Encyclopaedia Analytica

2015-11-28 12_52_17

It seems as though every season the footprint of the analytics movement in hockey grows a little larger, to the point today where possession numbers aren’t just bandied about but rather there are entirely different strains of what used to just be called “Corsi”. The wealth of information available to us on every player’s performance eclipses what was available even five years ago.

With that in mind, I’ve run through a few of the more commonly-used analytics sites to give readers here a quick introduction not to the numbers themselves (because there’s really no way to make that quick) but rather to the sites themselves. My goal is to give you a walking tour through the landscape of websites who host the numbers in the hopes that this will encourage you to explore this world on your own.

This article is intended for the amateur and casual stats people. Fans who like some of the numbers but aren’t knee deep in them and still have to pause for a moment to remember the difference between Corsi and Fenwick (or who may not even remember there is a difference) but who want to learn more about them, reference them, do some leg work of their own on the numbers that are often thrown around.

If you are like me where there is a lot of rattle in the brain cage, you may often forget where to look for some stat or other. You know they are out there, people use the terms and some of the really established analytics people will quote them. But finding them is another issue entirely.

The underlying foundation of the analytics movement in hockey is, despite what some might believe, democratic in nature. The information is available to all and challenging the status quo and our assumptions is what birthed and fed the whole thing from the very beginning. So removing any barriers to that information and encouraging others to join the discussion is important.

I’m going to post a glossary of terms at the bottom of the article so that you can see what each term actually means (xGF, Corsi, HDSCF and so on). This may not include every term, but it covers the basics and you can probably figure out a few others that come up from those listed below.

To begin, we’ll start with brief descriptions on the sites themselves with some details about visualizations, searchable formats, player comparison tools and such that give each one a unique strength that recommends it.

Below is a chart that shows each site mentioned below and what statistics they include. Click to enlarge the image as there is a lot of information to try to get across.

Encyclopaedia Analytica

Let’s begin our statistical safari! We’ll proceed alphabetically.

Oh, The Sites You’ll See

Corsica

The catch-all of stats sites, hosted by Emmanuel Perry (@mannyelk on twitter) , Corsica gives you nearly every major statistical category (and more than a few minor ones). Closed for the summer for some coding upgrades, Corsica is, for all intents and purposes, your one-stop shop for team and player data.

Includes information on Teams, Goalies, Skaters, Combos (lines), Highlights, Games (data on specific games), Data, Blog posts, Similarity Calculator (player comparisons which you can search and contrast players, provides a % of similarity based on a number of metrics).

Stats include GP, TOI, Corsi, Fenwick, shots, sh%, sv%, xGF/A (expected goals for and against based on league average percentages against the possession numbers), low, medium and high-danger sv%, PDO, faceoff %, penalty differentials.

The site provides boxscore data as well as graphs based on game-data of score, expected goals, possession numbers, and allows for those to be adjusted by score or venue. Corsica likely carries the most wide-ranging and thorough collection of possession data available as well as offering a variety of different tools by which to arrange and compare that data.

DataRink

Provides searchable and sortable data sets for both teams and players by points, point percentage, games played, goals for and against, goal differential, team shooting %, sv %, Corsi raw and percentage and adjusted. All of the above can be filtered for PP and PK situations also.

There are both the raw numbers and visuals to help illustrate player and team performances in various categories. While not as comprehensive as Corsica, datarink helps fill in some of the gaps such as WOWY (With Or Without You – how a player does with and away from other players) and special teams data.

The skater information includes position, team, minutes, minutes per game, goals, assist, 1st assists and shooting %. You can add information such as on-ice goals or corsi events as well as goal differential, shooting % and save %. And again, all of the above can be filtered for PP and PK situations.

@datarink on twitter.

Hockey Reference

An encyclopaedic site that features a vast array of hockey-related stats from games, goals, assists, team standings, to the more in-depth statistical measures.

Excellent source for both player and team-focused information, searchable and sortable, the data is presented in table form with less focus on visualizations. Includes some particularly useful tools including Game Finder, allowing one to look up specific games for a player based on a wide range of variables including age, season, opponent, regular or post season, home or away, even the player’s performances on their birthday.

Includes Corsi and Fenwick metrics, PDO and their contributing measures on a player-by-player basis, zone start information including takeaways, giveaways and percentage of shots taken which get through to the net.

Hockey Reference is a possession-based database which also gives a good deal of time to the more traditional stats such as boxcars, blocks and hits. It also includes the percentage of shots a player takes which get through to the net, which while not unlike Fenwick, is a subtle but worthwhile distinction.

@hockey_ref on twitter

Hockeystats

Includes game data, shot-share, graphs, player in-game breakdowns, corsi data and zone starts. Excellent source for visual presentations of game-data which can be adjusted to Corsi, Fenwick (both also adjusted to score) shots, scoring chances and GRITCHART which adds blocks and hits to shots and shot attempts.

Very useful source for post-game analysis and a regular resource cited by Lowetide as he tracks player performance and development, though the data is game-focused rather than player/team focused meaning they are better served as a resource for mining data on a specific game rather than a season or career.

They also have some estimated puck possession data and such for AHL teams, which can be a difficult thing to find.

@hockeystatsca on twitter

Natural Stat Trick

Searchable and sortable information for both teams and players as well as individual games (regular and post-season), this site is ideal for player and team possession numbers with an eye to WOWY and game state (up a goal, down a goal, powerplay, penalty kill, etc).

Team, player and game data and a favourite of many who routinely delve into possession metrics. Lowetide uses this for his regular updates on player performances, both in-game and over the season.

Stats include, GP, TOI, Wins, Losses, OTL, ROW, Corsi and Fenwick metrics, Shots For, Goals For, Scoring Chances For and Against, Scoring Chances Goals For %, High Danger Scoring Chances For, Faceoffs

The site also produces game charts with shots, goals, and heat maps (shot location by frequency), player WOWYs in-game (to tell which lines are doing well and who might be helping the most), player vs opposition data in-game (to see which players are dominating play against which opponents).

@NatStatTrick on twitter

Puck on net

Good resource for checking team possession stats, especially if you want to investigate score effects.

Provides stackable rankings on team corsi (shots against, score adjusted and venue) and fenwick (same) data as well as more detailed information on Corsi and Fenwick by game state (up by 1,2,3 goals, even, down by 1,2,3 goals). Includes data on shots, goals, misses, hits and blocks as well.

Puckalytics

Searchable, sortable, skater and goaltender statistics as well as WOWY components, Puckalytics is a useful site the hosts possession data while allowing sortable rankings of player ranks relative to their teammates.

Includes skater stats, goalie stats, WOWYs (With or Without You – player comparisons and linemate impacts), comparisons. The old site was Hockeyanalysis.

The stats included are Games Played, Time on Ice, Goals For, Goals Against, Goals/60, Shots For, Shots Against, Shooting %, PDO, Corsi metrics For and Against, Corsi %

The original site (still preferred by some including Darcy McLeod, aka Woodguy, from BecauseOilers) is here: http://stats.hockeyanalysis.com/

@puckalytics on twitter

PuckIQ

The new kid on the block and the brainchild of Oilers bloggers/fans Darcy McLoed (@woodguy55) Ganesh Murdeshwar (@oilersnerdalert), Zsolt Munox (@oilfaninYYC) and Derek Blasutti (@dawgbone98). The site contains many of the fundamental metrics of the analytics field like Corsi, but it is the only site that features a newer possession model, Woodmoney (an amalgam of Darcy and Ganesh’s blog handles from their time visiting the Oilers blog, Lowetide).

Woodmoney is also referred to as Dangerous Fenwick, a possession metric based on Fenwick but which adds data on where the shot was taken on the ice, giving it an element of scoring chance – the closer the shot, the higher the likelihood of a goal.

The full value of that metric is realized when looking at the player tables, wherein player stats are displayed relative to the time they played against certain levels of opposition (ie: player x had 35% of their total ice time against middle tier competition and fared well, but in 30% of their ice time against elite competition they struggled badly). Dangerous Fenwick is, for the time being, available only on this site.

Searchable and sortable, PuckIQ provides Dangerous Fenwick data on players from the 2016-2017 season onwards. While not as comprehensive a resource as Corsica or Natural Stat Trick, Puck IQ covers the basics and offers a very workable resource for it’s own unique statistic.

Sporting Charts

Easily searchable site that is good for the visually-inclined. Wide array of data categories including goals, goals per game, goals against per game, goal differential per game, shots for and against per game, shooting and save %s, shutouts, powerplays per game, assist to goal ratio, hits per game, goal type breakdowns and so on.

This site is perhaps one of the better resources for those looking to pull visualizations or focus on specific statistical categories from a wide menu, friendly to beginners and useful to the experienced.

Includes player and team data not only for the NHL but also the NFL, NBA and MLB.

@sportingcharts on twitter

Vollman Player Usage Charts (Sledgehammer)

Old standard of visualizations that allows for team views and player history or comparisons placing players on a grid that measures zone starts against Corsi, either straight Corsi or Relative to opposition or teammates.

Rob Vollman is the author of the annually published book on hockey analytics, Hockey Abstract, a worthy read for anyone who wants to get more in-depth information on both established and currently researched analytical approaches. Vollman is also the author of Statshot, a more thorough guide to hockey analytics overall.

You can follow Rob @robvollmanNHL on twitter.

Xtrahockeystats

Hosted by Nick Abe (@nickabe on twitter), this site provides large blocks of sortable, searchable possession data including things such as points percentage for players, expected goals and delta goals (a calculation of the difference between a player’s percentage of points registered against what their expected points estimates).

Provides corsi and fenwick data, goals for %, rebounds, expected goals for and against, game stats including shot animations for specific games archived back to 2015.

Another insightful tool is the animated player shot map, which can be calibrated to show every shot taken by a designated player over a chosen length of time, colour coding them to show shots, goals and missed shots.

Analytics Blogs (sites which offer material more on the nature of analytics than their end product)

Crowdscout

A scouting site predominantly, but featuring blog posts by Cole Anderson (@crowdscoutsprts on twitter) that delve deeply into issues such as Wins Above Replacement (WAR), game theory and various other topics in the field of analytical work (that is, actually compiling and interpreting data). This is not easy reading for those who are either unfamiliar or uncomfortable with a deeper understanding of mathematical terms and statistical work, but it is intriguing and in all likelihood you will read here topics which will eventually find their way to the larger hockey stats conversations.

If you’re a fan of Oilers Nerd Alert’s blog, you will likely enjoy this site.

HockeyAnalytics

This is some deep(er) math for those interested in digging into new theory and exploring some old ones. Not a reference source, updates irregularly often with gaps extending beyond a year, but a good site for those looking to investigate new sports statistical measures and theories through articles and blog posts. Updates infrequently.

I’ve presented each site in alphabetical order, preferring not to prioritize any particular one ahead of another. Each has useful tools and ways of communicating the information that may endear you to one over another. There is a great deal of overlap in terms of the basic metrics they cover (Corsi, for instance) but they all provide their own perspective and there is often at least one measure, feature or category unique to each.

In Closing

This isn’t the first time someone has tried to compile all of this information into one place. Rob Vollman’s article on the subject, which is far more in-depth and exhaustive than this piece, is here. I recommend you bookmark it (in addition to this one, of course). If there is a stats site that you use that you don’t see listed here please drop me a line on twitter (@codexrex) and I’ll take a look at it.

If you find yourself with questions about any of these always feel free to ask any of the writers here at the Rig, you can find most of us on twitter and if we can’t answer it directly we’d be happy to send the question on to someone who can.

I hope you find this useful.

Glossary

GPs – Games played

TOI – Time on ice

TOI/Gm – Time on ice per game

GF – Goals for

GA – Goals against

G/60 – Goals per 60 minutes played

SF/60 – Shots for per 60 minutes played

SA/60 – Shots against per 60 minutes played

SCF – Scoring chances for

SCA – Scoring chances against

SCF/60 – Scoring chances for per 60 minutes played

SCA/60 – Scoring chances against per 60 minutes played

HDSC – High danger scoring chances

CF% – Corsi for represented as a percentage

FF% – Fenwick for represented as a percentage

CF – Corsi for represented as a straight number

FF – Fenwick for represented as a straight number

Sh% – Shooting percentage

Sv% – Save percentage

xGF – Expected goals for (derived from shots and shooting percentage)

xGA – Expected goals against

PDO – Combination of sv% and sh% which, over long term, tends towards 1000

FO% – Faceoff percentage

GpG – Goals per Game

PPG/Gm – Power play goals per game

PKG/Gm – Penalty kill goals per game

ScEff – Scoring efficiency

Assist/Gl – Assist to goal ratio

HDCF% – High Danger Corsi for percentage

HDGF% – High Danger Goals for percentage

HDSH% – High Danger shooting percentage

Ozs% – Offensive zone start percentage

Dzs% – Defensive zone start percentage

Thru% – percentage of shots taken that get through to the net

WOWY – Numbers that compare a player with and without his teammates, to determine if his results are his own or the result of linemates

Arrow to top