Facebook TwitterText
Share This: |

What Advanced Stats Say About the Sabres Pt. 7: Ryan Miller

Buffalo Sabres' fans have several months to think about what went wrong. Was it the injuries? Was it a combative relationship between players and coach? Was it lack of determination? Or the ever-talked-about “core?”

Often times, those aspects of failure are tough to put a finger on. The only ones who know the answers to those questions are behind closed doors. But that doesn't mean we can't search for answers within the resources we have. One of those untapped resources is advanced statistics. Yes, they exist in hockey, not just baseball.

In part 7 of our series, we focus on goaltie Ryan Miller.

The question: Why hasn't Miller been the Vezina-caliber goalie the last two years that he was in 2009-10? Is just still an elite goaltender? Is he overpaid? Will he be better next year?

The majority opinion: Miller can still play like an elite goaltender at times, but he is too inconsistent and will never be one of the league's best again. His cap hit isn't worth his average play.

What the stats say:

The problem with mainstream goaltending statistics a la save percentage, goals against average, wins, is that they don't provide any context. When you look at Ryan Miller's career .915 save percentage, you might consider that to be pretty mediocre – same goes for his .916 last season. But there are so many factors that affect save percentage and how we view the number as being “good” or “bad.” Consider factors such as the quality of defenseman and defensive forwards in front of him, the quality of his team's penalty kill, how often his team is on the power play. They all can bump up or down this catch-all statistic.

We'll get to all those factors as they pertain to Miller in a minute, but first, let's address that career save percentage. You see, the league has undergone massive changes since Miller first stepped in between the pipes as a full-time starter. Sure, .915 might not be that impressive this year, but it wasn't always that way.

Here, I've done a breakdown of Miller vs. the NHL average save percentage since he became a full-time starter in 2005-06. I've taken his save percentage, compared it to the league, then determined the number if goals allowed above what an average goalie would have allowed if that average netminder had faced the same amount of shots.* Take a look:

*This is figured by taking the amount of shots faced, multiplied by the goals-per-thousand shots above average.


Miller - .914

League - .901

Goals above average: 18.72


Miller - .911

League - .905

Goals above average: 11.32


Miller - .906

League - .909

Goals above average: -6.31


Miller - .918

League - .908

Goals above average: 17.73


Miller - .929

League - .911

Goals above average: 37.76


Miller - .916

League - .913

Goals above average: 5.89


Miller - .916

League - .914

Goals above average: 3.58

Goals per 1,000 shots above average – 7

Shots faced per season – 1,865

Total goals above average per season – 13.06

What we find is that Miller has been well, well above average throughout his career. In fact, his average goals above average per season since becoming a starter is 13.06. That puts him in the “elite” category since the lockout. Only Henrik Lundqvist, Tim Thomas, Tomas Vokoun, Carey Price and Roberto Luongo are in the same ballpark as Miller in terms of goals above an average netminder since the lockout.

What these goaltenders have in common is a big enough sample to determine that they have had an array of different teammates, some have had different coaches, systems etc., and yet year after year they are above average. Look no farther than Mike Smith or Brian Elliott as goaltenders who have only had success under one type of system and with one set of teammates. Lundqvist, Thomas, Vokoun, Price, Luongo and Miller have proven they are not simply a product of outside factors.

Now – take a look at last season. It was Miller's second lowest total of goals above average. Let's look at the outside factors as they pertain to Miller's save percentage and compare them to the Vezina Trophy Finalists:

Ryan Miller

Puck possession rank - 23rd

Shots against rank - 26th

PP time - 23rd

PK shots vs. - 22nd

Blocked shot rank - 14th

Jonathan Quick

Puck possession - 4th

Shots against - 5th

PP time - 5th

PK shots vs. - 28th

Blocked Shots – 26th

Henrik Lundqvist

Possession - 14th

Shots against - 6th

PP time - 9th

PK shots vs. - 5th

Blocked shots - 3rd

Pekka Rinne

Possession - 29th

Shots against - 20th

PP time - 29th

PK shots - 10th

Blocked shots - 4th

Let's talk about how these factors can alter save percentage:

Possession – Simply, the more the opposition has the puck, the more chances they have to get a “good bounce” or create a scoring chance

Shots against – While there is no statistical way to differentiate quality chances from a shot from the blue line, it stands to reason that the more shots a team faces, the more scoring chances they will have

PP time – The more a team is on the power play, the less time the opposition's best players are being put in a position to score. It's also a low stress point in the game

PK shots – Similar logic to shots vs., the more PK shots are taken against a goaltender, the more chance there is for a quality scoring chance. It also makes for more high-difficulty saves

Blocked shots – This one can can be tricky because often poor possession teams block more shots, so the numbers can come out as deceiving. But shot blocking is still generally helpful to a goalie as we see year after year in the playoffs

You will notice that while each Vezina Trophy finalist had a low ranking in one or several of the categories, Miller's team ranked poorly in all the categories. His team was abysmal at possessing the puck, they had very little power play time to give him a rest, he faced a very high number of difficult shots on the penalty kill and his team, despite poor possession numbers, did not block shots for him.

Lindy Ruff's system plays a role in Miller's save percentage, too. The statistics suggest that Ruff believes that Miller will bail his defenseman out when they are overly-aggressive and allow 2-on-1 breakouts.

I wrote this in my article about Ruff's system, but it's worth repeating to fill out the point:

Take a look at the Sabres' NHL rank in shots against per game

2011-12: 26th

2010-11: 17th

2009-10: 21st

2008-09: 22nd

2007-08: 12th

Now look at where they rank in shots for per game:

2011-12: 18th

2010-11: 4th

2009-10: 9th

2008-09: 10th

2007-08: 7th

The Sabres take a lot of shots and allow a lot of shots. Their defenseman are often jumping up in the play in attempts to contribute on the offensive end. This, many times, leads to more chances for the opposition.

Luck also might be a factor. When Miller won the Vezina Trophy, he had some good luck. Last year, he had some bad luck. How can that be determined? By penalty kill save percentage. Consider that far more goals scored on the penalty kill have nothing to do with the goaltender. Whether it be a tipped shot, a screen or a player being left wide open back door, there is often little a goalie can do to stop those shots. The luck part is whether his teammates block the shots, whether the tips hit his pads instead of going over his shoulder etc.

Take a look at Miller's even strength save percentages over the years compared to his PK save percentages:

Even Strength Save %

06 - .919

07 - .928

08 - .915

09 - .927

10 - .928

11 - .924

12 - .922

PK save %

06 - .899

07 - .846

08 - .866

09 - .879

10 - .919

11 - .886

12 - .885

His even strength save percentage was only .006 behind his Vezina Trophy year, but PK save percentage was its lowest since 2007 and .034 behind his Vezina year.

To demonstrate the inconsistency/luck factor of PK save percentages, look at Lundqvist's PK percentages over the past four years:

2008-09: .913

2009-10: .883

2010-11: .873

2011-12: .905

If you take into account that Miller faced 1,788 shots and only 262 were on the power play, you can figure that he was only slightly below same goalie last season that he was during his Vezina Trophy season the vast majority of the time.

The two questions we have no answered yet are: is he overpaid? And will he bounce back next year?

The overpaid question is a complex one. If Miller played for St. Louis, he'd be massively overpaid. Ken Hitchcock plays a hard puck possession game that allows an incredibly low number of shots/scoring chances to the opposition. Essentially any NHL goalie could have an above average year under Hitchcock. However, with Lindy Ruff's system, the Sabres need a top tier goalie. The “goaltending doesn't matter” catch-all only really applies to teams that ask very little of their goalies but also do not play aggressively on the offensive end.

In short – the Sabres' money is better spent on Miller than it would be if a defensive-minded team spent the same. Is he worth a cap hit of $6.25M to save three goals? No. Last season the Sabres did not get their bang-for-buck with Miller. That doesn't mean they never have or won't next year.

Speaking of next season, will he come back strong?

The health of Tyler Myers and Christian Ehrhoff certainly play a role here. Miller's health does too. Overall, the stats suggest his 5-on-5 save percentage will likely jump back from .922 closer to .926 to .928 range (or in the range of 10 goals difference) where it had been the previous two seasons and PK save percentage from .885 to around .900.

Also consider that Miller admitted to coming back from his concussion too soon. The numbers indicate as such. In December, which is when he came back from the mid-Nov. incident, he had an .893 save percentage – his lowest month.

So the short answer is yes, Miller should improve from last season.


Ryan Miller has been one of the best goalies since the lockout. His coach relies on him to make more saves than most coaches across the NHL do. Given the same situation, most average goalies would perform below average under the same circumstances. His contract, like all contracts, is only worth it if he plays at the top of his game. And the stats suggest he should bounce back from last season – which was still a better season than an average for an NHL goalie.

**One note about the "Lindy should rest Miller more" fallacy. On zero days rest, Miller's save percentage was .918 in 2011-12. On one day rest, it was .921. If you notice, those numbers are higher than his average for the season, meaning when he had more rest, he was worse!**


Do you think the Bills Stadium Committee meetings should be public?
  Yes, the public should know what's going on
  No, being public might affect talks in a negative way
View Results