Why popular advanced stats are bad at predicting who wins hockey games

Check out this post on the new version of Rink Stats.

A few days ago, Jonathan Willis at Bleacher Report had a great article introducing his readers to a few of the most common NHL advanced statistics. He breaks the stats up into team statistics and player statistics. For team stats he mostly focuses on Fenwick and Corsi, which are the two most popular way that hockey analysts like to measure shot attempts.

Fenwick and Corsi are commonly thought of as a good way to capture how much a team is dominating offensively, since teams only attempt a shot when they're in their offensive zone. In other words, Fenwick and Corsi are thought of as proxies for how much time a team spends in their offensive zone. A natural conclusion to draw from this is that if your team dominates your opponent in the Fenwick and Corsi stats, then most of the game was likely spent in your offensive zone and your team probably won the game.

In this post I'm going to dig into Fenwick and Corsi and show that neither stat is very good at predicting who wins hockey games in the NHL. In fact, I'll show that the simple shots-on-goal stat does a better job at predicting who will win a game than Fenwick and Corsi.

What wins hockey games?

Check out this post on the new version of Rink Stats.

What wins a hockey game? We often hear from announcers things like, "The team that wins in the faceoff circle will win this game" or "They've got to get the puck on the net if they want to win tonight." But which of these statements are true? More generally, which team stats are best at predicting who wins hockey games?

I take a first stab at that question here by looking at what correlates with winning hockey games. I use seven years of NHL play-by-play data to generate statistics like shot-differential, hit-differential, faceoff win-differential, etc. and then use those statistics as variables in logistic regression in order to evaluate who wins games and why.

What surprised me is that the statistic that I would have guess correlates most strongly with winning (shots on goal) is highly correlated with winning, but in the wrong direction. That's to say, the team that takes more shots in a game is, on average, less likely to win the game. More predictably, I also find that winning faceoffs and being the beneficiary of turnovers both positively contribute to a team's chances of winning.

Rink Stats in the Boston Globe!

Check out this post on the new version of Rink Stats. I was really excited yesterday to see that my work on NHL realignment had generated some publicity beyond the internet. At the New England Symposium on Statistics in Sports a couple weekends ago, I talked with Fluto Shinzawa, the Bruins beat writer for the Boston Globe, about my work on why the realignment is unfair for Eastern Conference teams.

In yesterday's Boston Globe he had incorporated some of my research findings into his preview of the 2013-2014 NHL season. Check it out here. And just to prove that it really appeared in print, I bought a copy of the paper and took a picture of the article:

Realignment gap conference poster for NESSIS

Check out this post on the new version of Rink Stats. Tomorrow I'm going to be doing a poster presentation about the work on NHL realignment at the New England Symposium on Statistics in Sports. The conference is hosted by the Harvard Statistics Department and should be really fun. There's presentations about just about every sport, as well as discussions with people who have worked in statistics departments for teams.

Here's a bigger version of the poster that's easier to read. And check out my original post about my realignment work here.

Should college football teams ever bring kickoffs out of the endzone?

Check out this post on the new version of Rink Stats.

The end of the summer was pretty busy for me, so I haven't had time to do any posts on here. But I'm back now, although with a short post about something that's not hockey related. Hopefully you'll still find it interesting.

While watching college football with friends last weekend, I was reminded of how frustrated I get when teams bring kickoffs out of the endzone but don't have at least a 25 yard return. For anybody unfamiliar with college football rules, any player who catches a kickoff in the endzone can kneel down, and his team will start their drive on the 25 yard line (20 in the NFL). The player also has the option to run the ball out of the endzone, and his team starts their next drive where he is tackled.

The main question here is, should a kick returner ever bring the ball out of the endzone, if he catches it there? If the average length of kickoff returns that start in the endzone is at least 25 yards, then it probably makes sense to run the ball out. If the average length is less than 25 yards, then he probably should kneel down and take the 25 free yards. From some basic analysis, I find some pretty compelling results that suggest that the latter strategy should be the dominant one. Specifically, I show that about two-thirds of all kickoff returns that start in the endzone do not make it to the 25.

Why the NHL's new conference alignment is unfair for Eastern teams

Check out this post on the new version of Rink Stats.

On Friday the NHL released the 2013-2014 regular season schedule. This season will be the first played under the new conference and division alignment which sent Winnipeg west and Columbus and Detroit east. A key aspect of the new alignment is that there is an imbalance in the number of teams in the East (16) and in the West (14). In this post I'll discuss a paper I wrote a couple months back (How the West will be Won: Using Monte Carlo Simulations to Estimate the Effects of NHL Realignment). In the paper, I show that because 8 teams make the playoffs from each conference, the new alignment and playoff qualification rules unfairly disadvantage Eastern Conference teams.

Specifically, I'll show that the 8^th seed in the East will (on average) be 2 or 3 points better in the standings than the 8^th seed in the West. And I'll show that about 40% of the time, the 9^th seed in the East would have made the playoffs if they were in the West (compared to just 20% of the time when the inverse is true).

Should you shoot first or second in a shootout?

Check out this post on the new version of Rink Stats.

I've always been interested in the strategy that goes into shootouts in the NHL. Each coach must first choose three players to take the shootout attempts for their team, and then they have to decide the ordering of those players. In addition, the home team gets to decide whether to shoot first or second.

80% of the time, the home team chooses to shoot first in the shootout. In terms of strategy, this seems like the proper course of action. Somebody better versed in economic game theory could probably prove this better than me, but the basic intuition is based on the fact that the probability of scoring on any shootout shot is less than 50%. If the team that shoots first scores, then the other team's shooter probably feels more pressure to score. Likewise, if the first team fails to score, the odds are still in that team's favor that the shootout score will remain tied (since there's still less than 50% chance of the second team to score).

In this post, I want to see whether the team that shoots first actually wins more shootouts. By looking at the 900+ shootouts that have happened in the NHL since 2007-2008, I find that there appears to be some evidence that shooting first gives you a slightly better probability of winning. Somewhat surprisingly though, I also find that teams tend to do worse in shootouts at home than on the road.

Are shorthanded goals momentum killers?

Check out this post on the new version of Rink Stats.

Shorthanded goals are powerplay killers. You hear this all the time when you watch hockey games. But how true is this conventional wisdom? It certainly doesn't seem very common for a team to concede a shorthanded goal and then quickly follow it with a powerplay goal of their own. But it's also true that a majority of powerplays don't result in a goal, regardless of whether a SHG was conceded or not.

The question of how much SHGs kill a team's momentum on the powerplay is one that I've always thought about. Below, I take a first crack at answering the question. I show that having a SHG scored against you lowers your probability of scoring a PPG by about 10 percentage points. And early shorthanded goals have a bigger effect on a team's powerplay momentum that late ones.

How unusual was the ending of Monday's Game 6?

Check out this post on the new version of Rink Stats.

As a follow up to yesterday's post on the latest Cup-winning goals in history, today I'm going to look at little more at the craziness of the end of Monday's game. Earlier today I was thinking about how often it is that a team comes from behind to tie and win the game at the end of regulation, like the Blackhawks did a couple nights ago. My inclination was that it can't possibly be a common thing. So with my six seasons worth of regular season play-by-play data I explored the question. Here's what I found:

The Blackhawk's Cup Clincher, in Perspective

Check out this post on the new version of Rink Stats.

Last night's Stanley Cup Final Game 6 will almost certainly go down as having one of the most memorable in Cup history. But how does the Blackhawks' comeback measure up against history? As you might guess, their game tying goal with 1:16 left in the 3^rd period is the latest GTG to come in a Stanley Cup clinching game. Their game winning goal with :59 remaining is the latest Stanley Cup winning goal to come before overtime. Here's the list of the latest Cup clinching goals:

Hitters and Hittees in 2013

Check out this post on the new version of Rink Stats.

In Game 5 of the Stanley Cup Finals last night, the Blackhawks came away with a 3-1 win over the Bruins. Chicago outshot Boston, took fewer penalties, and won more faceoffs. But one interesting battle won by the Bruins was in the number of hits delivered. Boston players were credited with 53 hits, whereas Chicago only registered 22. Interestingly, Boston has outhit Chicago (often by a large amount) in every game except Game 1, in which Chicago has 61 hits to Boston's 59.

For each NHL game, you can find an event summary (like this one) which provides the total number of hits credited to each player. What these summaries do not provide is the number of times each player is the recipient of a hit from the other team. These hit-recipient data do exist, however. In this post I'm interested in looking at which players received the most punishment in the 2013 regular season, and whether there are distinct patterns which can teach us about in-game strategies of teams and players.

Testing two common adages about when goals are scored

Check out this post on the new version of Rink Stats. Hockey announcers and fans seem to have a lot of adages about when goals are scored during games. I just finished collecting and cleaning play-by-play data from every NHL game from the 2007-2008 season through the lockout-shortened 2013 season. In this post I'm going to look at some very basic attributes of the data. I also hope to convince you of the truth behind two common adages.

Canada's Cup drought and the salary cap

Check out this post on the new version of Rink Stats. The lockout-shortened 2013 NHL season marks the 19th straight season in which the Cup will not be raised by a Canadian team. In a blog post yesterday, Nate Silver took on the question of why Canada is on such a dry spell. In the article, he puts forth three explanations. First, that Canadian teams have just been on a bad luck streak. Second, that changes to the NHL's salary cap rules changed at a time when the Canadian dollar was gaining strength against the US dollar. Third, that there are too few Canadian teams, relative to the sport's popularity in different markets. The article is well put together, there were a couple things that stood out to me about his arguments.