Saturday, July 30, 2011

A Contract I Actually Like, Part I

For the hockey geek, July 1 is Christmas, Easter, and St. Patrick's Day rolled into one. General managers, fat off expiring contracts and a rise in the salary cap, plunge their available cap space into whatever the market offers. No general manager, unless he's Brian Burke, wants to come up empty. Whoever spends the most often gains the least - the Buffalo Sabres are bound to regret the decisions they made earlier this month. I spent most of July 1 snickering at other teams' overspending. However, there were a few transactions I did like, and as we've gotten closer to training camp, I've liked more and more of the moves made. Usually the closer to the season we get, the better a contract is bound to be. There's fewer suitors for the stragglers. The contracts being offered are inevitably shorter, which usually makes them better - the team takes on less risk. So I've decided to start a feature entitled 'A Contract I Actually Like'. We'll see if I follow through on more (or if general managers do).

An incredulous tweet from June 30th whose source I cannot recall (likely @Kent_Wilson) referenced an agent referring to the 'Big 4' of unrestricted free agency as Brad Richards, James Wisniewski, Tomas Kaberle, and Ian White. Now while the other 3 players each got fairly large deals, Ian White took a more curious route. He signed with Detroit for only 2 years and 2.875M per season. Of players who signed in the calendar year 2011, White's contract ranked 38th in cap hit. 13 defensemen signed for more dollars per year, and 16 signed for more years. It's clearly a bargain when you consider that White has played over 20 minutes per game the last 3 seasons, he's a combined +17 playing mostly for mediocre teams, and has averaged .39 points per game over that time (31 points over 82 games).

From Ian White's Side Of Things

Ian White only turned 27 in June, so he's one of the youngest possible unrestricted free agents (aside from players who have accrued 7 seasons). He'll be 29 when this contract ends; plenty of time to get a more generous unrestricted contract. He's going to a team that needs to replace the 20 minutes a game of Brian Rafalski. He's also going to a Detroit team that generates a lot of offense:

Detroit's Offensive Rank in NHL:

2008-09: 1st (of 30)
2009-10: 14th (of 30)
2010-11: 2nd (of 30)

Detroit's Power Play Percentage Rank in NHL:

2008-09: 1st (of 30)
2009-10: 9th (of 30)
2010-11: 5th (of 30)

Detroit generates a lot of goals. That means more points for White. Presumably, it also means a generous contract after this one. I don't think I need to do a study on the correlation between a defensemen's points and his salary, but it's rather high - perhaps not as high as it is with forwards, but defensemen who put up a lot of points tend to be compensated quite well. Look no further than Tomas Kaberle's contract, which pays $4M per season for a player who struggled throughout this year's playoffs. If Ian White is able to score 40 points this season, who knows what kind of deal he gets when his contract expires?

Detroit's Side

Detroit isn't risking very much with this contract. There isn't a no-trade clause, or a no-movement clause involved. We've seen Ian White be traded 3 times over the last 2 seasons; this contract will move if they want it to. The dollars are reasonable and the years are reasonable.

They get a player who's played capably for 20 minutes a game who is capable of playing on the power play. Moreover, what they avoid is the potential overpayment of a point-generating defenseman; teams give too many dollars and years to a player who simply doesn't do much outside of the power play. As a comparison, the Calgary Flames gave Anton Babchuk 2.5M per season for 2 years, plus a no-movement clause - Babchuk has piled up 35 points in each of the last two seasons he's played in North America, but he only averaged around 17 minutes per game, and has mostly played against weak competition at even strength.

Conclusion

Detroit may have erred by giving Jonathan Ericsson a 3 year deal worth nearly 10 million dollars in total, but they certainly made up for that by getting White to sign such a discounted contract. Whoever gets White after the Red Wings, assuming he does go elsewhere, is likely buying high. White took a discount on this contract, but he's almost certain to make that money up on a future contract. Everyone involved gains besides the rest of the league, who once again have to deal with a formidable Detroit team.

Wednesday, July 20, 2011

Part I: The Aftermath of the Mike Richards and Jeff Carter Deals

On June 23rd, 2011, Flyers General Manager Paul Holmgren sent shockwaves through the hockey world when he dealt arguably the two most notable faces of the franchise – Mike Richards and Jeff Carter – to Los Angeles and Columbus respectively. Since then, much has been (and will continue to be) written about possible motives behind what he and team owner Ed Snider were pondering to make such bold moves. So far, speculation has included both off-ice issues as well as the need to create salary cap space for newly signed goaltender Ilya Bryzgalov. Starting with the former, Richards’ tumultuous relationship with members of the Philadelphia media is no secret. For the last few seasons, there have been accusations that he (and Carter) enjoyed a lifestyle where partying was the main focus, leaving hockey on the back burner. Richards’ leadership inside the locker room hasn’t been looked upon any more favorably. It has been rumored that a longstanding rift between the team’s young stars and seasoned veterans – most notably Chris Pronger – could also have played a factor in the stars’ departure from the team. Whether or not these accusations have merit, it is certain that the moves will accomplish one of the team’s likely intended goals: a culture change inside the dressing room.


In the aftershock of what happened almost four weeks ago, another highly-debated question has naturally emerged: are the Flyers still one of the premier Stanley Cup contenders in the Eastern Conference? In order to answer such a question, I am going to break my study into three parts. Part one will look into the deals for Richards and Carter themselves, evaluating what it was that the Flyers added to their lineup. Part two will evaluate Philadelphia’s signings on July 1 and speak to where Jaromir Jagr, Maxime Talbot, and Andreas Lilja fit into the equation. Finally, part three will decipher what the Flyers lost when they made the decision to deal their captain and his swift sidekick.


Of course, trying to answer our question is a bit of a double-edged sword – only time will tell if Holmgren’s return of Wayne Simmonds, Brayden Schenn, a 2012 second round pick, Jakub Voracek, the 8th overall selection in this year’s draft (Sean Couturier), and a third round selection in this year’s draft (Nick Cousins) were an adequate return for both superstars. Fortunately, we can still attempt to decipher what the numbers tell us about these players (and even draft picks). In order to do this, I like to start by looking at players’ average ice time per game. This tells us 1) what situations the players are being used in, and 2) how often they are being used. Per nhl.com, here are the numbers for Simmonds and Voracek, the two players coming to Philadelphia who saw significant ice time at the NHL level last season:



From these numbers, we can conclude that Simmonds played a bottom-six checking role in Los Angeles, the same role that he will most likely see in Philadelphia. Voracek, on the other hand, was one of the Jackets’ top forwards, ranking in the top 5 on his team at even strength and on the PP. This is good news for the Flyers – they will need him to replace any and all minutes in both situations without Richards and Carter in the fold.
With our ice-time analysis complete, we can now attempt to give this raw data its proper context.

 

What do these numbers tell us? Starting with Simmonds, he was asked to play a moderately defensive role against the toughest competition of any forward on the team (min. 20 games played). However, his zone start percentage should probably see an expected Corsi score of zero or slightly worse (see chart), and instead Simmonds sits at -4.02 . DobberHockey tells us that he most often played with Michal Handzus and Alex Ponikarovsky (21.25%), and Handzus and Kyle Clifford (15.33%). Of these forwards, Handzus and Ponikarovsky both have Corsi scores around what we would expect for someone given their roles, but no player seems to be “carrying the water” as we like to say. Unfortunately for the Flyers, Simmonds is no exception and his low scores in just about every category show that he cannot send the play in the right direction on his own against the opponent’s best players.


Voracek, on the other hand, is an interesting case in and of himself. Once again using the expected Corsi graph linked, Voracek actually slightly over-performs what we might expect from somebody given his zone start percentage. However, his impressive Corsi scores and Fenwick percentage are perhaps correlated with a few points of interest. First, his aforementioned high zone start percentage gave him an immediate advantage in generating shots towards the opponent’s net as he quite often started his shifts in prime scoring position. Second, his competition was anything but impressive, actually averaging a negative relative Corsi score. Finally, DobberHockey shows us that among his three most common line combinations, Rick Nash was on the ice a healthy 57.81% of the time. I hardly think explaining why playing with Nash would be beneficial to Voracek, but it is worth noting that Nash was among the league’s leaders in shots last season – his total of 305 ranked 6th in the entire NHL. Though Nash’s Corsi score of 4.49 may be lower than expected considering his own 57.1% zone start, taking into consideration how often he shoots it is easy to see how Voracek’s own score was undoubtedly affected for the better. It will be most interesting to see if Voracek can repeat such gaudy scores without a line-mate sporting the credentials of Mr. Nash.


We have already noted Voracek’s 2:57 of average PP time per game in ’10-11 which will be immensely valuable to Philadelphia in the absence of Richards and Carter. Though he ranked amongst the team’s leaders in said category, however, he only registered 8 total points for his efforts on the man advantage. Perhaps it is unsurprising that Columbus’ Power Play ranked 8th in the league in shots for/60 minutes with Nash in the fold (remember: shots are a better indicator than goals), but for a team that saw success in generating pressure on the opponent, Voracek’s totals still seem low. However, some of this effect can be explained when we realize that Columbus’ opponents saved 91.1% of all shots while on the PK according to Behind The Net. Thanks to mc79hockey, we know that the historic average is around 86.6%, a full 4.5% disparity. Had Columbus’ opponents not been so lucky, both the team and Voracek most likely would have sported slightly higher power play production.


Moving on to the relative unknowns of what the Flyers got back in the deal, on the surface Brayden Schenn and three draft picks may seem like an appetizing return. However, Derek Zona’s study on draft picks and their value tells us something slightly different. Putting Schenn aside for a moment, what can we realistically expect from the first, second, and third round picks that the Flyers gained? Zona’s study is particularly excellent because it shows the historical chance of drafting a “top” player with a certain selection. He notes to “...consider the 'Top Players' to be top five forwards and top three defenseman [on their team].” Knowing what we know about Richards and Carter, I don’t think anybody would argue that the Flyers subtracted two established “top 5” forwards from their lineup. However, the article also notes that the odds of drafting such a player with the number 8-13 selections (they selected Couturier 8th overall) is a mere 41.2%. Looking at the other two picks, the 68th overall selection this season which turned into Nick Cousins has a 7.4% chance of turning into a Richards or Carter-esque player. Considering the Kings figure to be among the top western conference contenders next season in the wake of their offseason, I would most likely expect the 2012 second round pick to fall within this same range. The fact that the Flyers didn’t receive higher than a 50% chance to replenish their lineup with two established stars puts a bit of a hindrance on their returns.


Adding this to the fact that both Richards and Carter were on long-term, cap-friendly deals, and I’m not sure that there is a net positive to be found here. Perhaps Sean Couturier or one of the other selections will make a difference, saving the Flyers money in the short-term should they produce while on an entry level contract. So far, all indications are that Brayden Schenn will be given every possible opportunity to make the final roster, but much like Couturier, there are still question marks surrounding his development. Unfortunately, prospects are called prospects for a reason – there is no guarantee that they will meet development expectations. Considering what the Flyers gave up in these deals, while the return could most certainly prove lucrative, the odds simply do not stack up in their favor.

Tuesday, July 19, 2011

On Why Entry Level Contracts Are More Valuable Than Ever

I guess I should make this a regular feature here at Driving Play - Pointing Stuff Out You Might've Known But Have Forgotten. I'll work on the title. The first entry in this series was about the decline in goal scoring since the lockout. This one is about the effect a rising salary cap has had on entry-level contracts.

Every player who is drafted and signs with their drafting team signs an entry-level contract (ELC hereafter). The rules about entry-level contracts are fixed - there's certain bonuses that can be given out, but there is a clearly defined maximum salary that depends on what year the player was drafted. Here's the list:

2005: $850,000
2006: $850,000
2007: $875,000
2008: $875,000
2009: $900,000
2010: $900,000
2011: $925,000

So, from 2005 to 2011, the maximum salary in an ELC has increased by nearly 9%. That sounds almost substantial, but we have to remember that the salary cap has increased by 64%, from $39 million to $64 million. Leaving bonuses aside, a player making the rookie maximum in 2005 would be making 2.1% of their team's possible salary allotment. In 2011, that player will make 1.4%. That sounds insubstantial, but in effect, it serves as a large pay cut. The difference in the two percentages (.7%) is almost enough for an additional minimum salary player on the roster (.82%).

Okay, so you might say, 'Well Triumph that all sounds well and good, but what if teams are paying pretty much everyone that maximum, thereby reducing that effect?' But they're not. I did a study on 15 teams and the ELCs they signed that go into effect next season. I chose 15 teams because it took a long time and because I'm kinda lazy, but I can assure you the numbers won't change much if I did all 30. Here's what drafted players are getting paid, on average (numbers courtesy of capgeek.com):

1st round: $909,000
2nd round: $797,000
3rd round: $631,000
4th round: $677,000
All Others: $625,000

Not only will any of these contracts provide a substantial discount if the players make it to the NHL, but we've seen that negotiating a 2nd NHL contract can be very tricky. If the player has only played 3 years in a professional league, he does not have arbitration rights. This means that his only negotiating recourse is to get an offer sheet from another team, which are rather rare. We saw recent RFA Karl Alzner sign a contract at a massive discount because he essentially admitted that he had no choice.

We shouldn't expect to see this situation change in a future CBA, either. Players who are draft-eligible aren't part of the union, and no doubt teams don't want to give young, unproven players more money. The value of entry-level contracts should increase in the NHL going forward, making the NHL Entry Draft even more important than it already is.

Friday, July 15, 2011

On Why Adam Larsson's New Entry Level Contract Is Not The Steal It Appears To Be

In 1993, Alexandre Daigle made headlines with the contract he signed - 5 years, 12.25 million in total. He then used this money to buy a dress. Or something - that was ages ago, and I can't be counted on to remember every single detail. Whatever the case, that was an absolutely outrageous sum at the time, and led to the creation of a rookie salary cap in the NHL's next CBA. The rookie wage system has been even more refined since then.

Note: Some of you will already know all of this stuff, but I think the fullest explanation is usually best, especially when starting a blog.

In 2011, Ryan Nugent-Hopkins was drafted first overall. He signed a contract that pays him $925,000 per season in base salary. $925,000 is the maximum rookie salary allowed by the CBA. He can also receive a potential $2,850,000 in bonus money each season. Mika Zibenijad was drafted sixth overall in 2011, and he signed a contract which pays him $925,000 in base salary per year, along with $850,000 per year in possible bonuses. Adam Larsson was taken fourth overall, and he signed a contract which pays him $925,000 per season, along with $000,000 in bonuses. Yes, you read that correctly - he is receiving no player bonuses. How could the 1st overall pick and 6th overall pick get bonuses, while the 4th overall pick gets none? Let's delve into the murky underworld of the collective bargaining agreement, where even Gary Bettman fears to tread.

A: The Bonus Cushion Does Not Refer To Weight Gained Thanks To Expensive Dinners Out

Section 50.5.h of the Collective Bargaining Agreement sets out the definition of the bonus cushion. If you prefer speaking Lawyer, here is the full text:

"A Club shall be permitted to have an Averaged Club Salary in excess of the Upper Limit resulting from Performance Bonuses solely to the extent that such excess results from the inclusion in Averaged Club Salary of: (i) Exhibit 5 Individual "A" Performance Bonuses and "B" Performance Bonuses paid by the Club that may be earned by Players in the Entry Level System and (ii) Performance Bonuses that may be earned by Players pursuant to Section 50.2(b)(i)(C) above, provided that under no circumstances may a Club's Averaged Club Salary so exceed the Upper Limit by an amount greater than the result of seven and one-half (7.5) percent multiplied by the Upper Limit (the "Performance Bonus Cushion")."

Translating, it means that the club is permitted to go over the salary cap by as much as 7.5% so long as the excess is made up of money that is to be paid via bonus. So let's posit a $60 million salary cap, a team could go up to $64.5 million in total money, if $4.5 million of that is money that may be earned via bonus. Simple enough, right? Any money that was actually earned via bonus, if it was in excess of the $60 million salary cap, would carry over to the next season. However, there's this little catch written in at 50.5.h.iii.C:

"If the NHLPA exercises its option to extend this Agreement to September 15, 2012, consistent with Section 3.1(b) of this Agreement, then for the 2011-12 League Year, all of the above-described Performance Bonuses that could be earned by the Players under SPCs with a Club shall be counted against such Club's Upper Limit for that League Year (with no opportunity for the Clubs to "carry over" any charges to their Upper Limit for the following League Year)."

The NHLPA has exercised their option to extend the agreement to 2012, and as a result, there is no bonus cushion this season. Ryan Nugent-Hopkins therefore has a $3.75 million cap hit this season.

B: The New Jersey Devils Are Up Against The Salary Cap

I don't want to delve too much into this issue, as it complicates things. Let it be said that the Devils, according to capgeek.com, have $6.7 million in salary room with 21 players signed. That sounds like a lot of room, but that player yet to sign is Zach Parise. He will no doubt command a large salary. Had Adam Larsson taken the normal amount of bonuses in his contract, his salary cap hit would have been somewhere between $2 million and $3.5 million. The Devils would likely not have been able to afford him this season unless they got rid of other players in order to make room. That's not a move that would make a whole lot of sense, given how poor the typical 18 or 19 year old defenseman is.

C: The Entry Level Contract Slide

Ok, so we're Adam Larsson now, and we've signed a hypothetical $3 million dollar per season deal, with most of that money coming in bonuses. We get sent down to the AHL and we stay down there the entire season. We lose a year off our contract, right? Nope - if a player is 18 or 19, and has a signed contract, his contract doesn't begin until he plays 10 NHL games in a season. The contract 'slides' - a contract can slide a maximum up to twice. So even under this scenario, Larsson would still have 3 years left on his entry-level contract. Even Karl Alzner, who pundits are saying left a lot of money on the table in his current deal, now makes $1.3 million per season, a step up for $900,000 and bonuses. Larsson would make $925,000 if he plays the full season in the NHL, but significantly less than that if he plays in the AHL (probably $70,000).

D: Performance Bonuses - What Are They?

Exhibit 5 of the Collective Bargaining Agreement sets out what items a team can give bonuses for.

For a defensemen, they are as follows:

Top 4 On Team Among Defensemen In Ice Time (min 42 GP)
10 Goals
25 Assists
40 Points
.49 Points Per Game
+/-: Among Top 3 D On Team (min 42 GP)
All Rookie Team
All Star Game
All Star Game MVP

None of these bonuses can exceed $212,500 for any individually, or $850,000 in total. Let's examine, since 1995-96, how often defensemen aged 18-21 have hit some of these plateaus. There are 165 player seasons in total. As a proxy for top 4 in ice time, I am instead using 1500 total minutes played.

1500 minutes played - 47 times
10 Goals - 21 times
25 Assists - 31 times
.49 PPG - 18 times

So, he might have hit some of these performance bonuses - certainly 1500 minutes played seems pretty likely by Year 3, if he is as good as they say he is. Victor Hedman, the player that people compare him to endlessly, has done that for both his seasons in the NHL so far.

The other performance bonuses are more outlandish, but the relevant ones include winning the Norris Trophy, Conn Smythe Trophy, or Calder Trophy. He would also get a bonus for finishing in the Top 3 of Calder voting. Any or all of these are seriously unlikely to happen.

E. Getting A Head Start On A Contract Is Massively Valuable

Let's go over some rules on free agency now.

A player becomes an unrestricted free agent either when he is 27 or older on July 1 of that calendar year, or has played 7 NHL seasons. If Larsson were to begin his contract this season (by playing more than 10 games, remember), he would be eligible to become an unrestricted free agent in July of 2018. That's of course assuming that the rules governing free agency don't change. If Larsson's contract began next year, he would be eligible to become UFA in 2019, and his contract must begin the year after; he would be a free agent in 2020.

Anyone who was paying attention on July 1 saw the exorbitant salaries that even mediocre players command when they become unrestricted free agents. Larsson playing this year would mean he'd be able to take part in that frenzy two years earlier than his birthdate would indicate.
Larsson would also be a restricted free agent in 2014 if his contract began this year - this means a higher salary, although had he signed for, say $3M, he might've had more potential bonus money. Still, we've seen how difficult it is to hit some of these bonuses. A star like Drew Doughty will certainly earn significantly more in his next contract than he ever could have with bonuses.

F: Conclusion

Without the specifics of each entry level contract and what bonuses are placed in there, it's hard to say just how much 'potential' money Adam Larsson passed on by not having those bonuses in his contract. Regardless, with New Jersey's potential salary cap difficulties, Larsson has secured himself a chance to play on this year's team. If he does play 10 games this year, he will almost certainly manage to make up the potential money lost in bonuses with a higher salary in 2014-15 and 2018-19, as those are the years he is eligible to be an RFA and UFA respectively. While the Devils are presenting this deal as a kind of sacrifice, it's also rather shrewd. If Larsson is as good as people are saying, he'll have the opportunity to make up this cash elsewhere. If he's not that good, he wouldn't've made that bonus money regardless.

Yeah, But: QualComp

"Going by these stats, X is better than Y." "Yeah, but…"

In the Yeah, But series, I will be taking a closer look at the stats lurking in the background. We don't care about them per se, but they provide context and help us compare and rate players. My initial plan was to start the series with a discussion of shooting percentage, but after reading this article at arcticicehockey by Dirk Hoag this seems like a good time to take a deeper look at QualComp.

If you've ever argued with someone about whether one player is better than another, you've probably had a conversation that went something like:

Pens fan: Letang had a better year than Lidström. Look at how much better his +/- and Corsi stats are!
Wings fan: Yeah, but Lidström faced much tougher opponents!

Here are their Corsi/60 and Corsi QualComp stats, courtesy of Behindthenet:

PlayerCorsiCorsi QualComp
Lidström2.651.205
Letang12.010.281

Roughly speaking, Lidström put up decent numbers against the toughest competition in the league, while Letang did very well against opponents that were a little above average. How much should we boost Lidström's numbers to compensate?

The obvious way to figure this out would be to list the Corsi QoC and Corsi rate for each player (maybe above some time-played threshold) and run a regression. Unfortunately, that won't work very well and will drastically underestimate the importance of QualComp.

Say you are watching your favorite team play on home ice. It's tied in the second period and, with both teams at full strength, you see the opposition put their best players on the ice. Let's stop there. What should you expect in the next minute or so?

The first thing is that the opponents will be good, so you shouldn't expect much. That's what I'll call the competition effect - the better the opposition, the worse your expected results will be. This is what we would like to measure. On the other hand, your coach will see who is on the ice and will probably put out one of the top lines and a good defense pairing. This means you should expect good results. I'll call this the matchup effect - the better the opposition, the better your players on the ice, which will raise the expected results.

Which one will win out depends on how important the competition effect is and, importantly, how much the coaches focus on the matchups. In the regular season, they put some importance on them, but they have to take the long haul into account. In the playoffs, matchups get a lot of focus by everyone from casual fans to bloggers to on-air analysts to the coaches themselves. Barring serious injuries like Crosby's concussion, coaches do close to everything they can to win the game they're in. A top line facing a bottom line is rare, especially on face offs that don't follow icing. In the playoffs we would expect the matchup effect to dominate the competition effect. More on that in a bit.

To see how this all works, let's oversimplify things and say that Corsi rates are simply additive (subtractive?) - if your line is +5 and you face my line, which is +2, then your Corsi rate will be 5 - 2 = 3. Here's a graphical representation a possible matchup between two teams:


You can see the competition effect within each line - the slope of the line connecting the points is -1, which comes from the (over)simplifying assumption that Corsi rates are additive. You can see the matchup effect by noting that as you move from left to right from the opponent's fourth line to their first, you will tend to also move up toward where your first line is because of matchups.

Let's look at a real-world example - the first-round series between the Vancouver Canucks and the Chicago Blackhawks. There are several things about this series that make it a good one to demonstrate the matchup effect. It's in the playoffs, when coaches go out of their way to avoid randomizing matchups. The series went seven games, with a couple overtimes for good measure, which provides more data. The sample size is in a sweet spot where it's small enough for me to be able to break it down but big enough for there to be evidence of the effects I'm trying to demonstrate. The teams also fit the bill nicely - both feature two very strong lines with a significant dropoff. As usual, I will restrict attention to 5-on-5 play where both goalies were in net.

Overall, Vancouver dominated this series 5-on-5. Chicago had no answer, apart from Crawford in goal, for either of the Canucks top lines. Meanwhile, the Blackhawks' best struggled. Looking at the matchups, the only place where Chicago had an edge was when the bottom lines faced each other. The Blackhawks show the competition effect quite well - those players, including their best, that faced the best on the Canucks had awful stats while those playing mainly against the bottom lines did better. On the other hand, the Canucks were an example of the matchup effect swamping the competition effect. The players that faced the toughest opposition were their top players, and they did extremely well. The less skilled players struggled, even though they played far worse opposition.

Here are three plots giving the QualComp and Corsi rates for all players with at least 15 minutes of ice time against both the top and bottom two lines (combined) of the other team. For QualComp, I took the weighted average of the regular-season Corsi from BTN.


It doesn't look like much of a relationship and this is confirmed in the regression, with similar results to what Dirk Hoag discussed:

Coeff. t R2
Corsi QualComp-0.54 -0.230.0016

Instead of averaging everything out, let's look at within-player results. That is, after all, the information we want to know - e.g. if Lidstrom played against competition similar to Letang's, what would we expect his Corsi to be? To do this, we need to create some kind of split in the data. Vic Ferrari compared the first half of the season and the second. I will instead look at performance when facing the opponent's top two lines compared to the bottom two. Keep in mind that I'm only trying to demonstrate the matchup effect. Trying to determine the importance for the league based on a 7-game sample between two teams would obviously not work all that well.

Here is a similar scatterplot to the above, but this time with the data split:


Now that we're looking at differences for each player, we can see a much stronger relationship. It is stronger for the Chicago players but the most impressive change is on the Vancouver side. Aggregating all the minutes, there was a positive relationship between opponent strength and Corsi rate due to the matchup effect. Splitting it up we see that the Canucks players tended to be less productive against better opposition, although the effect was quite weak - this almost surely would be bigger over a more reasonable sample.

Running the same regression on the pooled data gives us:

Coeff.t R2
Corsi QualComp-1.92-2.870.1054

Splitting the data up so we capture the change in Corsi for each player gives us a regression indicating that the effect is over three times as large, is statistically significant (before it wasn't even close) and an R^2 over 65 times larger.

When looking at correlations, you have to be careful to think about how coaching decisions affect everything. Good players tend to play with other good players and against other good players. This influences basically all the "yeah, but…" stats including quality of competition and teammates, zone starts and special teams. I'll explore those more in future articles. To measure effects like Qual Comp it is important to use a method that captures changes for a player or group like what I've done here, WOWY or the method Vic Ferrari used in his article on Qual Comp.

In a future article I will try using a similar methodology on a larger scale to figure out how large the effect actually is.

Tuesday, July 12, 2011

On The Goal Decline Since 2005-06, and Its Effect On Individual Goal Scorers

The non-famous Greek philosopher Heraclitus said ‘All is in flux’. He also said ‘You can never step into the same river twice.’ These two notions should be our guiding principles through this examination. I will be demonstrating how NHL scoring has changed over the last 5 years. It’s a story not often talked about in the media, and as a result, fans aren't aware of it either. I'll see fans predict X number of goals for the top players, this guy'll score 40 goals, that guy should get 30, and I want to reach into their computer to show them these numbers. At the rate offense is declining, there's a limited number of players who I would consider favorites to score 30 goals in a given season.

To first examine this issue, let's just count up the average number of goals scored per team since the lockout. The NHL, in its infinite wisdom, decided to count shootout wins as ‘goals’, so to examine lea
gue-wide offense, we’ll have to get those out of there. We've done so in this helpful chart (all stats in this entry are courtesy of nhl.com and hockey-reference.com):


I also included the number of overtime games per team - it shouldn't change the numbers very much at all, but it is worth noting that teams play an average of one more OT game per season. We see a general decline in goals - 24 per team since the lockout.

So, goal scoring is down. Why has goal scoring gone down? It's pretty simple, actually - power plays. Power plays provide the best opportunities for offense, and power plays per game have dropped considerably.




As we can see, power plays per game per team have dropped by over two since 2005-06. Each team had on average, 33 fewer power play goals in 2010-11 than in 2005-06. Power plays are still falling - where should we expect next year's number to land? The league office has yet to issue one of its 'Hey, We're Actually Going to Call The Rulebook' decrees that were so common in the late 90s and early 00s. Those 'tight' calls would last until December and be a dim memory by the time the playoffs rolled around. Without one of those mandates, I see no reason to expect power play opportunities to rise next season, and I would expect to see them between 260 and 280 per team on average.

Goals are down, power plays are down, so how does this filter down to the individual player? It's certainly in a more erratic fashion than goals or power plays per team, but we cannot help but note the downward trend. This chart shows how many players scored a minimum of 40, 30, and 20 goals in a given season:


Injuries are going to foul this chart considerably - Sidney Crosby would have been a lock for 40 goals in 2010-11 had he stayed healthy. Regardless, only 5 players scored 40 goals last season, and there have been only 19 40+ goal seasons in the last three years. I attribute this to the decline in power plays - with fewer power plays, it means the top players get fewer opportunities to score. Average power play time for the top players has declined considerably. However, while 40 and 30 goal scorers are down considerably, there were only 16% fewer 20 goal scorers in 2010-11 than 2005-06. That's where to make your safe bets - your best players are still pretty damn likely to get at least 20 goals.

So yes, all is in flux, and no, you can't step in the same river twice - we don't know exactly what's coming next season. Offense could go up. Regardless, I see no reason to expect it to rise. So, when an overeager fan offers you a bet about his favorite player's goal scoring, take the under and spend that money.

Friday, July 8, 2011

On Power Play Points By Defensemen And The Limits of Knowledge

Is there any more of a buzzkill blogpost title than 'The Limits of Knowledge'? No, probably not. Still, I like exploring questions more than I do necessarily getting the answers.

One of the things that bothers me about hockey in general is that single unit of scoring, the point. Do we know why secondary assists are given on the vast majority of goals? Why is that the standard? How often have you watched a game and seen assists given for a blocked shot or goalie's save? Yet, these points are the way in which most hockey fans judge a player. If he's got a lot of points, he's good, if he doesn't, he's not. While this is largely true for forwards, for defensemen it creates a problem, because there is no true 'defense' stat. There are hits and blocked shots, but measurements of these things vary wildly from arena to arena. Plus/minus measures offense relative to defense, so that's really no help. What I see fans doing too often is relying on points, even though I think most of a defenseman's job is to prevent goals, not to contribute to the scoring of them.

So, as is my custom, I made a spreadsheet. I wanted to examine power play points, as it is my belief that not all points are created equal. Why should we necessarily reward defensemen who get a lot of points on the power play? Are they better than their peers at the power play - so much so that they should be heavily compensated, as James Wisniewski was this off-season? So this spreadsheet lists every defenseman who played 40 or more games last season and got at least 2 minutes per game on the power play, the number of points they got, the number of total PP minutes they got, their points per 60 minutes of PP ice time, and the shooting percentage* while they were on the ice *note: shooting percentage only covers 5 on 4 PP time. Voila. (Stats courtesy of nhl.com and the incomparable behindthenet.ca)


That looks all well and good, but what does it all mean? As is all the rage these days, I ran a Pearson's correlation on some of the important numbers. A Pearson correlation takes two sets of variables and shows how closely related they are. A Pearson correlation ranges from -1 to 1 - -1 means they are inversely correlated, 1 means totally correlated.

First, I looked at PP Time and Total PP Points, which is silly - they are of course highly correlated, and I got a Pearson value of .84, which duh. Next I looked at the relationship between Points/60 and Power Play S%, and I got a high value there - .66. Next I looked at the relation between total power play time and Points/60, and I got a reasonably high value - .55. Lastly, I looked at the correlation between power play points and shooting percentage. All of these, according to the good people at Wikipedia, are strong correlations. Let's list them out:

PP Time/Total PP Points: .84
Points/60 - Power Play S%: .66
Total PP Time - Points/60: .55
Total PP Points - Power play S%: .50

Okay, so to put this in simpler terms, I was asking, aren't guys who get a lot of points on the power play also just guys who get a lot of power play time? This suggests, yes, that is pretty much the case. Okay, and aren't a lot of power play points related to power play shooting percentage, and this says, yeah, a lot of the time, that is true. We've generally seen that shooting percentage at even strength is more luck than skill, so why shouldn't that hold true on the power play? With all that being true, I looked at power play time and points/60, and basically the more time on the power play a player gets, the more points/per 60 minutes he scores. And back again to the beginning, the points a D man gets on the power play are related somewhat closely to the shooting percentage while he is on the ice.

So didn't I prove what I set out to prove? Not really. We've once again run into the biggest problem in hockey: Usage. It's especially true on that third correlation - total power play time is correlated with a player's points/60. What this means is that the more points a player gets, the more time he will end up getting. All this time we've been looking at individuals, but the problem with hockey is that A: it's a team game and B: the coach chooses at almost all points of the game what players he wants out on the ice. So the fact that the players better at generating more points get more ice time isn't at all an accident - their coaches are putting them out there more often because ostensibly they're better. And the fact that the power play shooting percentage is higher with those players on the ice might not be an accident either; the first unit power play shooters may be more skilled. We also have the problem of taking such a limited sample - 300 minutes of ice time is equal to around 20 games' worth of even strength ice time for these defensemen. We'd need multiple years of data to try to prove anything substantial.

The overall relationship between defensemen and offense is nebulous, and this study hasn't really helped to clear anything up. Still, unlike some sites that will just publish anything even with faulty logic at its heart, Driving Play will admit when its studies don't reach a logical conclusion. I may follow up on this next season.