As instructed in the previous School of Stats lesson entitled, “Welcome to the School of Stats,” more owners of professional sports franchises are managing their teams like every day businesses rather than as a hobby to keep them entertained in their off hours. This movement is what has provided the momentum for statistical analysts to continue their efforts to quantify performance.
One of those efforts resulted in the Puck Prospectus’ advanced statistic, Goals Versus Threshold (GVT), which blends multiple facets of on-ice performance into one value that can be used to compare and contrast any and all players regardless of position. For a refresher on that particular lesson, here is the link to the article: http://www.insidehockey.com/columns/4739
GVT is one of several statistics that attempts to quantify performance. But there is still something missing from the overall equation. While we can now compare players and their respective performances, how do we attach a value (in terms of dollars) to that performance?
Managers and owners of every day businesses clamor to know what the “Return on Investment” (ROI), is for any ad campaign, new product or project undertaken. As such, Managers and owners in hockey (and all sports) desire to know whether their outlay of cash to a specific player was met with results that merited their investment.
Well, now that we have a number in GVT we feel reasonably comfortable accurately measures performance, it isn’t a big reach to use GVT to help find a solution to how much that number is worth in terms of dollars. One such attempt is Puck Prospectus’ “Goals Versus Salary,” (GVS) which obviously (as its name implies) is tied directly with GVT.
In terms of the mathematics involved, GVS is probably one of the simpler calculations we’ll cover here on School of Stats. Here is an example of how to calculate a player’s GVS rating. For purposes of this example, we will use Marian Gaborik’s salary ($7.5 million) and GVT (+5.6) from last season:
Step 1: Subtract the player’s actual salary (not salary cap hit) by the salary of a replacement-level (i.e. minimum salaried player, for which we will assign a value of $500,000) $7.5 million – $500,000 = $7 million.
Step 2: Multiply the remaining # of millions of $ by 3 (an explanation for why we use 3 will be offered later in the article). 7 x 3 = 21
Step 3: Take the player’s actual GVT rating and subtract it by the product from Step 2. 5.6 – 21 = -15.4. Gaborik’s GVS rating for 2008/2009 was a -15.4.
Before we weigh the merits of this formulation, a brief explanation of why we used the value of 3 in Step 2 must be offered. If we had a roster of 20 players, all of whom were paid the league minimum, our total “threshold” payroll would be $10,000,000. The average actual team payroll in the NHL last season was roughly $50,000,000. Since we are utilizing a stat that values the “performance above the threshold level,” we must deduct the “threshold level” payroll ($10,000,000) from the average team payroll amount. This leaves us with $40 million.
Next, Puck Prospectus tells us that the average team GVT rating was +120 goals above threshold. This means that the average NHL team paid $40,000,000 for 120 goals worth of above threshold production. That works out to 3 goals per $1 million and yields us the factor of 3 which we utilized in step 2.
Hopefully the explanation of the actual mathematics made sense to the students of School of Stats. Now I will offer some additional factors to consider when utilizing this rating.
One thing to consider is that a players GVT rating is based in large part on “counting” statistics or the numbers that are accrued throughout the season (like goals, assists, etc.) and not on “rate” statistics (such as GAA, Save %, shooting %, etc.). This means that if a player misses significant time during the season and thus records lower values in those areas then his GVT rating will be adversely affected.
This is precisely why I used Gaborik for the above example. Gaborik missed a total of 65 games last season due to groin and hip issues. In limited time, Gaborik recorded an astonishing 13 goals and 23 points in only 17 games (numbers which are ironically virtually identical to those posted through 17 games to start this season). If you were to take his goal-per-game and point-per-game rates alone then it is inconceivable that Gaborik would have a GVT rating as low as his +5.6 rating from last season.
However, I think that this particular characteristic of the GVS rating is necessary and makes the stat more useful. No owner or Manager wants to pay a player to sit in the press box whether due to injury or ineffective play. If a player is not on the ice then that diminishes the on-ice ROI that the team is receiving from their investment in that player. That’s the whole point of this statistic.
Now, to further illustrate the potential usefulness and validity of this stat, here are the 10 worst, followed by the 10 best GVS finishers from a year ago:
Player Salary (in millions) GVT GVS
Mats Sundin $8.6 1.6 -22.7
Daniel Briere $8.0 4.3 -18.2
Wade Redden $8.0 4.4 -18.1
Joe Sakic $6.0 0.6 -15.9
Dany Heatley $10.0 13.0 -15.5
Marian Gaborik $7.5 5.6 -15.4
Brad Richards $7.8 6.6 -15.3
Sergei Zubov $5.4 -.04 -15.0
Mike Fisher $6.0 2.1 -14.4
Scott Gomez $8.0 8.2 -14.3
A quick look at the above players and I’m sure most can agree that all of them underperformed their salaries last season. I’m sure we could even agree on that without actually seeing it in black and white with the GVS ratings given. Of course several of the players were victimized by injuries which negatively impacted their GVT ratings and thus their GVS ratings as well.
However there is one aspect of this rating which skews the numbers slightly and somewhat unfairly. The GVS rating uses actual salary instead of a player’s salary cap hit. Most GM’s would probably say that they are more concerned with a player’s cap hit rather than his actual salary. GM’s can convince an owner to go beyond the team’s budgeted payroll but no GM has ever been successful in getting the NHL to raise the salary cap ceiling just for them.
The salary cap ceiling is fixed and it provides a tangible line that no GM can cross for any reason. In fact, with the approval of their ownership, many GM’s are frontloading long term contracts to star players and then providing for smaller compensation at the back end in order to both entice players to sign by offering huge, up-front money while spreading out the annual cap hit over the long life of the deal. This is a blatant attempt to work around the limitations of the salary cap and is likely going to be addressed during the next round of collective bargaining between the NHLPA and the league.
Looking again at the list above and substituting salary cap hit in place of actual salary, the GVS ratings of two players (Dany Heatley and Mike Fisher) improve significantly.
Player Salary GVT GVS
Heatley $7.5 13.0 -8.0
Fisher $4.2 2.1 -9.0
While the adjustment didn’t make either player’s GVS a positive, a jump of +7.5 points and +5.4 points respectively can’t be ignored and provides cause to possibly adjust the criteria with which GVS measures players.
Now let’s peruse the list of the 10 best GVS ratings of 2008/2009 to see if we can pick up any patterns there:
Player Team Salary GVT GVS
Evgeni Malkin Pittsburgh 0.98 23.4 +22.0
Zach Parise New Jersey 2.50 24.2 +18.2
David Krejci Boston 0.83 18.3 +17.3
Nicklas Backstrom Washington 0.85 17.4 +16.4
Phil Kessel Boston 0.85 15.3 +14.3
Johan Franzen Detroit 1.15 15.2 +13.3
Devin Setoguchi San Jose 0.85 13.9 +12.9
Loui Eriksson Dallas 1.50 15.5 +12.5
Anton Babchuk Carolina 1.00 13.7 +12.2
David Booth Florida 0.68 12.3 +11.8
The one thing that sticks out immediately is the lower salaries relative to the players on the list of 10 worst ratings. The highest salary on this list is the $2.5 million paid to the Devils’ Zach Parise. The lowest paid player on the list of 10 worst was Sergei Zubov with a salary of $5.4 million. Obviously players that earn a smaller salary have a better chance of recording a good GVS despite recording moderate GVT ratings.
Something less obvious to the casual glance but infinitely more important to the relevance of this stat is the youthfulness of the players on this list. Scrolling down this list you’ll find that seven of the 10 players listed just completed either their second or third NHL season and thus were still playing on their entry level contracts. The other three just finished with their fourth season and none have been eligible to cash in on the big UFA bucks.
This chart illustrates that GVS ratings reward players whose production trends are on the rise and punishes players whose production trends are falling. Now we all know that eventually, as these players near free agent status, their salaries will skyrocket thus lowering their GVS ratings. In fact, Evgeni Malkin’s extension just kicked in for the 2009/2010 season. If measuring last year’s production against his cap hit for this year ($8.7 million), Malkin would have finished with a -1.2 GVS rating instead of +22; a swing of -23.2 points.
I think that Puck Prospectus and the GVS rating are on the right track even if this stat has been shown in this article to favor youth and punish older players. GM’s have regularly overpaid players for past production and these numbers show that tendency quite clearly.
While moving in the right direction, there is still clearly some more work to do. In fact, the GVS stat has sparked some thoughts within my own mind. I have undertaken to create a stat similar to baseball’s VORP stat myself. I’ll have more on my personal crusade in an upcoming article.
For now, I sign off and I sincerely hope that you find the lessons that the School of Stats offers as intriguing as I do. Please email me at firstname.lastname@example.org if you have any questions, comments about past lessons or suggestions for future research or articles. Thanks for attending.