In most fantasy leagues, the crutch of your team will be your goaltending. In any sort of rotisserie or head-to-head league, there’s a good chance that goaltending categories are at least 30 percent of your total categories, sometimes up to 50 percent. And it’s one of the most volatile and unpredictable positions in sports; of the top-5 goalies in a standard Yahoo! league for the 2012-2013 season, none of them were top-5 in 2011-2012. Only Henrik Lundqvist and Cory Schneider were Top-10 in both seasons. That’s an 80 percent turnover rate at the top.
So far this early season, only Tuukka Rask and Antti Niemi are the only top-10 goalies from last year that are also there again this year.
The question is how to figure out which goalies will maintain their hot starts, which will not, and why this happening.
One way to look for regression in goaltenders is to look at their team’s penalty kill rates and their own power play save percentage
Power Play Save Percentage (PPSV%)
One of the biggest factors in a goalie’s final numbers will be the number of goals that the goaltender allows while on the penalty kill. An elite PPSV percentage is anywhere .900 and above; only six goalies achieved that mark in 2011-2012 (min. 50 starts) and only two goalies in 2012-2013 (min. 30 starts). It’s also very dependent on the penalty kill of the team in front of him:
Of the bottom five teams in penalty kill percentage last year, the combined overall save percentage of the teams’ five regular starters (Braden Holtby, Ryan Miller, Cam Ward, Pekka Rinne, Jacob Markstrom) was .912. A top-20 mark for goalies that year was .915.
Of the bottom five teams in penalty kill percentage the year before, the combined overall save percentage of the teams’ five regular starters (Mathieu Garon, Corey Crawford, Jonas Gustavsson, Antti Niemi, Steve Mason) was .904. And it could have been much lower, as Gustavsson only started two more games than James Reimer, and Reimer allowed 13 more power play goals against than Gustavsson on only seven more shots. A top-20 mark for goalies that year was .916.
In a full season, the elite-to-very good starters in PPSV% will allow 25-35 goals on the penalty kill. Even that spread though, just 10 goals on 2,000 shots, is enough to move a goaltender’s save percentage from .925 – which is fringe top-5 – to .920 – which is fringe top-10. Many goalies only get to about 1,800 shots against, thus in several cases the effect is exaggerated. So when you talk about allowing maybe 20-25 goals which the very elite allow, to allowing 35-40 goals like the many of the worst goalies on the penalty kill do, you could be looking at around a full percent lost in your save percentage (~.925-~.915). That’s huge.
It’s important to remember that many bad penalty kill teams are bad teams defensively overall. There are instances were the penalty kill outperforms the quality of the defense (Buffalo, Toronto), but that’s not usually the case.
The last point is that about 20 percent of a goalie’s goals allowed in a year will be from the penalty kill. So while you want goalies that have good even-strength save percentages, as 5 on 5 play is when the vast majority of goals are scored, that 20 percent makes the difference between a good season and a great season, or a good season and a bad season.
Just to show you visually the impact of the save percentage while short-handed on a goalie’s final numbers, this is every shot and every save of every goalie who started 50 games in each of 2010-2011 and 2011-2012 as well as every goalie that started 30 games in the lockout year. The horizontal axis is their final PPSV%, the vertical axis is the overall save percentage for that year. It’s pretty much a linear relationship.
 
It’s all well and good to have a good goaltender on your team. But if the penalty kill in front of them is bad, there’s little they can do.
This is what you can expect to happen with a few of the hotter starters for the season
Ryan Miller (BUF) – .941 Power Play Save Percentage
Miller is off to a great start this season, posting a .926 save percentage through his first seven starts. The problem is that his .941 PPSV percentage is at least 40 points too high.
Incidentally, Miller’s even-strength save percentage is .925, which is about his talent level considering his team. Buffalo’s penalty kill has been surprisingly good, coming it at seventh in efficiency (84.4 percent) to start the year. So while his PPSV percentage won’t plummet down to Rinne-In-2013 levels, it has to go down. Even if he finished at a .900 PPSV% for the year, which is very good remember, he would still allow 25 more goals than he’s currently on pace for. At that rate, the expected drop in overall save percentage will be around 1.5 percent, and that’s an enormous spread (going from .926 to .911). Considering Buffalo is one of the worst teams in possession and shots allowed, and Miller’s last three seasons have been save percentages of .915, .916 and .916, it’s not unreasonable to think he’ll end up with a .911 this year.
The problem with Miller is that you can’t sell high on him because his team has managed one win in 10 games. At best, you can expect a season similar to last year, when he was a fringe Top-30 goalie.
Jonathan Bernier (TOR) – .926 Power Play Save Percentage
Determining the level of regression with Bernier is a little more of an uncertainty because the Leafs penalty kill was pretty good last year, finishing second in efficiency at 87.9 percent and just 18th-worst in times shorthanded at 157. This year, the Leafs are fifth in penalty kill efficiency at 86.8 percent but are tied for seventh-worst in times shorthanded at 38.
If Bernier wants to maintain an elite save percentage, the Leafs need to stop taking penalties. The efficiency of the Leafs penalty kill is much less meaningful for Bernier if they take more penalties, as his numbers are tied to the volume of goals allowed; the rate at which goals go in on the penalty kill than at 5 on 5 increases about 60 percent. If he has a higher percentage of shots against on the penalty kill compared to all of his shots, both his goals against average and save percentage will almost certainly suffer.
It’s not to say that Bernier’s .926 on the PK is going to go down a lot, it really may only go down to about .910, which is still brilliant. But if the Leafs keep taking penalties at their current rate, regardless of the efficiency of their penalty kill, Bernier’s numbers could take a big hit.
Ben Bishop (TBL) – .900 Power Play Save Percentage
Another goalie off to a great start this year, Bishop seems to have finally found his home in Tampa Bay.
The problem is that Tampa Bay’s penalty kill isn’t very good; at 78.8 percent, the Lightning are tied for 20th in penalty kill efficiency. Once you get below the 80 percent mark, you start getting very close to the bottom-5 territory that I talked about earlier. With 33 times short-handed in just eight games, the Lightning are averaging four penalty kills per game. This is the same problem that Bernier will run in to, but to an even more severe degree as the Leafs have played 10 games; even if Bishop stays solid at even strength and the penalty kill stays static, the volume of goals are going to kill his numbers.
The final problem is that penalty kill efficiency. If that doesn’t improve, Bishop’s numbers will plummet:
Every regular goalie that finished with a .900 penalty kill save percentage or better from 2010-2013 had a top-10 penalty kill in front of them.
For these reasons, the Tampa penalty kill has to improve drastically in efficiency, along with the rate at which they take penalties, in order for Bishop to maintain anywhere close to his current .924 save percentage and 2.11 goals against average. If you can sell high on him now and get a goalie that had a slower start like Jonathan Quick, I’m all for it.
Of the three goalies, I would say Bernier has the best chance of finishing the season as a top-15 goalie. Tampa and Toronto might be close in overall team play, but Bernier’s team boasts the best penalty kill. That small edge could mean the difference between a top-15 ranking and top-25 ranking.
Remember, the majority of hockey games are decided at 5 on 5 play, so that’s where the majority of goalie stats come from. That’s why you want to draft a goalie that consistently has a good even strength save percentage, it creates a floor for his downside (Rinne’s .910 last year is about as bad as it can get for him). However, the penalty kill save percentage, over which the goalie has almost no control, is about the team.
It’s important to keep an eye on penalty kill rates to see what direction your goalie’s stats might go in as sample size is the issue here. While Bishop can’t maintain his level at the current TB efficiency, the Lightning are just a 7/7 stretch away from getting to 82.5 percent on the penalty kill, which would be league average or better most years.
I’ll have periodic updates through the year on how these goalies, and others, are doing while their team is short-handed.
