Fantasy Hockey: Eastern Conference Goaltenders and Blow-Up Rates

Last week, I wrote about how a team’s possession rates probably won’t make a goalie’s stats much worse unless the goalie sucked to begin with (Steve Mason), but how a team’s penalty killing can affect just about any goaltender for better (James Reimer in 2013) or for worse (James Reimer from 2010-2012).

There’s a concept that I came across a few years ago and was recently expanded upon by a blogger at the ‘Nucks Misconduct section of SBNation. This concept was that of goaltender blow-ups.

A goaltender blow-up in general is when a goaltender performs so poorly that it gives his team a very low chance of actually winning the game. This stands to reason in other sports as well; I would doubt that a football team’s winning percentage is very good if their quarterback throws four interceptions and a baseball team wouldn’t have much of a chance of winning if their starting pitcher gives up six runs in the first two innings.

Specifically, a goaltender blow-up is when a goalie achieves one of the following conditions:

1. A save percentage under .850 for a specific game.
2. Allowing five goals or more while facing less than 40 shots in a given game.
3. To avoid double-counting, games where a goalie allows 5+ goals on less than 40 shots and has a SV% under .850 is only counted as one blow-up, not two (and gets counted as a 5+ goal BU).

There’s a reason for this. In the blog I linked earlier, the writer states that if a goalie suffers a blow-up, this gives his team a 15% chance of winning the game. The 8^th seed in both conferences (the play-off cutoff) the last two years won 131 of a possible 260 games, slightly over 50%. So if a mediocre team has their all-things-equal chance of winning a game cut by more than two-thirds simply because their goaltender had an off-night, it’s worth exploring from a fantasy perspective.

However, all this is at the macro level, e.g. how a goalie performed last year or a two-year average. The problem with fantasy hockey is that a macro-level analysis doesn’t necessarily translate and this is the pitfall of small sample sizes. This is not more evident anywhere than in head-to-head fantasy hockey.

This seems like a huge tangent to essentially say this: you can estimate who will be a leader in goalie wins (Corey Crawford-CHI) and who won’t (Jacob Markstrom-FLA), but you don’t know when they will happen. A goalie that wins 60% of their starts should help your team more than a goalie who wins 50% of his starts, but if that 60%-goalie goes 0-3 in your fantasy hockey championship week, it’s all for nothing.

So we can’t predict when a goalie will get their wins, but we can estimate how many they will get. We also can’t avoid goalie blow-ups, but we can track their frequency as well.

As anyone who has played head-to-head fantasy hockey or baseball can attest to, one terrible start by a goalie or pitcher can ruin a week. In a league that has 10 categories, a goalie’s SV% and Goals Against Average will account for 20% of your stats. One horrendous start can eliminate you immediately from winning 20% of the categories. That is a pretty hefty price.

So while we can’t predict when these blow-ups will happen there are three indicators I use when looking to project how many blow-ups there might be:

• A goaltender’s track record. If a goalie has a history of seven or eight blow-ups a year, it’s pretty safe to say we can expect four or five next year.
• Teams that take a lot of penalties. There were 11 teams that were short-handed at least 167 times last year and four of those teams were under 80% on the penalty kill. As I showed yesterday, team penalty killing has a good correlation with a goaltender’s overall SV%. A bad penalty killing night can lead to a blow-up even if the goaltender is “on his game”.
• Team possession. While good goalies are typically good wherever they play, bad goalies need a good team in order to have good numbers. It may not correlate strongly with save percentage but it will in goals allowed; of the bottom nine teams in goals allowed at even-strength/60 minutes in 2013, all but one (Carolina) were a sub-50% possession team. The year before, of the bottom 12 teams in GA/60 again, four were a plus possession team. In short, if a team has bad possession numbers, the odds are overwhelmingly against the goalie to post a good GAA.

That last point is most important. A bad possession team will give up more shots, which can actually help a goalie’s SV%, but won’t be conducive to winning. I’ll give you an example of a hypothetical game:

• Goalie A saves 50/55 shots, giving him a .909 SV% but a goals against of 5.00 (assuming a full 60 minutes).
• Goalie B saves 25/29 shots, giving him a SV% of .862 but a goals against of just 4   .00 and in a head to head match-up, the win.
• Despite Goalie A having a better game than Goalie B, Goalie B would “win” two out of three head-to-head categories.

This isn’t a random example. This happened to Colorado’s Semyon Varlamov on February 16^th of this year against Edmonton. The combination of Devan Dubnyk (who got pulled) and Nikolai Khabibulin performed much worse than Varlamov, yet skated away with a better combined GAA and the win.

What this tells us is that you can’t predict goaltending outcomes on a specific day/week, but if your goalie(s) play well you should win at least 1/3 goaltending categories, and that’s better than 0/3. You run into the problem of going 0/3 when your goalie blows up.

With that in mind, here are the blow-up rates for Eastern Conference goalies that were their team’s starter in 2012-2013 over the last three years:

Sorted by Team CorsiFor% (the percentage of all shot attempts by both teams in a game that belong to one specific team. This percentage shown tallies for the last three seasons)

Some few notes about what’s presented above:

• I excluded Tampa Bay from the conversation entirely. Who gets the most starts for them is a toss-up at best. Ben Bishop would seem to have the inside track, but saying he’s a lock for 60+ starts is a stretch. Also, they have had nine different starters in the last three seasons.
• Starts/BU in red are below what we expect for that quality of team, green is better than what we should expect, black is about par for the course.
• I included Braden Holtby at the bottom but don’t really read too much into his numbers, the sample size just is not big enough. Generally, 100 starts is a decent baseline – you will see others with less than 100 starts, but they’re more established goalies than the 23 year-old Holtby.
• This conversation starts with Henrik Lundqvist. We can talk about how John Tortorella’s system might have an effect, but Lundqvist is so far ahead of the rest of the field that there’s much more at work than a team philosophy i.e. there’s a reason they call Henrik ‘The King’.
• None of the top CorsiFor% goalies have an exceptional blow-up rate. Of the top five CorsiFor% goalies in the West, four of them had at least 7 games between blow-ups. We don’t see a goalie with at least 7 games between blow-ups in the East until Cam Ward.
• Speaking of Ward, it seems he’s as good as he ever was. His numbers get skewed a bit because of the workload he usually takes in – Carolina allowed the most shots on goal/game in the NHL from in both 2010-2011 and 2011-2012, also were fifth-worst this year.
• Craig Anderson’s numbers are also skewed a bit; he had 12 of his 22 blow-ups come in 2010-2011 splitting time with Colorado and Ottawa (49 starts). He’s only had 10 blow-ups (84 starts) in the last two seasons.
• Ryan Miller and James Reimer are two goalies that take a lot of criticism (probably because of their markets) but manage to keep their mediocre-at-best teams in games better than most. Maybe it’s time to stop looking between the pipes and start looking at the other 20 guys on the roster, Sabres and Leafs fans.
• Jimmy Howard leads the list in blow-ups in terms of a SV% under .850. This is not a stat you should lead in playing in front of the best possession team in the NHL for the last three seasons.
• Martin Brodeur and Ilya Bryzgalov are tied for worst blow-up rate on this list (Holtby excepted). One is a 41-year old hall-of-fame goaltender, the other doesn’t have a contract.
• Carey Price is tied with Bryzgalov for most 5+ goal, < 40 shot blow-ups and that’s not good for Canadiens fans. This type of blow-up tends to mean bad team defense – there’s a reason Nabokov, Bryzgalov and Theodore are at the bottom of this category as a rate – so there should be cause for Habs fans to not get too caught up in strong possession numbers and young, talented forwards.
• While goalies can control their SV% more than their Goals Against Average, don’t read too much into lower goalies having better < .850 SV% blow-up rates – many 5+ goals, < 40 shot games are also < .850 SV% games, so they don’t get double counted (also a reason for Howard’s high < .850 SV% BU rate).

I hope this helps you in your fantasy hockey draft prep. Remember conference realignment means Detroit and Columbus got their stats in a different Conference (for the most part) so these numbers won’t hold constant for the next couple of years. However, it gives us part confirmation of what we already knew – Rask, Lundqvist are good – but also shows us some goaltenders that we might think are not-so-good are constantly bailing out their team – Reimer, Ward, Miller.

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 FantasyTrade411.com. He can be reached on Twitter @SlimCliffy for any fantasy hockey questions. Follow @SlimCliffy !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.id=id;js.src=p+'://platform.twitter.com/widgets.js';fjs.parentNode.insertBefore(js,fjs);}}(document, 'script', 'twitter-wjs');

See Full Bio