# Applying CUSUM to hockey prediction models

This slideshow requires JavaScript.

The NHL season is a long and grueling affair and most teams will experience some ups and downs over the course of 82 games. Even a team that had a 67% chance of winning every game it played, would still have a 20% probability of putting up a five-game losing streak. And this is just straight probability theory with fixed probabilities. What happens when you consider all of the factors that go into determining the probability of winning an individual game, let alone predicting performance over an entire season?

Well, I’m not here to answer that question.

What I am here to do is to try to apply an analytical technique that was developed in the 1950s for the purposes of quality control in industrial and manufacturing processes to the game of hockey. Continue reading

# 25 Games In, What Does the Corsi Say?

Happy Max Corsi Productivity Day! We’ve reached the point in the season where Corsi best predicts future winning percentage. There’s plenty of more advanced ways to better predict how the rest of the season will go, but Corsi offers a simple baseline in a way that helps explain why it is so important.  I’ll first explain what that means and why it matters, then take a look at how we can use it to predict basic shifts in the standings for the rest of the NHL season.

# 2016-17 Hockey Graphs Top 50 Players

Embed from Getty Images

Welcome to the second annual Hockey Graphs Top 50 Players in the NHL list.

The main reason I put this together last year (you can view that here) was as a basis for comparison against the other, more famous, top 50 players lists. The annual list is a season preview staple for TSN and THN and the rankings are usually slightly controversial. Both lists are created via a poll of various people inside hockey, who are generally very smart people, but who are also prone to old-school thinking with value sometimes being shaped by recency bias, reputation and a winning pedigree.

This list is a bit of the opposite as it comes from mostly outsiders, people who study and analyze the game in the public sphere. That’s not to say these are necessarily smarter people, they just approach the game from a different angle based mostly on underlying trends and numbers over more traditional stats and what is immediately seen on the ice.

# Just How Important is Quality of Competition? Very. Also, not much. It’s All Relative.

*This post is co-authored by DTMAboutHeart and Ryan Stimson*

Recently, the topic of Quality of Competition has been at the forefront of Hockey Twitter. This post hopes to articulate some of the nuance associated with Quality of Competition, as well as Quality of Teammate, metrics and how impactful they are. To do that, we will revisit methods outlined here by Eric Tulsky, namely splitting the competition and teammate quality by position and measuring the impact of each split. Ryan recently wrote about this at the NCAA level, but it has not been looked at with much rigor at the NHL level.

Both Quality of Competition and Quality of Teammates matter. They also don’t matter. It depends on the position and metric you’re looking at. All TOI data is 5v5 and from Corsica. Ryan had the game files of who was on the ice during each 5v5 shot from Micah Blake McCurdy, so that data was used as well. Also, thanks to Muneeb for feedback during this process. Thanks to all!

# Tactalytics: Using Data to Inform Tactical Neutral Zone Decisions

Last time, I showed how using data and video evidence can be combined to inform tactical offensive zone decisions. Today, I’m going to do the same thing in the neutral zone. Neutral zone play is something that has been a hot topic among analysts for many years, going back to this paper written by Eric Tulsky, Geoffrey Detweiler, Robert Spencer, and Corey Sznajder. Our own garik16 wrote a great piece covering neutral zone tracking. Jen Lute Costella’s work shows that scoring occurs sooner with a controlled entry than an uncontrolled entry.

However, for all the work that goes into zone entries, there have been few efforts to account for how predictable these metrics are. At the end of the day, what matters is how we can better predict future goal-scoring. Also, in looking at our passing data, what can we also learn about how actions are linked when entering the zone? Does simply getting into the offensive zone matter? Does it matter whether it’s controlled or not? Or, does what happen after you enter the zone matter exponentially more? Lastly, what decisions can we make to improve the team’s process using this data?

# Measuring Single Game Productivity: An Introduction To Game Score

Embed from Getty Images

Who had the best game last night?

That’s a rhetorical question obviously… because it’s July, but when hockey is actually being played from October to June it’s an important question to ask – one that’s currently not very easy to definitively answer.

Some will look at points, some will look at shot differentials, some will watch the game, but rarely is there any consensus. Different people value different things. At the top and bottom of the spectrum the answer is sometimes obvious. If Connor McDavid has a five point night, he was very likely the best guy on the ice. If Pekka Rinne lets in five goals against on 21 shots he was very likely the worst. But for many games the answer is neither obvious or simple and is generally up for debate depending on an observer’s personal value system.

What we don’t have in hockey is a standardized measurement for single game productivity. It’s not something that will end any debate, but it can provide a much better framework to answer the question over what’s currently available. And that’s what I’m going to introduce in this post.

# Neutral Zone Playing Styles

Player A is a sniper. Player B is a playmaker. Quick: If the two of them get a 2-on-1 break, what do you expect each of them to do? Odds are you would expect the playmaker to pass and the sniper to shoot. You may not know how good each of these players is, but the monikers give you a rough idea of this player’s relative strengths and how they generally try to succeed.

We have plenty of different names that explain a player’s general “role”. We use words like sniper, dangler, two-way player, and power forwards (even if we can’t agree on what that last one actually means). However, these names are usually limited to the offensive zone. We have no easy way to describe what a player does in the neutral zone.

# Entry Generation and Suppression

### Intro

Hockey analysts have repeatedly shown the value of neutral zone play. If a player performs well in the neutral zone, he or she is helping generate offense for their team and limiting the opponent’s chances. In addition, neutral zone play is repeatable, and the player is likely to continue to drive possession for their team. If you can identify players who thrive in the neutral zone, you are in a position to help your team improve.

But while neutral zone play is important, we still have a very limited understanding of it. Between the distance from the goal, the fluidity of play, and the relative scarcity of data, most people don’t know how players perform in the middle third of the ice. Furthermore, we don’t even have a complete idea of how to make those evaluations. When figuring out how good a player is in the neutral zone, should offense and defense be evaluated separately, or are overall results enough? What skills translate to strong neutral zone play? What playing styles?

# Relationship of Possession to Shots-For

Here you go, Mike, you old stocky codger.

That is meaningful.

# How Can We Quantify Power Play Performance In Formation?

Last week I wrote about a new metric, ZEFR Rate, which measures zone entry success on the power play and is relatively repeatable and predictive of future goal scoring efficiency. The metric was based around the idea that getting into formation efficiently — most frequently a 1-3-1 — is a catalyst for power play success.

But now let’s say you’re a team that has perfected your entry scheme, and you find yourself setting up in formation at a consistent rate. What now? How can one maximize one’s use of possession in formation to score goals at the highest possible rate?