For awhile now I've been wondering how much of a disadvantage players are at when they start their shifts predominantly outside the offensive zone. In order to create a better sample and use as many players as possible, I determined that I would have to use Relative Corsi as my possession metric in order to account for different team dynamics.

For the purposes of this exploration I ignored quality of competition, mostly because I didn't want to create a convoluted calculation. I'm not sure what conclusions can be drawn from the following information but I found it extremely interesting.

Continue reading this post »

It is fascinating how much things can change in 20 games. After the first quarter of the season the Montreal Canadiens were primed to be a top 10 team in the NHL if they could get just a bit of luck on the powerplay and if Price could play like himself. Well they never did get that powerplay luck, and to make matters worse the team suffered a myriad of devastating injuries to key players.

This all resulted in a record of 18-20-4 over 42 games, firing an assistant coach early in the year, a trade to acquire a powerplay quarterback in Kaberle, firing of a head coach who had managed to keep the team near even despite only 4 games where the Canadiens iced a roster over the salary floor, and right after the 42 game mark; a panic trade giving up a sniper and good 1st line player for an older, less effective player in Rene Bourque and a second round pick in 2013. It has been a tumultuous time to be a Habs fan.

Never the less, Chris Boyle and I will be continuing our in depth breakdown of every Habs player to play significant time for the CH this season. In fact, we're going to up the ante this time and include five more statistics to add increased context into the evaluations of each player, while contrasting their play with the first quarter.

Last time there were some complaints and confusion about the grades given being too high for players who hadn't produced much (Gomez and Cammalleri in particular were mentioned), but you have to remember that results is not what we're analyzing here. We're looking at how the players are playing overall, everywhere on the ice. This is more important than goals and assists, and sometimes players just don't get rewarded for good play. Call it luck or random variance or unrepeatable performance, but over the amount of time we're covering, results aren't a good indicator of quality of play.

Before I explain the new statistics we'll be bringing in, I recommend you read my primer on possession and random variance based stats, as well as the explanations of the rest of the rest of the statistics we used previously in the first quarter review if you're new to advanced hockey analytics.

The five new statistics we will be using will give us both context for how difficult a player's job is, as well as measure a couple of intangible qualities.

The first two new statistics deal with quality of competition:

Traditional (or at least as traditional as microstats get) quality of competition stats are based on the plus/minus (+/-) scores of the opposing team's players. The calculation begins by creating a +/- score for each opposing player relevant to his own team. For example, if Player A on the other team has a +/- of -3, but his team has a +/- of -14 when Player A is off the ice, his relative +/- is +11 better than the team average. Once all the relative +/- scores are tabulated for the opposing team's players, the Qualcomp for any player on your team is the weighted average of all the relative +/- he faces during games, the weighting being based on what proportion of his total TOI is spent against each opposing player. This resulting metric is an indication of how the player on your team is being deployed by the coaching staff; i.e., players with high Qualcomp scores are generally playing against other teams' best players every night.

However even with relative plus/minus scores there are limitations to that specific stat. A player can have a poor relative plus/minus rating due to bad luck. So in order to add more context to matchups, Relative Corsi Quality of Competition was created. The statistic works in the same manner, calculating the the opponent's Corsi score relative to their own teammates, and averaging out all the opponents a player plays against. This way you can see how difficult a player's job is from a possession standpoint.

The next two statistics are very simple; zone starts are all about usage by the coaching staff, while zone finishes are all about driving the play. If a player starts in the offensive zone more than 50% of the time, they are either being used to create offense or being sheltered defensively. If a player finishes in the offensive zone more often than he starts there, chances are he is driving the play up the ice in a positive manner.

Lastly we'll be looking at penalty ratios to get a feel for how disciplined each player is, combined with how effective they are at drawing penalties. We'll do this by dividing the number of penalties a player takes by the number he draws per 60 minutes of ice time. This will tell us how many penalties that player will take per 60 minutes before drawing a penalty for his own team. The lower the number in this category, the better.

As with the last review, all statistics are at even strength only, aside from risk/reward which includes all situations.

Continue reading this post »

With half a season gone and some disappointing results for the Habs the time is ripe to take a look at who has met or exceeded expectations offensively and who has disappointed. Also it is important to look at scoring by game situation to look at which results are due to poor general play, poor 5 on 5 play or the specifically the result of the a bottom ranked power play. For our purposes here, scoring other than even strength and powerplay is treated as irrelevant since the main drivers of the traditional counting stats are these two game states.

Continue reading this post »

This piece is a collaboration between Chris Boyle and myself. What we aim to do with this piece is create a comprehensive template to evaluate the Montreal Canadiens, specifically each player, each quarter of the season. This first entry will be taking in data from the first 22 games. Only players who have played 10 or more games up to that point will be included. We will post a similar breakdown for each quarter of the season.

Before we get into it, let's break down the specific stats we'll be using to evaluate the players. The statistics I explained in this piece on November 14th will be used on an individual level instead of team level and while we will display PDO, we will also break it down into on ice shooting percentage and on ice save percentage. Shots, Fenwick, Corsi and PDO will be expressed in the same style as in that piece. As opposed to that piece, we'll be using those statistics for while the game is within one goal either way instead of just tied, or put another way, while the score is "close". Doing this expands our sample size by a large margin and allows for a more conclusive look at each player.

We will also be including scoring chance data from Olivier Bouchard's En Attendant Les Nordiques. For scoring chances we will be displaying scoring chances for and against the team at even strength while each player is on the ice. This will show how Fenwick, Corsi and shots correlate to scoring chances, and how effective the team is while each player is on the ice. Olivier counts scoring chances as shots that hit or miss the net within the "home plate" area:

A fairly new statistic (to fans but certainly not coaches) we'll be including here is Chris Boucher's risk/reward rating. What Chris does is watch every single game, and break down each and every play that occurs within the game. Boucher refers to these plays as events. Specifically he looks at the following plays;

successful or unsuccessful passes, successfully or unsuccessfully beating an opponent 1-on-1 (deke), successful or unsuccessful shot-attempts (puck on net), successful or unsuccessful dump-ins, successful or unsuccessful puck-battles, successful or unsuccessful loose-puck recoveries, successful or unsuccessful blocked passes, successful or unsuccessful blocked shots, successful or unsuccessful stick-checks.

What this allows Boucher to do is create a metric for how successful each player is at their specific job. For example, if a player has a risk/reward of 1.5 then that player makes 1.5 successful plays for every unsuccessful one. From reading Chris's site over the last several months I believe Jacques Martin uses a very similar metric when making personnel decisions. Usually ice time during games highly correlates to Boucher's risk reward ratings.

Finally, we'll be fiddling with the +/- statistic in order to get a figure that more accurately represents a player's performance. We'll be eliminating +/- data during empty net situations as well as short handed goals against. The reason for this is simple, neither of these situations tell us very much about player performance. Both situations exclusively punish the team's best players. Doing this allows us to focus solely on even strength play.

Why the extensive focus on even strength?

The main reason for this is that most of every game is played at even strength. Because of this, the teams that are most successful in this situation have the best chance at success over time. Look no further than last year's Stanley Cup champions, the hated Boston Bruins. They were incredible at even strength with a terrible powerplay and mediocre penalty kill. While special teams are important, they can be overcome with great even strength play.

Another reason is because due to how little time is played on special teams, 1/4 of the season is not a long enough time to get a solid sample size to critique the play in those areas. At least not conclusively.

Legend

Continue reading this post »

Early in the year its very easy to get caught up in the number of goals and points a player has in a very limited sample size of games, with the likes of Phil Kessel scoring like Gretzky and Eric Staal on pace for a 19 goal, 34 point campaign. However this does not mean that we cannot make an assessment of how well a player is performing and how productive they might be expected to be over the course of this season. It merely means that our projections should be based on the stable components of a production and the player's historic abilities.

Here I will attempt to get a good idea of what kind of goalscoring production we may expect from Montreal's players by basing our projection on the most stable component of a player's goalscoring given their current usage and competition, shots per game, while combining it with the element that best reflects his style and historical goalscoring ability, career shooting percentage. However, as the existing scoring results from this year are in the past and thus unchanging, the projections are only for the balance of the season with the player's current goal totals added afterwards.

Continue reading this post »

Our even strength preview started with the defense and continues with the forwards here.

Even strength play has ever been the bug-bear of the more recent editions of the Montreal Canadiens. Since the lockout, Montreal's performance at 5 on 5 has tended to range from mediocre to terrible and only somewhat redeemed by superior goaltending. The skaters in front of said goaltender has tended towards getting heavily outplayed. As a result of poor ES play Montreal had been out-shot every season from 05-06 to 09-10.

In that sense last season was a major departure from the previous experience. For the first time in a long time Montreal out-shot their opponents, but more importantly they dominated shots, and therefore puck possession at even strength. In shot differential at even strength with the score tied, considered by many as the most reliable puck possession stat and best indicator of sustainably good even strength play, Montreal ranked 4^th in the league, behind only Chicago and Vancouver. While some of this may be due to spray and pray tactics by Gomez's line, the fact remains that the majority of players for Montreal were in the black for Corsi, which suggests that they were at least a better than average even strength team last season, albeit one that had abnormally poor shot conversion (behind only New Jersey at even strength). Much of that improvement can be attributed to a better defensive group, especially the graduation of P.K. Subban who proved to be a puck possession monster on even strength.

Thus there is a great deal of optimism in certain quarters that Montreal is on the cusp of a breakthrough in 5 on 5 play, bolstered by the strong even strength abilities of newly incorporated players like Pacioretty and Subban, the resurgence of quality even strength players like Mike Cammalleri, Scott Gomez and Brian Gionta and the addition of Erik Cole, a player with a strong track record of ES point production.

One of the most exciting aspects of the upcoming season is the prospect of Montreal having the forward depth to construct a strong offensive third line. 6 bonafide top six forwards, emerging talents in Max Pacioretty, Lars Eller, and David Desharnais and a good offensive depth player in Mathieu Darche means Montreal will have the option to victimize the opponents weaker players.

These reviews include the point production rate stats per 60 minutes. A top six forward can be expected to produce at around 1.8 points per 60 while players that get above 2.1 tend to be either stars or good forwards having a strong or lucky year.

Continue reading this post »

Whether you're in a pool, or just like a good analysis of what to expect in the upcoming NHL season, "Hockey Prospectus 2011-12" might be worth picking up

This handy 489 page guide, with info and projections on all 30 teams, 1000 players and a top-100 prospect list, is from the creators of Baseball Prospectus.

The guide introduces and makes projections based on Vukota and GVT stats. I'm not one for relying on the stats to predict who finishes as there are so many other variables to take in, but it does go to show how fortunate last season's Stanley Cup winners really were.

For those into advanced statistics, such as quality starts, Corsi and advanced face-off stats, this would be the guide for you with plenty or articles and tables to dig into.

The paperback guide retails for $19.95 (USD) and can be purchased through Amazon via the link below.

The guide can also be purchased in PDF form from their website for just under $10

The Montreal Canadiens will be hosting  several former members of the franchise, during a four-game run, this week. The most notable comes this Saturday, when the Anaheim Ducks and former captain Saku Koivu pay a visit.

With the 2010-11 NHL season now into it's second half, I thought I would scout around the league to see how the "alumni" are doing in their current colours.

Hopefully I tracked all of the ex-Habs down, and the stats listed are based on this NHL season up to games ending on January 15.

As much as I toyed with it, I'm sticking to just current NHLers. Therefore, those who had brief NHL stints, such as Cedrick Desjardins and Steve Begin, etc. are excluded. It also keeps me from making any cracks about Sheldon Souray (still in the AHL) and Cristobal Huet (on loan to Europe). I've also excluded Dan Ellis, who was merely with the team on paper for 24 hours, as he has enough "Twitter problems" of his own to deal with.

Continue reading this post »