Friday afternoon was a time of celebration for many Columbus fans. The Blue Jackets and Sergei Bobrovsky avoided any summertime headaches by reaching a contract agreement. To no real surprise, Bob’s deal puts him in the upper echelon of NHL goalie salaries. As reported by Aaron Portzline, that’s a cap hit of $7.425M per year over the next 4 years.
It is at this point where I hit a great philosophical problem. A few of the questions that spring to my mind: just how much should a goalie be worth? Is having an elite goalie at an elite salary all that important? Is keeping Bobrovsky worth the kind of cap hit percentage he’ll eat up each year?
In the most basic way: Is Bobrovsky’s new contract good for the Blue Jackets?
Considering Bobrovsky’s Greatness
As preface, I’d like to point you toward the most important goaltending article out there. It’s Eric Tulsky’s work on predicting future performance and finding true talent level. In essence, we can look at even strength save percentage and then refine our out look as the goalie adds to his resume (more time played or more shots against). It’s a Bayesian approach to the situation and I recommend reading the whole thing. At least glancing along will be useful while I walk through this exercise.
In the context of Tulsky’s article? Bobrovsky is among the league’s best and has been that way long enough to indicate his true excellence. First, let’s grab Bobrovsky’s career even strength numbers from NHL.com. This data is current as of 5:00 pm 1/9.
Bob Contract Table 0
Translating from Tulsky’s work (take a look at the table on estimating talent distribution), the data for Bobrovsky puts him nearest the 5000 SA, .925 SV% coordinates. We can estimate Bob has near a .924 SV% talent at even strength, with confidence of about .006 in either direction. Taking a step further back, we can see Bob in comparison to his peers. Using Hockey Analysis, we can look at 5-on-5 SV% for goalies between 2009 and 2014. I’m limiting this to 5,000+ minutes played. Here, Bobrovsky comes out as the 3rd best goalie of the era and has incredible peers (only Rask and Lundqvist above, the underrated Vokoun and Schneider below).
Bobrovsky’s Vezina season? Well-deserved. The years following? Also mostly brilliant. This season’s red-hot December for the Blue Jackets? The handiwork of Bob.
Make no mistake: Bobrovsky is among the NHL’s best goalie. The problem: does that statement justify a $7M cap hit? Is such budgeting even sensible?
The prolific Mr. Tulsky also explored some aspects of goalie valuation in his time with SB Nation. In particular, he used Henrik Lundqvist as a case study, stepping through his thought process on the issue. It’s another rewarding piece and I’d suggest following his article and this part concurrently. We’re going to work through his concepts and consider Bobrovsky’s value.
Following his general method, we’ve already done Step 1 (projecting performance, or getting an estimate of true even strength SV% talent). Step 2 is comparison to replacement level. We’ll use goalies with under 10 games played in 2011-13, 2012-13, and 2013-14 and use data from the excellent War on Ice to get an estimate for replacement level goaltending. On average, this comes out as an even strength SV% of .911.
Comparing to the .924 we estimated for Bobrovsky? That’s 1.4% goals more than a replacement in a regular year. Using the 1,750 shot example from Tulsky, that’s around a 23 goal difference over a replacement goalie.
Estimating 6 goals as 1 win, that means Bob is worth about 3.8 wins over a replacement level goalie. Gabe Desjardins estimated a value of $2.65M per win in a year with a $59.4M cap (considering value from UFAs signed in 2010-11). Scaling that to the 2014-15 cap ($69M) and the projected 2015-16 cap ($73M), we get a value per win of between $3.08M and $3.26M. This mean’s Bobrovsky’s value is estimated to be between $11.7M and $12.4M per year.
Considering the contract structure as reported by Portzline, the $8.5M over year 1 and 2 of the contract might be worthwhile. If Bobrovsky is to continue his near-.924 level, then the estimated value of the player is above the actual salary.
As a final element of this estimate, we should also consider aging and eventual decline for goalies. Tulsky’s estimate of the goalie dip in the Lundqvist valuation article displays the inevitability of time. Garik’s article at Hockey Graphs also gives an estimate of drop in save percentage performance. Garik’s conclusion is more heartening for Blue Jackets fans and the Bobrovsky case (he notes that the strongest decline happens after age 30, Bob will only be 30 at the end of this extension). I won’t attempt an estimate of value over time here, but I’ll note that we should expect depreciation going forward.
Widening the Scope
This all sounds quite promising for Columbus, a possibly wise use of cap allocation. However, I would submit that we should consider a wider angle. Consider the league average even strength save percentage for each of the past two years: .921 in 2012-13, and .922 in 2013-14 (calculated using team data from War on Ice). Bobrovsky was well above the average in those years (.941 and .931, respectively). Yet we already considered the distribution of talent and the impact of age. There’s not much certainty he’ll repeat those kind of herculean performances in the future. There’s even less certainty that he can do that in any given game.
That kind of goalie inconsistency is precisely what makes this so tricky. Even Sergei Bobrovsky bounces all over the place. Consider a plot of his 2013-14 rolling 5-game average save percentage at even strength (again with data from War on Ice). I’ve included that year’s league average even strength SV% as the red line (.922), and that’s the center of the y-axis.
Bob Contract Plot 1
The quick takeaway: Bobrovsky was above average much of the time (especially in his insane 25-30 span). The more concerning takeaway: even one of the greatest goalies of this era is terribly inconsistent. Even on a 5-game basis, a great goalie spends time below average. There is unavoidable uncertainty in net no matter the selection.
To that end: even if we “minimize” risk by selecting Bob, is the risk substantially higher with a “lesser” goalie at a lower price?
Let’s Play Blind Goalie!
To illustrate an example, it’s time to play everybody’s favorite new game: Blind Goalie! (Feel free to play some game show music if you’d like)
I’ve selected 4 Goalies, and I’m naming them Goalie A, Goalie B, Goalie C, and Goalie D. For Round 1, I’m going to list their career even strength data below (from NHL.com). Bobrovsky is included for reference. I will admit at the outset: I’m stacking the deck somewhat in my favor (this is my game, after all), but that it was fairly easy to pick 4 goalies is part of the point.
Bob Contract Table 1
It’s early in the game, but what do you view as the most glaring difference? At this point, it’s number of shots faced. Bobrovsky has a larger resume than all four, suggesting we can be more confident in his level of success.
Let’s move on to Round 2 of the game. Let’s pick up Eric Tulsky’s table again and use the estimation process for valuation. We’ll use the same values from the Bobrovsky estimation steps, but we’ll stop short of putting money values here. You’ll get estimated talent level and confidence, goals over replacement level (over .911 using 1750 shots against per year), and estimated wins over replacement.
Bob Contract Table 2
The goal difference doesn’t loom so large considered this way. The difference between Bobrovsky and our Four Candidates? 0, 1.75, 5.25, and 5.25 goals over the course of a season. That’s something between zero and one win worth of difference.
Now we are less confident in these values (the Goalies do have smaller resumes than Bob), so the risk does increase. These guys may not actually be quite so close.
For our Final Round, let’s see what kind of cap hits these Goalies have this year and next, and also see what age they are right now. Contract data is from NHL Numbers (and included is a moment of silence for CapGeek).
Bob Contract Table 3
This is where the game is tough for Bobrovsky. The Blue Jackets are paying that much more? Is the increased certainty really that valuable? Even if we’re worried about aging and want to throw out Goalie A and Goalie B, I have a hard time seeing many problems with Goalie C.
The Blind Goalies? A – Anton Khudobin, B – Thomas Greiss, C – James Reimer, D – Jhonas Enroth.
Risk and the Goalie Market
Sergei Bobrovksy is a really good goalie. He adds real value to his team, and we can feel reasonably certain that he’s near a talent level that is decently above replacement.
Then again, James Reimer is also a really good goalie. There’s somewhat less certainty about his talent level, but there is a $5.125M gap in cap hit between the current Maple Leaf and Bobrovsky next season. We have to ask: is Bob worth $5M more? Could that money be better spent to improve the rest of the team?
There are some gaping, serious problems for the Blue Jackets this season. The team right now is not playing in a way that sustains winning. Is Bobrovsky a big enough difference from more-affordable options?
Surely for Columbus, losing a franchise face, a shining beacon of hope would not be ideal (especially after the disaster years of Steve Mason). But is the mitigated risk in net worth compromising future salary power at all other positions? It’s a question of asset management and relative goalie value that teams must consider carefully. Simply because the goalie market suggests a high cap hit for a Vezina winner does not mean the payout is the right option.
Sergei Bobrovsky is great. I also can’t help but worry about the rest of the team suffering from a contract disproportionate to other (potentially available) resources in net.