Freelance Friday: Salary Value Rating

Freelance Friday is a regular feature at The Nylon Calculus where we solicit contributions from readers. This edition is a discussion of a new metric focused on tying player value to their salary within the context of the salary cap by Neil Johnson. Neil is a programmer from Columbus, Ohio who watches entirely too much basketball and approaches life as one big optimization problem. Growing up in LeBron’s shadow led to a love for the NBA and basketball in general. Follow him @neilmjohnson. Questions, comments or submissions for Freelance Friday should be directed to@NylonCalculus on twitter or via email to TheNylonCalculus at Gmail dot com.

Introducing SVR

In the modern NBA, the Association is now in the news year-round. Why? There is as much interest in what happens in the front office as with what happens on the court. The Decision was a Thing, and contract negotiations are live-tweeted by fans, journalists, and even the people involved[1. Using the emoji language of course]. Roster construction is in the public interest, and transactions are happening more than ever.

Whether people want to admit it, contracts affect the on-court product a team puts out. If a player takes up a large portion of your cap space and doesn’t contribute, that matters more than a guy making the league minimum being unplayable. That is where this tool comes in. To quantify this impact of contract value so as to be able to compare players, teams, and eventually trades, I created Salary Value Rating (SVR.)

The goal of SVR. The Salary Value Rating.

The Salary Value Rating blends three components into a simple formula:

• Relative salary size
• Player output relative to salary
• Available minutes played

Salary size for each player is expressed as a percentage of the salary cap for that year. This number normalizes nominal salary numbers to account for inflation[2. The equation itself is parabolic to hurt the guys who take up almost all of a team’s salary cap.]. Measuring a player’s output relative to salary is the simplest part of the formula: win shares[3. Win shares were chosen because it allows us to look back past seasons without play-by-play data.]  multiplied by salary impact. The last component is a modifier assigned by the proportion of available minutes played to emphasize the importance of a salary for a player who plays a lot over the salary for a player who rarely plays.

The formula

So the formula is relatively simple but requires a different data set than the one NBA statisticians typically use. Here is a look at it with a more detailed explanation in the footnotes[1.

We see the three components and a coefficient. The third component is more complex because it uses a special equation to weight the percentage of minutes played. Here is that formula:

Which doesn’t mean much right now without this:

So what does function g do? Function g helps weight the percentage of minutes played in a more accurate fashion. Here is what the plot of function g looks like:

The goal of function g was to penalize the players who rarely played, reward the players who played too much, and to leave the guys in the middle alone. Why? Because the middle is when minutes are more likely to be awarded by strategy than by merit. Quantifying a player’s fit within a coach’s strategy is near impossible unless teams were to divulge their plans.

So as we can see we the three components of this equation factor in the three core aspects of a player’s value: their salary, their on-court performance, and how much they were contributing throughout the season.].

How can we use this?

The goal for SVR is to provide a foundation upon which the basketball community can build salary-related metrics. At this point, with this stat we can quantitatively determine which players provide the best value for their contract on a seasonal or multi-season basis, This can even apply on the team level. The average salary value rating for all of the player’s on a team would provide proper context into which teams maximized all of their roster spots annually in a vacuum[4. A possible next step is to build a contract value rating, which would evaluate the impact a player had throughout the life of their contract. That will be coming up in the future, and will allow for quantitative evaluation of trades.] Let’s look at how the salary value ratings stack up to conventional wisdom.

Some SVR highlights

The full SVR for all players for which I have data and all teams since 1990 are at the end of this post To give an idea of how the numbers look, mean team SVR score is around 175[5. The distribution of ratings is almost exactly linear].

Individually, here are the top 5 players with the best SVR for the 2014-2015 season.

Coincidentally, 4 of these guys ended up in the top 6 for MVP voting this past season. If the MVP award were decided by the Salary Value Rating, James Harden would have won[3. Ed – for LAST season…]. Having played a staggering 75% of the minutes a player could play for the Houston Rockets last season gave him an advantage above everyone else. It’s also worth noting how much value Chris Paul and DeAndre Jordan gave the Los Angeles Clippers last year. The barebones bench and Blake Griffin’s injury were clearly offset by their contributions.

What to remember

The salary value rating is just that and is best looked at in combination with other stats. The best way to use it is to see the value of a player’s salary instead of just someone’s salary. That way, you can compare across generations Twenty years ago, $25 million was the salary cap. Next year, $25 million will be a run-of-the-mill max contract, and you can see how well a player “earned their contract” or wasted a team’s money.

One thing to note is this number does not account for injuries, and I like that. If a player can’t play, then that is detrimental to the team, and his salary negatively impacts the team. In the future, I will unveil a metric that builds off of this: the contract value rating. From there, we will be able to compare trades, depending upon how one weighs certain aspects of the trade.

Conclusion

So at this point, this article has been read by people smarter than me who are thinking about ways to build off of this metric. The goal is to establish a simple tool that we can use to quantitatively evaluate General Managers, teams, and even agents [7. At least retroactively for now]. General Managers make decisions in a very inefficient market, so it is important to evaluate their performance. I made this metric because I was surprised there was no standard, public tool out there already. I’m sure behind closed doors teams already have something similar, if not better, but this should get the ball rolling for the rest of us.