Fantasy Hockey: Evaluating and Projecting Goaltenders

New York Islanders goaltender Evgeni Nabokov
New York Islanders goaltender Evgeni Nabokov
Apr 18 2013 Toronto Ontario CAN New York Islanders goaltender Evgeni Nabokov 20 during the pre game warm up against the Toronto Maple Leafs at the Air Canada Centre John E Sokolowski USA TODAY Sports

How good a goaltender projects to be is a very precarious proposition. There are so many mitigating factors that play into how a goaltender performs at the NHL level that anyone who proclaims that they know a certain goaltender will be elite before they actually prove it over a large sample (there’s no real baseline, but around 100 starts is a good start) is really kidding themselves. Steve Mason won the Calder trophy in his rookie year as the NHL’s top rookie performer—the only goalie to do so since the 2004-2005 lockout—and then posted an .899 SV% over the next three seasons. Andrew Raycroft was traded from Boston to Toronto in exchange for prospect goalie Tuukka Rask. Raycroft would post an .890 SV% in his two seasons with the Leafs and Rask, well, you know. Montréal’s Carey Price is the last goalie to be taken in the top-5 at the NHL Draft (back in 2005). Price has posted as many seasons (2) with a .905 SV% as with a .920+ SV%.

My proposition—that I’m still testing—is that the goalies who can make the extra saves are the ones who prove elite. There’s a reason why Martin Brodeur, Tim Thomas and Dominik Hasek are perhaps the three top goaltenders of the last 20 years—their style lends to being able to make those crazy stretch saves on rebounds or cross-ice passes.

Here’s what I mean. Let’s say a goalie starts 60 games in a given season and faces 30 shots a game, that’s 1,800 shots against. At a .920 SV% rate overall, that’s 144 goals against. Suppose a goalie has a .915 SV% for the season, that would give us 153 goals against. Just nine goals over the course of 1,800 shots means the difference between a fringe top-10 goaltender and a fringe bottom-third goaltender. In 2011-2012, that one-half of one percent was the difference between Tim Thomas (10th in SV%) and Antti Niemi (21st), the year before it was the difference between James Reimer (T-10th) and Brian Boucher (T-17th w/ four others). I contend that the ability that Brodeur, Hasek and Thomas possessed was to make those extra saves, particularly rebound saves, which go in at a very high rate.

However, without a large sample like the careers of Brodeur or Hasek, this would be difficult to prove. So we are back to our original issue, how do we really evaluate goaltenders? Is physical size a factor? Sure, but remember that for every Pekka Rinne at 6 feet 5 inches there’s a Steve Mason who stands at 6 feet 4 inches. Is prior NHL performance important? I don’t think you can argue that, but at the time of the trade, Raycroft was one season removed from a .926 SV% campaign in 57 games (this is a perfect example of how one season is not a good sample size). Is pedigree an indicator? It can be, but there’s a long line of Finnish goaltenders who would argue otherwise: Minnesota’s Niklas Backstrom was undrafted, Rinne was an eigth round pick in 2004 and Miikka Kiprusoff was a fifth round pick in 1995. Henrik Lundqvist (a Swede) is regarded as possibly the best goaltender on the planet right now and was a seventh round pick in 2000. This is perhaps the biggest reason why Carey Price is the last goalie chosen in the top-5 of a draft; it’s so hard to project goaltenders that using an early pick in the draft on a goalie can be a very risky proposition (cc: Rick DiPietro).

So how can you possibly evaluate goaltenders? Save percentage is a good place to start, but it too has its issues. As can be expected, a goaltender’s SV% can be very reliant on the goalie’s SV% while on the penalty kill (Power-play SV%), despite a far fewer ratio of shots. Power-play shots have higher expected scoring rates and as I demonstrated earlier, even nine goals over the course of a large workload can greatly alter a goaltender’s value. This is why you will see very few goaltenders at the top of SV% stats list who play on teams with poor penalty killing rates; of the top six goaltenders in save percentage in this past season none of their teams finished lower than 15th in penalty killing (Lundqvist, tied at fifth with a .926%; Rangers, 15th on the penalty kill at 81.1%). Of course, it’s not a hard-and-fast rule; Philadelphia had the fifth-best penalty kill last year and they consequently bought out their starting goaltender. Usually though, it is a good starting point.

The best objective measure we are left with is even-strength save percentage. In fact, it’s already been shown that other measures we use to evaluate goaltenders (wins, goals against average, shut-outs) are more attributable to the entire team (personnel/philosophy) and random chance than the goaltender themselves (Hat-Tip to Cam Charron. Read more here).

So in this vein, I wanted to look at the even-strength goaltending performance from the starters (Minimum 2,500 shots) of the bottom-10 teams in terms of CorsiFor% – that is the percentage of all shots attempts that belong to that player’s particular team. A 45% CF% means your team had 45 shot attempts while the other team had 55 – over the last three years (2010-2013). Also, I wanted to look at how a team’s penalty kill can affect a goalie’s save percentage on the penalty kill, but won’t look at saves on the power-play (opportunities are far too few). This is how it shook out.

Even-Strength SV%

Power-play SV% (Avg. Team PK%)

Team CorsiFor%

Goalie SV% from 2010-2013

Niklas Backstrom (MIN)


.8612% (82.14%)



Devan Dubnyk (EDM)


.8836% (80.43%)



Jonas Hiller (ANA)


.8707% (81.56%)



James Reimer (TOR)


.8548% (79.82%)



Evgeni Nabokov* (NYI)


.8844% (80.42%)



Pekka Rinne (NSH)


.8834% (82.5%)



Ryan Miller (BUF)


.8795% (81.56%)



Steve Mason (CBJ)


.8481% (79.37%)



Kari Lehtonen (DAL)


.8795% (81.4%)



Miikka Kiprusoff


.8653% (82.38%)



*note: Nabokov didn’t play the 2010-2011 season, for reasons you can read here, so he only had 2,276 shots to draw from over the last two seasons.

A quick note about each goalie on this list (in no particular order):

  1. Steve Mason is not good (relatively speaking).
  2. Devan Dubnyk catches a lot of flak, but considering how poor his team has been for several years now, he’s out-performed names like Miller and Hiller.
  3. Pekka Rinne is really good. If he played on a good team like Los Angeles, Chicago or Boston, he would be at the top of the league’s goaltender categories annually.
  4. The league-average for ESSV% is ~.9210. This makes what Niklas Backstrom did even more impressive. He was on the worst possession team and maintained an EVSV% over three-tenths of a percent higher than even an average goaltender.
  5. Although all the goalies seem to be sport similar numbers, even fractions of a percent can greatly alter the value of a goaltender. Over the course of 3000 even-strength shots, or about two and a half season’s worth, the difference in expected goals at even-strength between James Reimer at 0.9253 and Evgeni Nabokov at 0.9166 (less than 1%) is 24 goals. Less than nine-tenths of a percent is a translation of 24 goals over 2 ½ seasons or about 10 goals a year. An extra goal every six starts is fairly significant. For this reason, save percentages are an indicator as to who might post a reasonable goals against average (although that’s still educated guess-work at best).
  6. James Reimer is a lot better than most people give him credit for. He’s posted better numbers than Ryan Miller over the last three seasons on a worse team and has actually out-performed Niklas Backstrom even though Reimer’s team has had the second-worst PK% on this list.
  7. Ryan Miller needs to find a better team.
  8. Kari Lehtonen is in the conversation for most underrated goalie in the NHL. His career .914 SV% might indicate otherwise, until you realize he’s only played for the Atlanta Thrashers and Dallas Stars.
  9. Jonas Hiller seems to be a good goalie but for whatever reason I still don’t fully trust him (I have no rationalization for this). It may be that I overrate the Ducks year in and year out, but he seems like he’s in the same company as Lehtonen and I can’t bring my brain around to understanding this.
  10. Miikka Kiprusoff was nearly traded to the Leafs last season. Let that sink in.

So if I were to offer some advice when evaluating your goalies next season, it would be to look for a team that’s been around league average on the penalty kill for the last couple of years and combine this with historical even-strength save percentages. While team puck possession certainly is a factor in a goalie’s statistics, good goalies will still usually be good (Rinne, Lehtonen, Backstrom, Reimer) regardless of their team while bad goalies (Mason, Nabokov) will usually still be bad.

author avatar
Michael Clifford
Michael Clifford was born and raised in Fredericton, New Brunswick, Canada and is a graduate of the Unviersity of New Brunswick. He writes about fantasy hockey and baseball for XNSports and He can be reached on Twitter @SlimCliffy for any fantasy hockey questions. !function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0],p=/^http:/.test(d.location)?'http':'https';if(!d.getElementById(id)){js=d.createElement(s);;js.src=p+'://';fjs.parentNode.insertBefore(js,fjs);}}(document, 'script', 'twitter-wjs');