contract is the only potential difference between the offers. I figured that meant longer deals for players signed using the mid-level exception.
Looking over the deals, I'm not sure whether that is the case or not. The average mid-level deal has historically been for a lengthy 4.5 years. Already, the 2005 CBA limited free agents changing teams to five-year deals, with those re-signing with their old team allowed to go up to six years. On top of that, teams seemed to find mid-level sanity last year; none of the six players signed for the mid-level got more than four years, and they averaged out to three years--virtually identical to the average for the group of players who signed for between $3 million and $5 million per year.
In sum, it doesn't appear there's anything unusual about the mid-level exception that is responsible for its poor track record. Free agency is simply a risky, expensive business, especially when it comes to signing players in the NBA's middle class. Managed correctly, the mid-level exception can be a huge weapon. The shrewd Dumars managed to use it twice to add two starters for the Pistons in Billups and Antonio McDyess, but even he later struck out on Nazr Mohammed. The Isiah Thomas-era Knicks represent the other extreme, having been saddled with the millstone contracts of Jerome James and Jared Jeffries. As always in free agency, buyer beware.
Five Best Mid-Level Contracts
Chauncey Billups, Detroit - Far and away the best player ever signed for the mid-level
Larry Hughes, Washington - Michael Jordan's best move
Gary Payton, L.A. Lakers - Took discounted deal to join star-studded roster
Hedo Turkoglu, Orlando - Emerged as Magic's best perimeter player last season
Antonio McDyess, Detroit - A bargain after hitting the market one year back from microfracture knee surgery
Five Worst Mid-Level Contracts
Jerome James, New York - Three-year totals: 684 minutes, 217 points, 177 fouls
Anthony Mason, Milwaukee - Wreaked havoc on chemistry as Bucks went from Eastern Conference Finals to the lottery
Michael Olowokandi, Minnesota - Too often overlooked amongst terrible No. 1 picks in favor of Kwame Brown; swapped in exchange for another MLE bust, Mark Blount
Troy Hudson, Minnesota - A hip injury has likely ended Hudson's career with two years left on his deal
Brian Cardinal, Memphis - Six-year deal was too long after breakout season
This Year's Mid-Level Signings
DeSagana Diop, Dallas - Useful role player, dramatically overpaid
Chris Duhon, New York - The money is high, but two-year deal limits risk
James Posey, New Orleans - Will help now, though he may not age well
Beno Udrih, Sacramento - Parlayed solid season as starter into long-term deal
Kevin Pelton is an author of Basketball Prospectus.
You can contact Kevin by clicking here or click here to see Kevin's other articles.