Blog Archives

Advanced stats are not scary

Advanced stats have gotten a lot of attention in the hockey world this summer, in part thanks to numerous hires by organizations around the league that indicate the debate is over as to whether NHL teams are using advanced stats as part of their evaluation process. While dubbing it the ‘summer of advanced stats’ (or ‘analytics,’ if you prefer) may discredit the wave that has been building in the hockey world for years, advanced stats are certainly part of hockey conversations more than ever.

A word on the term “advanced stats”: Does calling these stats “advanced” increase the divide, perceived or real, between the “watch the game” crowd and the “advanced stats” crowd? The current state of advanced stats within hockey uses a lot of the same basic math required to understand the more well-known stats we’ve all been familiar with for quite some time. While the stats may be different and, at times more in-depth, the math is largely the same. Using the word “advanced” may dissuade those unfamiliar with what “advanced” stats are.

Truth be told, both watching the games (scouting) and using numbers are essential, and you should be skeptical of anyone who tells you otherwise. Why are (advanced) stats important? Well, for a lot of reasons. First off, no one watches every game. On top of that, every fan I’ve ever spoken to uses numbers and stats as part of a conversation about sports. Advanced stats aim to provide the most accurate, telling, and/or predictive (depending on the stat and sport) statistics possible. Advanced stats do not aim to replace the sports that we all love with spreadsheets and data.  If we’re already using stats to quantify things about sports, shouldn’t we seek to quantify these things in the most meaningful way possible? Advanced stats are not the start of a new conversation. Rather, they are a continuation of and an attempt to better inform conversations we have been having about the games we love for as long as they’ve been around.

Today, Puck Daddy broke the news that the Maple Leafs have hired, among others, Darryl Metcalf, the founder of Extraskater.com. Extraskater.com has been the site I’ve used the most when combing through advanced stats. I’ve seen some people on Twitter wonder where to turn for their advanced stats needs now that Extra Skater is offline. The good news is, there are other similar sites out there. The two I have used the most (other than Extra Skater) are HockeyAnalysis.com and Behind the Net. With so many smart people out there that have an interest in advanced stats, I fully expect similar sites to pop up sooner rather than later (as I was writing this, I saw this tweet from Peter of RMNB). Editor’s Note:  More details are now available about Peter’s plan. Check them out and contribute to the conversation here.

The bulk of this post was initially intended for something else but it felt relevant today with the news about Extra Skater. To those of you opposed to or skeptical of advanced stats: Dig around online, ask questions, and don’t be scared of them. To those of you mourning the loss of Extra Skater: I’m with you, but there’s still useful resources out there and more help is likely on the way.

 

 

Who should play the point on the Caps power play?

When the Caps signed Matt Niskanen to a 7 year, $40.25 million contract, new coach Barry Trotz was given another intriguing option to play the point on the power play. Niskaenen is an offensively-gifted defenseman, finishing 12th among all defesemen in scoring last season with 46 points (10 goals 36 assists). While it remains to be seen if Trotz will use 2 defenders on the PP or skate one defender with Alex Ovechkin playing the other point, the Caps coach has 3 formidable PP options from his defensive corps in Niskanen, John Carlson, and Mike Green.

One consideration when deciding how to divvy up ice-time, as pointed out by Peter from RMNB, is that Carlson plays a significant amount of time on the penalty kill. We don’t yet know how the coaching and personnel changes in Washington will impact PK deployment, but last season Carlson saw the ice for 65.2% of the time the Caps were on the PK, while Green played 8.6% and Niskanen saw 13.8% in Pittsburgh.

Glossary for chart

FF% rel-The % of unblocked shot attempts the player’s team saw with him on the ice, relative to with him off the ice

SA per 2:00-The number of total shot attempts (SA) for a team with the player on the ice, per 2:00

Setup Passes per 2:00-Estimated number of passes by a player that led directly to a shot attempt, per 2:00

All of the numbers are PP only. Per 2:00 was obviously chosen to represent the time of a standard PP.

Player FF% rel SA per 2:00

Setup Passes per 2:00

Carlson

+6.3% 3.91

0.83

Green -8.3% 3.35

1.08

Niskanen +0.8% 3.54

0.31

-The Caps perform significantly better, FF%-wise, with Carlson on the ice during the PP while the Caps saw a significant drop in FF% with Green on the ice during the PP.  This is interesting, considering Green was the Caps best possession player in 5-on-5 close game situations. The Pens PP generated slightly more unblocked shot attempts with Niskanen on the ice.

-The differentials in SA per 2:00 may look insignificant but they add up quite a bit when given proper context. The Caps averaged 3.54 PP per game, so the difference between Carlson (13.8) vs Green (11.9) amounts to just about 2 additional SA per game in a hypothetical world where they each separately play 100% of the team’s PP minutes.

-Green has a significant advantage in setup passes per 2:00.  Discussing his numbers relative to Niskanen’s here is probably apples to oranges, since they played on different PP. There are all kinds of factors, such as set plays and who they are funneled through, that could impact this rather than passing ability and vision. For example, one possible reason Green’s numbers are so high is that he was often funneling the puck to shooting machine Alex Ovechkin for his patented PP one-timer. However, when comparing teammates, Green (3.82) would produce nearly one more setup pass per game than Carlson (2.93) were each to play 100% of the team’s PP minutes.

-One interesting tidbit not shown in the chart is that opponents’ average shorthanded shot distance with Carlson on the ice was 52.7 feet compared to 34.6 feet with Green on the ice. That’s the difference between 7.3 feet inside the blueline vs. 5.4 feet inside the top of the faceoff circle.  While not proof in and of itself, those numbers could serve as a launching point for someone who wants to show that Green is prone to take risks that can backfire (I’m skeptical of that theory).

Player

Individual SA per 2:00

Carlson

0.99

Green

0.91

Niskanen

0.91

-Carlson generates .08 more SA per 2 minutes of PP time than Green and Niskanen (Like above, reading too much into Niskanen’s numbers at face value vs. Green and/or Carlson is probably a bit faulty since they played on different PP). Last season, the Caps average 3.54 PP per game. Over the course of a season, at 3.54 PP attempts per game,  Carlson would generate 23.2 more PP SA than Green or Niskanen, were they each to play 100% of their team’s PP minutes (this also assumes, for the sake of simplicity, that each PP attempt is 2 minutes) . The Caps scored on 8.3% of their PP SA this past season. This would mean that, through his SA alone, Carlson would generate 1.93 more PP goals per season, in a hypothetical world where all 3 players assume the Caps average PP goals per shot attempt % from last season (8.3%) and the Caps average 3.54 PP per game for a season.

From all of this data, it appears that John Carlson should be first choice to play the point on the PP in 2014-15 for the Caps. The team generates more unblocked and overall shot attempts with him on the ice, and he personally takes  more shot attempts than Green or Niskanen. However, should Trotz choose to manage Carlson’s PP time in order to keep him fresher for PK and 5-on-5 play, and depending on what position Ovechkin plays on the PP, both Green and Niskanen offer Trotz viable options.

Agree? Disagree? Feel free to comment below or send a tweet. Follow us on Twitter here.

S/T to ExtraSkater.com for being such a great resource for this article (and in general). 

Advanced stats highlights from the Caps 2010-11 season

Advanced stats for the 2010-11 season are now available on ExtraSkater.com. I highly recommend visiting Extra Skater to see the stats for yourself, but below, after a quick refresher on the season, are some Caps highlights I pulled from a glance at the new stats on Extra Skater.

With a record of 48-23-11, the Caps were not only Southeast Division champs, but finished first in the Eastern Conference, with 107 points. The Caps discarded the Rangers 4-1 in the first round of the playoffs before being swept by the #5 seed Tampa Bay Lightening in the second round. This was also the season during which the Caps, under Bruce Boudreau, shifted to a more defensive-oriented system. The Caps defeated the Penguins 3-1 in the 2011 Winter Classic and were featured on the HBO series 24/7.

 

download

 

download (1)

 

download (3)

-The Caps finished 15th in Fenclose%, tied with the Kings at 50.4. The two teams have gone in opposite directions since, with the Caps (47.5%) finishing 25th in 2013-14 and the Kings (56.7%) finishing first.

-Nicklas Backstorm led all qualifying (41+ games played) Caps in FenClose rel at +5.0%, followed by Alex Ovechkin at +3.9% and Alex Semin at +3.5%. The top Caps defender was John Carlson at  +1.9%.

-Alex Semin’s PDO of 107.7 ranked highest on the team, aided by his teammates’ on-ice shooting % of 10.5%, 1.8% higher than any other Caps player.

-Marcus Johansson had the highest ZS% at 58.1%, while Boyd Gordon faced the toughest zone starts, with a ZS% of 41.8%

-As would be expected, the Caps top forward line of Ovechkin, Backstrom, and Mike Knuble faced tougher competition than any of their teammates. Carlson and Karl Alzner faced the toughest competition of any Caps defenders.

-To the surprise of absolutely no one, Mike Knuble had the shortest average shot distance (24.7 feet).

-Alex Ovechkin was on the ice for 82.4% of the Caps PP minutes during the season (this past season Ovechkin saw the ice during 93.2% of the Caps PP minutes).

This was after a quick look at the stats. If you look through Extra Skater and find any interesting tidbits, leave them in the comments below or give us a shout on Twitter.

Hockey writer and analytics specialist Neil Greenberg on The Tony Kornheiser Show

Hockey writer Neil Greenberg was a guest today on The Tony Kornheiser Show. Kornheiser has referenced Greenberg’s writing recently on-air, appearing to become somewhat of a fan of the advanced hockey stats expert and Washington Post, ESPN contributor. Here are some highlights from today’s appearance, Greenberg’s first ever on the show:

Greenberg on drinking from the Stanley Cup after the Rangers won it in 1994:

“I was frequenting some of the bars in Nassau County…there were people taking pictures with the Cup outside of one of the bars, so we stood in line. As soon as I turned the corner to enter the bar, someone said, ‘Do you want to drink from the Stanley Cup’ and I said ‘Absolutely.'”

On Alex Ovechkin’s decline:

“He’s getting older. There’s two things that are happening. One is just your general age progression. People have a misconception as to when hockey players, especially goalscorers, peak. It’s typically between the ages of 22 and 24 and then you start to see a down-slope at age 27. Ovechkin has pretty much been in that exact pattern.”

“If you look at how Ovechkin scored his goal early on in his career when he entered the league it was by a volume of shots. It was more quantity than it was quality. He led the league with 425 shots when he came in. He was just a completely dynamic player that tossed rubber at the [net] from ever angle. And as those shots per game decreased, it took the goal scoring along with it. And what we’re seeing now is a player who’s putting up 300 shots on goals as opposed to 500 and 400 shots on goal. So, the goal scoring numbers are going to come down. Now, to his credit those shots on goal have up-ticked a little this year but a lot of goal scoring is also luck. You have a clank of the pipe here, it goes through the wickets there. He hasn’t been seeing a whole lot of puck luck as maybe he’s had in the past. He’s just going to be the 50 goal scorer that he was probably ever again.”

On if Ovechkin is an assist guy, a set-up guy like Gretzky or Lemiuex, with room to elevate his game:

“Ovechkin is a good passer. I think that’s actually one of the most underrated parts of his game. However, he’s not Gretzky, he’s not Lemieux. He’s not going to tally a whole bunch of points from… [Kornheiser jumps in and says, "He's not Crosby."] He’s certainly not Crosby. So, he’s not going to be getting points that way. The switch to right wing was an effort to get him away from that overpowering move he had down the left side where he would go down the left wall and try to cut in and score the goal that way, because that just wasn’t working anymore. The defenses have caught up to that. So, you’re right. His points are going to come from goals and, unless he starts to adapt his game a little bit more on the right wing, we’re going to see some 25 to 30 goal seasons.”

On Ovechkin making “a lot of money”

“That’s where people’s expectations, I think, are coming unglued because they see on paper this 65, 50 goal scorer that’s making $10 million and they think that’s going to happen in perpetuity. But, look, the Ovechkin contract was a bad contract. When they signed Ovechkin to that contract, for that period of time, it was a bad deal. And you can never expect a goal scorer to score 60, 50, 60 goals a year for ten years. It’s just not reality.”

On if the Caps know what Greenberg knows about Ovechkin

“I think so. Hockey analytics has definitely become bigger in the past couple years, however you still have, and I’m not speaking specifically about the Caps, but in the NHL in general, there are some teams that are embracing it. I know that Tampa Bay has a hockey analytics guy on staff, Chicago, Pittsburgh, Calgary, some other teams are embracing it. But, as far as the Capitals are concerned, they seem to trust the coaches more than the numbers. An I’m not saying that that’s a bad thing. However, I think that there’s room for both, especially when you’re looking at player evaluation.

[I'm going to interrupt this transcription, as I've just read on Twitter that Russian Machine has sent out a transcription of the entire interview. So, rather than duplicate work, check the Russian Machine Never Breaks transcript of Neil Greenberg on the Tony Kornheiser Show by Peter Hassett and I'll go write something else.]

 

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: