Ville Leino and Why Chance and Circumstance Matter

Up until, well...this past week or so, Ville Leino had been seen as a bust. An overaged free agent from the Finnish elite league signed by Detroit, Leino spent a season putting up just okay results in the AHL before making the leap for the injury plagued parent club this year. He didn't do much. All of his stats were mediocre or worse across the board (11 points in 55 games between DEt and PHI) . He looked for all the world like a replacement level NHLer and not much else.

Lo and behold, he's a better than PPG player in the post season! What happened? Well, for one, his PP ice time per game has increased from 29 seconds in the regular season to just under 3 minutes in the playoffs. He's also probably most coddled Flyer at even strength - in the first 9 games played, the former healthy scratch has started more than 70% of his shifts has a zone start ratio of 70+%(!). Compare that to Mike Richards at 49% and poor ol' Blair Betts at 25%.

Of course, the real culprit behind his explosion is the bounces. His PDO, before tonight's performance, was a shade above 106. I'm sure that'll go up after this evening. (PDO during the regular season = 96.4).

The truth of Ville Leino is probably somewhere in between the two extremes of his regular season and playoff results. Luckily for the Flyers, Leino is signed through next season so they won't be forced to overpay for his string of luck and unrealistically easy circumstances come July.

Normalizing plus/minus: Carl Gunnarsson

James Mirtle tweeted about the Leafs suddenly stand-out rookie d-man Carl Gunnarsson today. The positive press made me check his underlying numbers because when you have unusually good results for a young no-name player, it typically means the hockey gods are throwing the guy some love.

Turns out that's the case with Gunnarsson this year. He has the leafs highest PDO (on-ice SH% and SV%) at 102 according to behind the net. The timeonice app, which restricts inquiries slightly differently than behindthenet, has it even higher at 106 (Sv5 of .950! and SV of 11.5%). The Leafs team-wide average is (90.1 + 7.1%) 97.2, putting Gunnarsson some 9.3 percentage points above the teams PDO - that's huge and almost certainly all randomness (unless you think the Leafs have found the biggest difference maker in the game since...I don't know, uh, Bobby Orr?).

It's easy enough to normalize the percentages to see what Gunnarsson's ES "expected" plus/minus would be.

ES shots on net with Gunarsson on the ice = 148
ES shots against = 121
Current plus/minus (17 GF - 6 GA) = +9

Expected GF = 148* .071 = 10
Expected GA = 121* .901 = 12
Expected ES plus/minus = -2

Delta between actual and expected = +11

So the bounces have inflated Gunnarsson's plus minus by 11 points thus far. If you wanted to compare this rate across the leafs, then you could normalize each guy's PDO to the team average. Repeat and rank.

*(Keep in mind, that's just using the timeonice information - things would be slightly different if we employed Gabe's info, which either includes or expels stuff like 4on4/3on3 data. The principle and activity would be the same, however).

On Corsi

This post is directed at a certain audience, but perhaps it'll be useful for others.

Corsi is becoming a bit more understood and mainstream these days, but there remains a lot of folks who are confused by it. This will be my attempt to clarify the stat. What it means and how it's useful in analysis.

What it is

Let's start with a simple definition. "Corsi" is the difference between all shots directed at net for and against at even strength. That is (shots+blocked shots+goals+missed shots FOR - SH+BLK+G+MS AGAINST). The purpose of the stat is to determine possession. It is, in fact, a proxy for "zone time". A positive corsi rate = more offensive zone time. Negative = more defensive zone time.

Here's an analogy that might help. Let's say a hockey game is a tug of war. Corsi is the how far right or left of center the rope is. On an individual level, it's an expression of which players are really pulling the rope. Therefore, if your team has a positive corsi rate, it means they are spending more time in the offensive zone at even stregth. It means they are pulling the rope harder than the opposition.

Why it's useful

In general, two things determine goals for (GF) and goals against (GA) in hockey: volume and frequency. Volume is the amount of shots a team generates and allows. Frequency is how often a team scores or allows goals on those shots. What we're learning in the NHL is that the former is far more repeatable and indicative of skill than the latter. Let's put it another way...

Goals are relatively random events in the game. On any given night, 60-80 pucks may be directed at the net at both ends. Maybe 5-10 goals will be scored. As a result, goals are statistically less powerful because the sample size is small. This means that randomness has a far greater influence. And what we've discovered at the NHL level is that percentages (SH% and SV%) tend to regress to the mean over the long term. As a result, a team that is winning via high frequencies is said to be "riding the percentages" and their success is probably based on randomness or "luck".

Another example. We all know that the chances of a flipped coin landing on heads is 50%. However, it's entirely possible that a coin will land on heads 7 or 8 times in a ten flip sample. This is not indicative of a special coin or special "coin flipping skill". It's variance. As such, we can say with confidence that over, say, 1000 flips, we'll get back down to the 50-50 split.

Volume, or outshooting (corsi) is far more powerful statistically, however, and therefore less skewed by randomness. So, whereas percentages tend to regress to the mean, outshooting is far more stable and therefore indicative of a team's (or players) abilities. The evidence of corsi's value is being investigated by smarter men than me these days, but the evidence continues to pile up. Corsi correlates strongly with scoring chances. It also correlates highly with outscoring (0.65) over the course of the season. From the latter link, JLikens explains that outshooting explains 40% of the variance in EV scoring. Almost half. That's regardless of of things like goaltending ability or the percentage of shots a team has blocked versus what they get on net. It also excludes randomness as we discussed above.

Corsi is a long range stat. A team can outshoot the bad guys in a single game or even a series of games and still lose. The hockey gods can be arbitrary. But, eventually, outshooting teams will win more than they lose. And the more time they spend in the offensive zone, the better they are, the more they'll win.

Evaluating individual players with corsi is a little trickier, because circumstances can elevate or sink skaters, depending. The checking center or shut down defender who starts every shift in his own zone against superstars is bound to have a lousy rate, for example. But that's probably a discussion for another time.

I hope this helped clarify things.

Offensive Defensemen are grossly overvalued

One of my expressed goals for the summer was coming up with a way to determine the value of defensive defenseman.

It hasn't gone all that well.

A result of this endeavor, however, was the realization that NHL GM's pay entirely too much for offense from the blueline. By way of introduction, here's the leagues top paid blueliners next season, by cap hit:

Zdeno Chara
Nicklas Lidstrom
Brian Campbell
Scott Niedermayer
Jay Bouwmeester
Dan Boyle
Wade Redden
Dion Phaneuf
Ed Jovomonkey
Kimmo Timonen
Chris Pronger
Brian Rafalski
Andrei Markov
Bryan McCabe
Lubomir Visnovsky

Some of these guys play against top notch competition. Some don't. They all play on the PP. Some got their dollars because of a single season of offensive brilliance. Some are more consistent.

That's not really the point, though. Here's the ranked ESP rates of all NHL defenders last year. You'll notice that only one rearguard topped 1.5 ESP/60 (Mike Green) and 1 point per 60 minutes of ES time was the 86th percentile. Most of the leagues defenders scored less than a point per hour at 5on5 (even those expensive guys listed above). The mean rate was a mere 0.70 ESP/60.

By contrast, most forwards who garner less than a point per 60 at ES are scrubs - 4th liners, goons, guys on the downslope or the rawest rookies. Of the Flames regular skaters, only Eric Nystrom scored at the defenders mean rate (0.76).

If the rates thing is too ephemeral for some, let's go by the raw totals. last year, only 13 defenders scored 50 or more points. Three scored 20 or more goals. Not even 30 guys broke the 40 point plateau.

Now, it's only human to covet that which is rare. The unique is frquently in high demand. Thus why blueliners with relatively big points score massive contracts. But to my eye, paying a guy 6M to score 50 points a year is poor use of cap space in most instances - because forwards are far more efficient producers of offense.

96 forwards scored 50 or more points last season. One hundred and seventeen of them scored 20 or more goals. Sunny Mehta recently posted this article on the subject of positional shooting percentage - in his analysis, he found that forwards averaged 134 goals over 1385 shots or 9.6% in 08/09. Defensemen averaged just 18 goals over 442 shots for a 4.1% average.

By the virtue of the nature of their position, forwards are more dangerous offensively - they get closer to to the net and therefore manage more shots from better positions. Plus, many of them are more adept offensively, that being a primary requirement of being a forward.

What's more, Defenders are even more circumstantially reliant than forwards when it comes to generating offense. Blueliners need lots of ice time (since they are much more inefficient at scoring), they need capable forwards (since a bulk of a defenders points come in the form of assists) and they need favorable deployment by the coach (PP time, offensive zone draws, weaker match-ups). Take a gander at the ranked defender offense by efficiency table: you'll notice a lot of names in the 90th percentile or above are 3rd pairing guys who didn't actually score a lot of points (Goligoski, Giordano, Bogosian, Lebda, Hunwick, Montador). All sheltered. All limited ice-time. They scored at a good rate, but didn't have enough volume of minutes to make much of an impact. And even with the best circumstances - think lots of PP time, second tier competition, lots of offensive zone face-offs - your 6.5M PP quarterback might not even break the 50 point barrier.

The flip side, of course, is that defensive defenseman are relatively undervalued in the league. Let's look at that list of cap hits again. How many inside the top 30 are "pure defensive" types? I count three: Hannan (4.5M), Komisarek (4.5M) and Brewer (4.25M). Jan Hejda took on some of the toughest competition in the league for the Blue Jackets last year. He wouldn't even register on most fans radars: not only because he isn't the flashiest of players, but because his counting stats are necessarily suppressed by the circumstances in which he plays - own zone draws versus tough assignments. How would his results look like if someone else was getting fed to the wolves? Because the range of production at ES for rearguards is so limited (1.50 - 0), it probably makes sense to have a team full of $2M Jan Hejda's: you'd likely find a production efficiency curve based on the level of competition (and concurrent circumstances, ie face-off position) each Hejda clone was facing. Put another way - Cory Sarich faced the second toughest competition on the Flames amongst defensemen in 07/08 - he had a negative corsi rate and scored 0.25 ESP/60. Last year, Sarich spent most of the season on the third pairing. His level of competition was on par with Mark Giordano. His corsi jumped to +10 and his ESP rate more than tripled to .89.

So when GM's shell out dollars for points from the blueline, a lot of the time they're probably paying for a perception based on circumstances. They're also putting money into a vastly inferior offensive option in terms of efficiency relative to forwards. This isn't to say that some defensemen aren't better than others offensively speaking - that's self-evidently false. However, even the best defenders offensive contributions are muted. Suppressed by the demands of the position, by circumstances and by the fact thet they're beholden to the men in front of them; defenders need a lot of things to go right in order to be (seen as) offensively capable. Even then, they probably won't outscore your better than average 2nd liner.

As such, I think it makes sense to invest the big dollars (assuming offense is the goal) in forwards and then target the guys who can quietly handle the tough minutes in the defensive end. Like goaltending, it may make sense to abandon this strategy if one of the truly elite players is available (Nik Lidstrom), but, overall, I think the market overvalues offense from the back-end.

Forwards and shot prevention

Pursuant to my earlier post, here's the Flames forwards from last year (minus Jokinen) and their shot against rates at ES:



I've included each guy's QUAL COMP from behindthenet, just to add a layer of context.

Im still trying to decide if this number is telling me anything useful or not. That said, the results seem to be sensible, given what I know and observed about Calgary last year and in light of the quality of opposition each player faced. Boyd and Bertuzzi at the bad end of the spectrum makes sense, as does Moss and Glencross leading the way. Nystrom's number also indicates that the "defensive specialist" label applied to him as a prospect might actually be accurate.

Again, numbers via timeonice.com (shots) and NHL.com (ice time).

Defensemen and shot Prevention

Capturing a defenseman's contribution to defense is tough to do quantitatively, given what we know about individual players effects on SV% (it's relatively negligible), which means relying on goals against is out. Also, does a stalwart defender cause more opposition shots to get blocked? Go wide? Or, in fact, do good defensemen prevent shots from happening at all by anticipating plays, intercepting passes and rubbing out guys along the boards (etc)?

It's this latter point I chose to explore recently. Having a lot of shots blocked or go wide while you're on the ice can also be a sign of weakness - for example, the defender that gives the puck away at his own blueline would get credit under such an assumption if the shot during the resultant break-away happened to miss the net.

It strikes me, therefore, that a measure of total shots against/60 will give us an idea of the relative strengths or weaknesses of a given player versus his teammates and, perhaps, the rest of the league (the lower the better).

Here's what I found for Flames defenders last year. All numbers are ES, of course:

click to enlarge

Keep in mind these numbers are absent context (quality of competition, quality of linemates, starting position). That said, this initial look passes the sniff test I think. The ranking roughly accords with who were the best to worst defenders on the club last year (in my estimation at least). I think if we were to map something like this on top of contextual factors, the rate becomes more meaningful. For example, Reggie's number is middling on the team, but given the fact that he faced the toughest opponents and started in his own end the most, his number becomes more impressive.

I haven't carried out this exercise for any other team, so maybe this is all coincidental and Im not really measuring anything here.

What so you folks? Useful? Not useful? Should it be applied to forwards as well? I'd like to see how some of the other clubs turn out...

PS - ES shot totals found via timeonice.com.

A rare link day

- As the play-offs wind down, *other* NHL news is starting to pick up again. Most significant for Flames fans is the fact the Brent Sutter announced his departure from Devils today. I have discussions going here and here, while Duncan (who wins the "best headline" contest with his effort) and Matt add their thoughts elsewhere. Take a trip to In Lou we Trust for a NJD perspective.

- Heatley apparently wants out of Ottawa. Let the rampant speculation begin! I'll get the ball rolling - Jokinen + for Heater. Flames are weak on the left side, strong up the middle and Im not a fan of Jokinen anyways. Plus Heatley is a native Calgarian. It only makes sense (E3)!

- Some of the statzis around the blogosphere have made progress in adding further context to the numbers recently. Mudcrutch discusses the effect of face-off location and offensive/defensive zone ratios. This work will likely help me in determining relative value of defensive defensemen, which is going to be a summer project round here I think. JLikens has done some work with the "scoring chance" stat Dennis King put together this past year for the Oilers. Vic has some additional material on scoring chances, SH% and the importance of possession and outshooting here.

- The scoring chance thing intrigues me and I'd like to start doing it for the Flames next year. However, I'd also like to settle on a meaningful definition of "scoring chance" which is reliable across observers before proceeding. Perhaps something else to kick around in the dull days of August...

- Finally, I think the Cup is going to be awarded (either tonight or later this week) to the Red Wings for the second straight year. Love 'em; hate 'em; envy them; you have to respect that organization whatever your perspective. Everyone else is so far behind at this point, I think Ken Holland has actually lapped the field.

Ranked even strength production by defensemen

I did something like this for forwards awhile back and it proved to be fairly popular. So, here's a ranked (by ESP/60) list of defensemen from the regular season. I've marked Flames players in yellow. Also included are QUAL COMP ratings, corsi/60 rates, ES plus/minus and the PDO numbers of guys inside the 90th percentile (plus a few others I bothered to look up). Excluded are players that failed to play more than 30 contests this year.

**(Spreadsheet won't publish for whatever reason - here's a handy link to the numbers instead)**

Looking at this list made me wonder: is there any way to rationally and quantifiably capture the contributions of low-offense, hard minute defensemen? I mean, none of Regehr's stats are impressive at a glance, but well all know he's the best defender the team's got. The issue for me is - how good is he, really? Is there any way to know outside of "he plays the big guns on the other team consistently and doesn't seem to get killed"?

Other items of interest: Mark Giordano's standing in ESP/60 and corsi rate, the ridiculous PDO's of Boston defenders and Jason Smith apparently not scoring a single ES point the entire year.

Tracking Kevin Lalande

Lost amidst the other, far more immediately relevant deals this past deadline day was a fairly innocuous swap between the Calgary Flames and the Columbus Blue Jackets: Calgary moved Kevin Lalande for a 4th round draft pick. At the time, I was surprised the Flames managed to get anything as high as a 4th rounder for what was a at the time an ECHL goalie who had been previously drafted in the 5th round. Seemed like a steal, even when the return was so modest.

The Blue Jackets installed Lalande as their AHL starter (since their previous starter and back-up were up in the NHL at the time) and he promptly won 3 straight games and was named AHL player of the week. Since then, I've kept a closer eye on the former Flames property and am starting to come to the shakey conclusion that Calgary may have traded away their best goaltending prospect.

First, some background. Lalande spent 3 seasons in the OHL with the Bellville Bulls. His SV% over those three years was consistently good to excellent: .920, .916 and .919. He finished 4th, 3rd, and 3rd in the OHL in terms of SV% in each of those years respectively. In 04/05, his rookie season, his .920 SV% was the best of any freshman puck stopper in the league. He was named to the all-star team in 05/06.

His first few games as a pro didn't go so well for Kevin and may be the reason he was relegated to a third stringer by the organization. In 7 games for Quad Cities, he won just twice and stopped just 88.8% of the pucks he saw. That was last year, most of which he spent in the ECHL with the Las Vegas Wranglers, where he put up far more respectable numbers over a longer period (.932 SV% in 27 games played). Since then, Lalande has managed sparkling percentages where ever he's played: ECHL (.925 SV%), Quad Cities (.929 SV%) and Syracuse (.927 SV%).

What's significant is, as far as I can tell, neither the QC Flames nor SYR Crunch were good teams this season. The former placed 5th in the West Division and had a GD of -4. The latter were 5th in the North Division and actually worse defensively, allowing 10 more goals against than the Flames and had a nominally worse GD (-12). So it's not like he was playing behind juggernauts that were limiting the opposition to perimeter shots and winning each game by 4.

In addition, Lalande outperformed all his peers this season. No other Flames 'tender posted an equal or better SV% than Lalande. Ditto Syracuse, aside from Calder shoe-in Steve Mason who played a mere 3 games in the minors before being called up to save the Blue Jackets season.

Now, to some degree, I'm reading tea leaves here: Lalande hasn't even seen 1000 shots at the AHL at this point in his career, so we're talking small sample, potential for misleading results, yadda yadda yadda. Still, looking at Lalande's results both past and present shows a lot of arrows pointing in the right direction. And while Im not big on the value of goalie prospects in the first place, I find it's generally a good rule to try to keep the best ones around. Of course, maybe he's just a really, really good Junior/ECHL/AHL goalie and little else. Something to keep an eye on, anyways.

Clarification (sorta)

In looking at the Flames seasonal percentages, I noticed that Vic's timeonice app was delivering a different ES SV% than the one stated in Desjardins study. In fact, the underlying shot and goal totals were different for each. Vic's site had Kipper seeing about 170 more shots and allowing three more goals than what Desjardins accounted for. As a result, Kipper comes out with a better SV% (90.9) in the former.

Unable to resolve the differences myself, I emailed Desjardins for some clarification. His response was:

Some possible differences:

- Vic includes 4-on-4 and 3-on-3 (this is certainly the biggest difference)
- Vic uses just the NHL.com data; in order for me to get shot
location, I need to combine ESPN with nhl.com. I've audited it in the
past and it looks like they have the same dataset, but there are some
mismatches
- Vic determines the number of skaters directly from the NHL.com
play-by-play (PL) whereas I pull it from the TOI charts (TV and TH).
For whatever reason, they're different.
- Neither one of us includes empty nets

So, seems to be differences in category and reporting source. Technically, then, Kipper's SV% at behind the net (89.9) is his "true" 5on5 save rate (as close as we can get with the data in question, I guess), while Vic's site is Kipper's "true" even-strength SV% across all such variants of that condition.

As of right now, I'm unsure which should be consider more reflective of a goalie's ability.

Kipper expected versus actual

Mirtle alerted me to the fact that Desjardins has published his yearly look at actual performance versus expected for goalies (based on these criteria).

James has the league wide run down, but of course it's Miikka we're mostly concerned with around here.

And regular readers of the space can likely predict the outcome of Gabe's analysis. Keep in mind this is ES only (and Kipper actually excelled SH this year, relative to other game states):

Kipper:

Actual SV% - 89.9
Expected SV% - 90.8
Diff - (-0.9%)

Actual GAA - 2.68
Expected GAA - 2.44
Diff - (+0.24)

Of all the NHL goalies that played this season, only Raycroft, Budaj, Legace and Lalime under-performed their expected GAA to a greater degree than Kipper according to this process. Toskala and Ersberg tied with Kiprusoff at +.24. The delta between his expected SV% and actual also placed him at the bottom of the list - tied with Marty Turco and ahead of only the Colorado tenders, Ersberg, Toskala, Legace and Lalime.

In theory, this analysis factors out team effects. Kipper's "expected stats" are based on the quality of shots he faced behind the Flames defense. His expected SV% of 90.9 placed him in the middle of the pack (tied with the likes of Chris Mason, Dwayne Roloson and Craig Anderson), so it's certainly arguable that the Flames didn't get good bang for the buck out of their pricey blueline this season. Still, it's not like he managed to outperform the club's mediocrity. If anything, he sank them to lower depths.

I hypothesized last summer that we'd see something like this out of Kipper this season based on his steadily declining stats, although looking at stuff like this now I'm still slightly surprised at how far he's slipped. The Flames probably didn't even really get replacement level performance out of him this season - and he's signed for another five more years. The question now becomes: is this as low as he goes?

A ray of light

A week or two ago, I was compelled to comment over at Sleek's place because the Good Earl was frustrated by his chosen team's money players: Getzlaf, Perry et al. Were being handily outscored by the opposition and the Ducks were in a seeming death spiral:

One specific area that could use some improvement? In the seven games since the trade deadline, Ryan Getzlaf, Corey Perry, and Eric Cristensen have been surrendering even-strength goals at twice the rate that Todd Marchant, Petteri Nokelainen, and Mike Brown have been surrendering goals while shorthanded. The Ducks seemingly would find some easier success if they could stem some of the GA-bleeding from the top line.

Upon reading his (understandably justified) rant, I took a look at ANA's corsi and percentages over the given period of time and discovered the big boys were outshooting the bad guys, but had lousy SH% and SV%. From my comment at the time:

The SV% with Getzlaf on the ice the last 7 games has been .822! For Perry….833. And the SH% for the Ducks with those guys on the ice has been 7 and 8.3% respectively. Those are TERRIBLE rates and it makes me think your big guns have been getting the bad bounces recently.

The day after I made that comment, the Ducks hammered the Avs 7-2 and they've been hot ever since. What happened? The percentages swung back around. Big time:

Getzlaf: .926 SV%, 20.4 SH%
Perry: .933 SV%, 21.6 SH%
Ryan: .915 SV%, 16.3 SH%

Now, I'm not writing this to pat myself on the back or say "I told you so". I bring it up because it seems relevant to the current situation the Flames find themselves in. As you all know, the club has been fighting through it's own bout of offensive doldrums, led in terms of ineffectiveness by the Flames own money players, who have themselves been handily outscored the last few weeks or so. Here's the club's corsi rates and percentages since March 18:

Jarome Iginla: +24, .886 SV%, 5.6 SH%
Olli Jokinen: +17, .884 SV%, 4.3 SH%
Mike Cammalleri: +26, .894 SV%, 4.1 SH%

Now that's bad - probably unsustainably bad. Overall during the period in question, Calgary is +92 in terms corsi, but has a putrid ES PDO (90.3+5.7=96). Kipper isn't making many saves and the opposition has enjoyed a .943% ES SV% in aggregate versus the Flames since the 18th of last month. Add in the sputtering power play and it's stunning that the Flames have won any games at all.

Basically, Calgary has had some shit luck recently. The injuries (and addition of Bertuzzi back into the top 6 ) haven't helped, but one has to assume that the correction is coming sooner or later and pucks are going to start finding the back of the net again. "Sooner" would obviously be preferable, especially with half of the defense corps on the sidelines.

New Numbers - Seminal work

Due to some excellent work by the more math-inclined bloggers out there, new metrics such as CORSI and on-ice SV/SH% have become very important in the way I think about and analyze the game. Unfortunately, the information about these various new measures and the assumptions behind them hasn't been collated in a single place yet (as far as I know), so, unless you've been able to follow the various discussions across a number of different blogs over the last 12 months, you might not have the whole picture.

As such, I've decided to track down the most influential posts as a reference for those with questions, complaints, comments, etc.

- Zone time, Corsi and Correlation to Winning by Vic Ferrari.

- Winning and Losing you games by Vic Ferrari (individual players contributions to corsi and SV%).

- On-ice Save Percentages By Mudcrutch.

- PDO Numbers By Mudcrutch.

- Outshooting By Mudcrutch.

- Even Strength Shooting Percentage By JLikens.

There's lots of reading there and some fairly in depth math, but I think this body of work provides the basis for my assumptions about variation, randomness, percentages and the value of possession and outshooting in the NHL.

If I've missed any other seminal posts or articles, please let me know and I'll add it to the list.

Ranked even strength production by forwards

I put together this list of NHL forwards, ranked by this season's ESP/60 rate. I removed anyone who hadn't played at least 30 games and added eah players associated corsi rate/60, just for a bit of context. I thought it was kind of interesting, so I figured I'd share it. Keep in mind, this is just for the current season.



The full list can also be found here. All numbers via Behindthenet.ca.

PS - I "marked" a few of the notable Flames players in red.

Why I look at the numbers

Michael Lewis, author of the seminal book "Moneyball", penned a piece on under-appreciated basketball player Shane Battier for the NY Times recently. It's a long, involved 8 pager that discusses the "new stats" movement in major sports in general, and provides a few quotables on why I - and some of the more numerically inclined bloggers upon whom I rely on so much - employ seemingly foreign and arcane figures and percentages for discussion and analysis of the sport. A taste:

In 2005, the Houston Rockets’ owner, Leslie Alexander, decided to hire new management for his losing team and went looking specifically for someone willing to rethink the game. “We now have all this data,” Alexander told me. “And we have computers that can analyze that data. And I wanted to use that data in a progressive way. When I hired Daryl, it was because I wanted somebody that was doing more than just looking at players in the normal way. I mean, I’m not even sure we’re playing the game the right way.”

The virus that infected professional baseball in the 1990s, the use of statistics to find new and better ways to value players and strategies, has found its way into every major sport. Not just basketball and football, but also soccer and cricket and rugby and, for all I know, snooker and darts — each one now supports a subculture of smart people who view it not just as a game to be played but as a problem to be solved. Outcomes that seem, after the fact, all but inevitable — of course LeBron James hit that buzzer beater, of course the Pittsburgh Steelers won the Super Bowl — are instead treated as a set of probabilities, even after the fact. The games are games of odds. Like professional card counters, the modern thinkers want to play the odds as efficiently as they can; but of course to play the odds efficiently they must first know the odds. Hence the new statistics, and the quest to acquire new data, and the intense interest in measuring the impact of every little thing a player does on his team’s chances of winning. In its spirit of inquiry, this subculture inside professional basketball is no different from the subculture inside baseball or football or darts.

The article is interesting, also, due to basketball's similarity to hockey, relative to say, baseball. The latter is more static than the other two, and the dynamic nature of hockey and basketball has created challenges in teasing apart individual effects on winning (and has often been used as a bludgeon by anti-new-stats folks to denounce efforts to understand the game quantitatively):

There are other things Morey has noticed too, but declines to discuss as there is right now in pro basketball real value to new information, and the Rockets feel they have some. What he will say, however, is that the big challenge on any basketball court is to measure the right things. The five players on any basketball team are far more than the sum of their parts; the Rockets devote a lot of energy to untangling subtle interactions among the team’s elements. To get at this they need something that basketball hasn’t historically supplied: meaningful statistics. For most of its history basketball has measured not so much what is important as what is easy to measure — points, rebounds, assists, steals, blocked shots — and these measurements have warped perceptions of the game. (“Someone created the box score,” Morey says, “and he should be shot.”) How many points a player scores, for example, is no true indication of how much he has helped his team. Another example: if you want to know a player’s value as a ­rebounder, you need to know not whether he got a rebound but the likelihood of the team getting the rebound when a missed shot enters that player’s zone.

There's no question that hockey analysis is just taking it's first few stumbling steps in this direction. We're only now starting to learn which stats are truly indicative (and predictive) of winning and which are merely effects of other processes (such as chance or randomness). When I started this thing a few years ago, I was introduced to the parsing of stats into situation (ES, PP, PK) and the tracking of offensive efficiency (ie: rates) rather than summing totals. Recently, the focus has become percentages, the role of chance and variance over spans of time, and how those can potentially result in the over or undervaluation of a player.

We aren't there yet and some mistakes are going to be made along the way, Im sure...but my education in the quantitative methods of understanding hockey has actually informed and improved the way I watch the games now, I think. Perhaps what might be most interesting in the long run, however, is how methods of analysis will converge across different (yet similar) sports:

One well-known statistic the Rockets’ front office pays attention to is plus-minus, which simply measures what happens to the score when any given player is on the court. In its crude form, plus-minus is hardly perfect: a player who finds himself on the same team with the world’s four best basketball players, and who plays only when they do, will have a plus-minus that looks pretty good, even if it says little about his play. Morey says that he and his staff can adjust for these potential distortions — though he is coy about how they do it — and render plus-minus a useful measure of a player’s effect on a basketball game. A good player might be a plus 3 — that is, his team averages 3 points more per game than its opponent when he is on the floor. In his best season, the superstar point guard Steve Nash was a plus 14.5. At the time of the Lakers game, Battier was a plus 10, which put him in the company of Dwight Howard and Kevin Garnett, both perennial All-Stars. For his career he’s a plus 6. “Plus 6 is enormous,” Morey says. “It’s the difference between 41 wins and 60 wins.” He names a few other players who were a plus 6 last season: Vince Carter, Carmelo Anthony, Tracy McGrady.

Corsi, anyone?

Diagnosing Iginla's ills

I don't think there's any denying now that Jarome Iginla in struggling. He has been for awhile, except the issues were rendered seemingly irrelevant by the teams success.

Well, now that the depth guys have cooled off a bit, the fact that Jarome has become Jarmoe is definitely influencing the Flames record to the bad. Before proceeding, perhaps I should firmly establish that Iginla is indeed having an off-year, since the guy is still inside the top 10 scorers in the league.

First, the recent struggles. Iginla has all of 2 goals in 2009. That's 19 games. He is currently on pace for 30 goals, his lowest total since 1999 (!). He has two points in his last five games and is a -5 five over that span.

What's more significant is the fact that Iginla isn't a difference-maker at ES this season. Here are his pertinent underlying 5-on-5 stats from last year and this one, just for the sake of comparison:

07/08: 2.85 ESP/60, 3.68 GFON/60, 2.21 GAON/60, +32 (!)
08/09: 2.16 ESP/60, 3.17 GFON/60, 2.88 GA/60, +4

Iginla is less efficient at producing points this season and more apt to get scored on. In 07/08, he scored the most ES points in the Western conference (65) and was third overall in the NHL by that measure ahead of guys like Joe Thornton and Pavel Datsyuk. This year, he's tied for 17th in the league with Travis Zajac and Rene Bourque. The difference between his ES ice time from the two seasons is nominal as well, with him averaging just 30 seconds more/game last year over this one.

What's more, Iginla is actually less effective on the PP this year as well. His previous PPP/60 rate was a team best 5.37. This year? 4.74. In addition, he doesn't appear to be the engine driving success with the man advantage anymore either: Iginla's GFON-GFOFF differential was +5.31 in 07/08, which has shrank to just +1.91 this season. The be fair, that has a lot to do with the team's improvement (the 2nd unit actually scores sometimes this year), but it still highlights the fact that his GFON rate has fallen by nearly 1 full point per hour (7.69 versus 6.77).

It's not just that the performance is there but the results are lacking. Jarome has looked ineffective by my eye for a large portion of year. Aside from the occasional bitch-slapping of a weak sister, Iginla's game has been a lot of frustration and perimeter play. He can't seem to bear down on chances like he has previously and his shots are all too frequently wide, blocked or easily stopped (accounting for his 0.57/game drop in shots on net). Iginla's board work and puck possession has really fallen by the way-side by my eye as well: separating Iginla from the biscuit used to be a herculean task, reserved for only the best checkers and defenseman on the other team. Now, Iginla loses puck battles and races more often than not. For my money, Rene Bourque, Curtis Glencross and David Moss have been superior to Jarome along the boards and down low in the offensive zone in terms of winning possession.

So what are the choices here in terms of explanations? "Lost-a-step", injury or circumstances...

The first possibility I dismiss immediately. The difference between 30 and 31 is marginal and he is the best conditioned athlete on the club by all accounts. And it's not like Iginla is sitting fat-and-sassy at home, enjoying his $7M pay-cheque and yawning at reports of his ineffectiveness - the guy isn't Alexi Yashin. He's roundly considered one of the most competitive players and best leaders by his peers.

The second option is possible I suppose, although one wonders what kind of ailment effects players in such a manner that a.) it sticks around all year, b.) they miss 0 games and play 20+ minutes per night anyways and c.) it only hampers performance against good teams.

Which leaves us with circumstances. What's different between this season and last? Linemates and competition.

Let's see who Iginla played with during 07/08 (via this site):



Now, from this season:



Let's look at the relative contributions of each unit at ES as well.

O7/08:



08/09:



By "frequency" I assume the site means "minutes", so let's standardize each unit's production so we can compare apples to apples here. I'll just apply the typical /60 correction we're all used to. Also, it looks like the production is just a "goal/line" type stat, meaning if a trio manages a goal, the site scores that as "1", rather than collecting all the points of the players in question for each event (goals, assists).

07/08:

Iginla/Langkow/Huselius - 0.65 goals/60
Iginla/Conroy/Tanguay - 0.92 goals/60
Iginla/Langkow/Tanguay - 0.99 goals/60

08/09:

Iginla/Conroy/Cammalleri - 0.58 goals/60
Iginla/Lombardi/Cammalleri - 0.57 goals/60
Iginla/Langkow/Bertuzzi - 0.34 goals/60

(Other combos are sullied by low sample sizes.)

As you can see, the lowest performing unit featuring Jarome in 07/08 outperforms the best one so far this year. And it's not like the club itself has taken a monumental step back in terms of scoring overall: with 108 ES goals, the Flames currently sit 6th in the league. Last year, they were 12th with 145 in 82 games.

Causes -

Difference in competition:

Iginla played behind Conroy and Tanguay a lot last year in terms of taking on the big guns. This season, that responsibility has mostly fallen on Jarome, although Keenan has been mixing things up a lot recently without much change in the outcome. Also, Iginla tended to face the big boys in 07/08 whenever Tanguay was on his line - and it just so happens that the combinations featuring those two players above are the best ES scoring ones in the analysis. So...

Linemates:

All arrows point to a reduction in quality of linemates playing a significant part in the variance in Iginla's output from one year to the next. Lots of kudos have been afforded to Cammalleri so far, except he's done a lot of his damage on the PP. Also, some muted acclaim has been heard for Bertuzzi in the media and such, but I think he just plain sucks. Neither of these guys are Tanguay - hell, I'd argue neither of them are Huselius at ES - especially not in terms of puck distribution in the offensive end of the rink.

Implications:

- Keenan probably owes both Tanguay and Huselius apologies, as does a portion of the Flames fanship. Both guys weren't crash-and-bangers and both could go through periods of frustrating ineffectiveness, but both were also excellent compliments to the captain (and probably better at 5on5 play than a lot of people - the taciturn coach included - gave them credit for).

- If Sutter wants to get Iginla going and seriously challenge the better teams in the league this post-season, he'll need to get Iginla some support in the top-6. Preferrably someone with good vision and an ability to move the puck into soft areas of the ice. This would probably also help the currently floundering power play as well. If not, Keenan will have to find a way to shelter Iginla a lot more.

- The answer to Iginla's struggles is certainly NOT playing him with Bertuzzi and Langkow. That's been the most ineffective trioin terms of output.

Im open for debate, but Im not sure how any other conclusion is possible at this time. Hell, we all some this coming, but it's fairly jarring to witness. If accurate, this look should also influence Sutter's decision making this summer as he wades through the league's UFA pool.

Also - thank you, lord, for Moss, Bourque, Glencross and the steps forward by Nystrom, Boyd and Lombardi.

Kryptonite

Jarome Iginla has played 4 games against the Hawks this year, managed zero goals, three assists...and a ghastly -9 rating. That's arguably the worst stats line for the captain versus any team in the league. Those terrible looking figures got me thinking back to something I noticed earlier in the year: Iginla seemed to be beating up on the lesser clubs (Nashville, Phoenix) but disappearing against the better opposition (Detroit, etc.). Luckily, CBS has those situational stats, which I've parsed into "strong" and "weak" to give us an indication if the perceived split in his performance is real or imagined:

click to enlarge

As you can see, the difference is real and startling: In 21 games against the relatively stronger teams, Iginla has just one goal, 18 points and is a putrid -17. His PPG pace is 0.86, he averages 2.95 shots on net and his SH% is .02!

In stark contrast, Jarome has made hay against the weaker-thans all year: 19 goals, 40 points, +16 and a PPG rate of 1.33, 3.93 shots/game and a SH% of 0.16.

One can quibble with my assignment of certain teams to certain conditions here, but I think I've caught the meat of the issue. Naturally, we can realistically expect a drop of just about anyone's stats in this kind of split, but...there's no way it should be such a steep drop-off, especially for a player of Iginla's caliber and pay-scale. I mean, ostensibly the reason to retain and pay a player like Jarome a huge chunk of dough is so that the team can compete with the other good teams in the league. Unfortunately, the captain has failed utterly in that assignment this year - so much so that I'm struggling to get my head around it still.

Causes? Fears? Does this make anyone wish for a better center at the deadline? I dont know what's going on here, but, whatever the cause, Iginla has to be better going forward. Calgary can't hope to do much in the play-offs if #12 continues to be little more than a detriment against the better teams in the league; there isn't any weak sisters in the post season, afterall.

Riding Percentages

JLikens has a post on even-strength shooting percentages which is of interest. It illuminates why CORSI rates can be more indicative of a strong team than GD.

The same rules regarding the x and y axes that apply to the yellow (actual) distribution also apply to the pink (random) distribution. The relative frequency for the pink distribution is the proportional representation of each artificial shooting percentage category. As an example, as there were 100 "seasons" and 30 teams, the entire sample consisted of 3000 artificial shooting percentages. 601 artificial percentages fell between 0.08 and 0.085. The relative frequency for the '0.0825' category is therefore ~0.2, as 601/3000 = ~0.2.

However, the pink distribution is itself fairly broad. In fact, it very closely resembles the yellow distribution. As would be anticipated, the yellow distribution is slightly broader than than its counterpart, but the difference is not large. This suggests that much of the inter-team variation in EV shooting percentage is the result of randomness.

...

The fact that the actual distribution is somewhat broader than the expected distribution shows that teams do indeed differ in their underlying shooting percentage at EV. Nonetheless, this variation is only very slightly larger than what would be predicted by chance alone. The underlying differences appear to be minimal.

The point is re-iterated in the comments of the post by another analysis found here. The conclusion?

"So after the whole season last year, if you gave an average team an average season and let them ride ordinary variance they could at the end of the year rank anywhere in the top 25 in goals per game."

That's a fairly massive spread, folks. At least to the human brain trying to interpret standings results during the season. When it comes to evaluating teams and predicting future success over the long term, it's becoming clearer that good teams dominate in terms of outshooting and offensive possession and pretenders ride percentages. This is why I was encouraged by what I found when I looked at the Flames road CORSI rates and percentages yesterday - the former was good and the latter was bad. So while Sutter is busy glowering at his charges after another oh-fer on a back-to-back, the truth of the matter is shit happens and the Flames will probably improve away from the Dome just because their modest road record is largely illusory.

Which isn't to say I think the team played well the two nights - on the contrary. But over the whole of the season, Calgary has positive offensive-zone results when sleeping in hotels. If they keep that up, they should be fine.

Calgary Flames - road woes?

Now cross-posted at FlamesNation...

After last night's loss at the hands of the bottom-feeding Colorado Avalanche, the Flames road record dipped to a mediocre 11-10-1; a stark contrast to their 19-5-3 home dominance.

With 18 of their final 33 games on the road - including a noxious 7 game trip at the beginning of March, it would be interesting to see what may be causing the inefficiencies away from the Dome.

First, let's get the obvious stuff out of the way: the Flames started the year on the road and they were terrible. They've also struggled in back-to-backs so far this season, often due to cruel scheduling.

Still, the issue has persisted somewhat beyond the ugly first couple of weeks. Calgary were monsters from January 1 till now, winning 8 of 12. Three of the 4 losses occurred on the road, however. Not a good trend for a team aspiring to be amongst the elite in the league.

Let's take a look at the clubs away corsi and ES percentages during the last month or so to perhaps illuminate what's been going on.

The good stuff:

Solid goaltending (.915) and strong outshooting (+50) over the 5 games in question. Doesn't look like the Flames were grossly outplayed in aggregate, which is good news: it indicates the team isn't reliant on certain home advantages to drive results (last change, familiar surroundings, etc).

The bad stuff:

Flames combined SH% and SV% over those 5 games was a below average 98.6, despite technically outplaying the opposition at ES. The culprit is the nominal 7.1 SH%, which is obviously driven by the terrible SH% with guys like Iginla (1.9 !!), Cammalleri (2.2), and Lombardi (4.3) on the ice. When your first line skaters are getting those kinds of bounces (despite typically outplaying the competition) it's tough to get wins.

The bright side, is, of course, the fact that there's little chance of Iginla et al. shooting blanks forever. Jarome managed 19 ES shots on net over five roads games during the period in question - and scored just one goal. Bertuzzi, for comparative purposes, scored twice on 6 shots. Iginla is still clearly moving the puck in the right direction and at some point it's bound to start going in the net.

Other stuff:

It's amazing how results can color perceptions, isn't it? I was starting to come around on the subject of Bertuzzi - but guess who was the biggest boat anchor in terms of possession at ES away from the Dome last month? That's right - ol' number 7 (one of only 4 players in red by the CORSI metric). His ineffectiveness was entirely hidden by astronomically good percentages at both ends of the rink. His combined SH%+SV% number of 119% is mind-bogglingly good fortune, given that we know that forwards don't really effect their own goalies SV% and the Flames were out-played with him on the ice.

Bert's CORSI rate has been mediocre or worse all year (despite good linemates and so-so opposition), but he was a victim of crappy ES % for a good portion earlier in the season. This is likely the correction. Once the hot streak ends, Bertuzzi may go back to looking like a questionable signing. You'll also note from that link that star players like Iginla tend to drive both shots and SH% over the long run, further suggesting that the team can't possibly be this bad with Jarome on the ice for much longer. Once Iginla gets the bounces, the team is likely to do better in other buildings.

Of course, five games is a pretty small sample size, so it might be instructive to look at the entire season on the road thus far.

The picture painted by these numbers might actually be a rosier one, if you believe that outshooting is valuable and percentages tend to correct themselves. The Flames are +128 away from the dome this year, with only youngsters like Boyd and Pardy in the red. Even Bertuzzi is above water by a good measure. The weakness, again, has been the SH% and SV%, with the team not being particularly efficient at either end of the rink. Wayne Primeau was the only semi-regular skater to enjoy a SH% above 10* when he was on the ice. The 89.8% SV is likely well below league average as well.

*(Poor Curtis MceLhinney - the Flames's ES SH% with him in net is 1.2%!!!)

All told, Im less worried about Calgary's road heavy schedule going forward: they're currently one game above .500 as the visitor thus far, despite some pretty crappy bounces and a number of ugly back-to-back trips. They've managed to play in the bad guy's end of the rink more often than not, which indicates they aren't necessarily being outplayed. If the team can continue in the same and vein and get Jarome on his horse, they'll be fine.

More acronyms and numbers

Interesting stuff over at MC79 recently (H/T to commenter Shep for the heads up). The discussion revolves around the "PDO number", which is a teams (or players) summed SH% and SV%. From the linked post:

A PDO number is a team’s ES S% plus their ES SV%. The idea, and I know that this will make some people uptight, is that there’s a lack of sustainability in that number - if it’s high, it’s going to come down; if it’s low, it’s going to go back up.

"High" being above the mean (approx. 100) by a few ticks and "low" being below the mean by a few ticks. Over the long haul, things tend to correct themselves in the percentages area, as Tyler shows in his first table: big differences (+/- 3 either way) above or below 100 tend to regress or progress back towards the middle over time.

The big "over-achievers" thus far? VAN, BOS, NJ and maybe PIT And MTL. The big under-achievers? CGY, COL, BUF, LA, DET (!!), NYI AND TOR.

Now, individuals obviously have some effect on SV% and SH% to a degree. Luongo, Regehr and Lidstrom will likely have a positive influence on a club's SV%. On the offensive side, Jarome Iginla and Sidney Crosby have the power to improve a club's SH% somewhat. Naturally, bad players will do the opposite.

Still...everything tends to gravitate to the mean in the end:

Overall, the effect is really startling...I’ve summarized the differences between the best 20 teams and the worst 20 teams in terms of the percentages for the first quarter and then the final three quarters. While there’s still some differences in the final three quarters, they’re much, much smaller.

This is good news for Flames fans. Calgary has one of the lowest PDO numbers in the league at 97.9, owing in part to Kippers limp start no doubt. Over the course of the season, we can probably expect the Flames to creep closer to the 100 mark - which, especially if they can keep up the SF/SA ration of 1.09, should result in some wins.

I find a look at individual PDO numbers helps to illustrate the sustainability (or lack) of certain numbers. Here are the Flames results so far:



Only three players hover above 100: Robyn Regehr (100.7), Andre Roy (101.2) and Curtis Glencross (103.3). Of the three, only Glencross' number seems unsustainably high, although one would think that Roy would be the type of guy to post lackluster percentages both ways (and so, is in line for a fall himself). Also: the SV% behind Regehr is ridiculous right now (94.5%), so expect that to fall a bit. Reggie is good - real good - but that's pretty inhuman, especially considering the quality of the competition he faces.

Which guys can we tab for a bound upwards? Dion Phaneuf (96.7), Mark Giordano (97), Brandon Prust (91.4), Wayne Primeau (95.4), Eric Nystrom (94) and...yes...Sore thumb (94.5).

Of course, some guys will be congenitally below or above average when it comes to either SV% and SH%. Bert, for instance, probably isn't going to have a massive swing upwards simply because he's not very good - he makes poor decisions with the puck and floats around the rink all game. Still, the club can't possibly continue to shoot at 5% while he's on the ice all year - that's absurdly low, even if you suck as much as Bert.

For a starker illustration, check out Boston's numbers. The teams SV% is 94.5! The lowest on-ice SV% of any player is 88.9% - and that's some unknown schlub who played maybe one game. In addition, 12 Bruins have on-ice SH% above 10. Only 2 Flames can make that claim. Now, Boston has some nice players and is playing well, but...really? I mean, are they really so much better than the Flames that half their line-up can drive percentages better than Calgary's entire roster?

Eh...probably not.

PS - Player #99 in the Flames link is the combination of Iginla, Cammalleri and Lombardi. Their corsi number is underwhelming (+8) while their PDO number is 109. Take from that what you will.