Category Archives: In-Depth Stats

A Tale of Two Teams

Today I’d like to discuss two very different hockey teams.

The first team is consistently one of the least penalized teams in the NHL, drawing very few penalties per game. Let’s call this team “Team Passive“. The other team, by comparison, is one of the most heavily penalized teams in the league, averaging significantly more penalty minutes than most — “Team Aggressive“.

Let’s compare each team’s penalties in minutes (PIMs) per game over the span of a season:

pims-per-game-passive pims-per-game-aggressive

The differences are pretty clear. Team Passive rarely exceeds the league average of about 9.7 PIMs per game while Team Aggressive does in more than half of their games. But I’d like to dig deeper than simply penalty minutes. As you can see, even Team Passive has the occasional 40 PIM tilt where all hell breaks loose. Let’s look at each team’s minor penalties per game. I want to remove the statistical noise of fighting infractions and coincidental minors, since these are always 2-sided penalties. Let’s focus on those tripping, hooking, boarding calls that might give us a better idea of each team’s play style. Here is how each team compares to the league average of 3.4 minor penalties per game:


These charts look fairly similar to the PIMs charts above, but they start to tell more of a story about individual games. Note that Team Passive has not exceeded 6 minor penalties once in the season while Team Aggressive has done so 11 times! Conversely, Team Passive played 6 games without drawing a single minor penalty; a feat that Team Aggressive could not accomplish a single time. Note that, as predicted, removing fighting majors and coincidental minors has pushed each team’s above/below league average frequency further to its respective extreme.

So what’s the real story here? Perhaps one team plays a more physical style and the other relies on speed and skill? Is it possible that these teams’ different styles of play simply lead to more/fewer penalties called in the games in which they participate? Let’s compare their minor penalties per game to their opponent for that game, rather than the rest of the league.


This view adds an interesting perspective. Specifically, games in which both teams and their opponents had the same number of minor penalties; a stat which, perhaps surprisingly, is almost equal between both. It’s worth noting that the discrepancy between teams is much smaller here than in the previous charts. The difference between teams in the “fewer minors than opponent” category is only 16% whereas in the comparison to league average it was 40%. This is due to the fact that to a small extent, there actually is a difference in how many penalties each team’s opponent draws on average.

Team Passive‘s opponents drew an average of 3.2 minor penalties per game, while Team Aggressive‘s drew an average of 3.8. Considering an overall league average of 3.4 minors, these are fairly significant variances, particularly in Team Aggressive‘s case.

Where are you going with this?

By now I’m sure some of you are saying “So what? Different teams play differently”. You’re right, but that’s just it; teams Passive and Aggressive are the same team.

Let me present to you, Team PassiveAggressive:


If you haven’t figured it out yet, Team PassiveAggressive is the Calgary Flames. As of January 19th game against the Nashville Predators, they have now played 82 games since the now-infamous “Wideman Incident” (which, ironically, happened in a game against the Predators). So yes, I played around with the facts a little bit when calling each team’s “season” a season. They are in fact two 82-game segments which have straddled the last 3 seasons, going back to the halfway point of the 2014/15 season. Now, I know that the point has been raised many times already, more recently even by team management. But I wanted to take a more critical look at the “Wideman Effect” as fans have coined it. How real is it, how drastic was it, and also, how can it still be a factor almost a calendar year later?

Looking Back

First, let’s try to understand where the Calgary Flames were before the incident. Here is a historical look at the Flames’ discipline compared to league averages over the past few years:


So many things jump out of this chart. You can see the steadily declining league average, indicating the less physical nature of the game, even over a 10-year span. You can see how much impact a coach and his system have on a team’s discipline (or at least, it’s propensity for taking penalties). Keenan’s Flames played a tough, physical game; Hartley’s Flames were a quick skating dump-and-chase team, and Sutter’s Flames were very… average.

There is no doubt that the Wideman incident led to backlash and more stringent penalties called against the Flames in the second half of the 2015/16 season. There is no other way to explain such a dramatic change in a team’s identity halfway through a season, in particular as the Flames were all but out of the playoff race by this time and the intensity of their games would have likely waned, not increased. Look at each coach’s “identity”, Hartley’s system did not change mid season, his approach did not all of a sudden cause in influx of penalties against.

This does lead to an interesting question however: what is Gulutzan’s identity? The Flames’ penalties per game have actually increased in the first half of 2016/17, under Gulutzan’s watch. Has officiating bias really followed the Flames through the off-season? Or perhaps Gulutzan’s system simply has the Flames playing a game more similar to Keenan’s, with the added calls being a byproduct of a hard-nosed system? To answer these questions, we’ll need to look at penalty calls against individual players, and hopefully find more insight.

Penalties by Player

Let’s first jump back to the immediate post-incident impact. Here is the player-by-player breakdown of penalty calls before and after the Wideman hit. We are looking specifically at the 2015/16 season, and only players that played at least 20 games before and after the incident.

It is pretty clear. 11 of the 15 players saw an increase in their minor penalties per 60 minutes (used in order to compare all players regardless of average ice time). Interestingly though, the average of all players went from significantly below career levels to just slightly above. This could indicate more of a “return to norm” rather than blind vengeance by the officials (more on this later).

Let’s see what’s happened to these players’ penalties per 60 minutes under Gulutzan (excluding players no longer with the team):


This chart paints a less clear picture. Some players have reverted back to their career average (both upwards and downwards), other players remain above or below theirs. The team average seems to have remained slightly above average, but I believe it is being buoyed by a small number of players. If we exclude Frolik, Hamilton, and Bennet from the average, we see an almost perfect return to the pre-incident average among the other 10 players. Is this a bit of a stretch? Possibly. But let me present one more item before drawing conclusions.

“Wideman Effect” or New Identity?

Let me clarify: I have no doubt that the final 34 games under Bob Hartley were officiated under a very different standard than the previous 48, or the final 34 games of any other NHL team for that matter. I am more interested in what is happening today, and whether the Flames have simply become “that kind of team” now, while the doubt cast by last season’s events has obscured that fact from most.

First, let’s establish something pretty critical: Hartley’s “Passive” Flames were much more statistically unusual than the current “Aggressive” iteration. If we take the same 164 game time frame and break it down to two 82 game segments for every team, the resulting 60 “Teams” have had a minor penalties per game distribution that looks something like this:


Note that the Flames on the low end were significantly further from the average than the Flames on the high end. In fact, only the Carolina Hurricanes even come close to Calgary’s level of discipline. By comparison, the Aggressive Flames are outdone by Winnipeg and Columbus, both of whom took more penalties per game (both during the Passive Flames period). My point here is not to diminish the reality of the Wideman Effect, but to set some parameters around it. Statistically speaking, the Aggressive Flames are much more likely than the Passive Flames, all other things being equal.

But enough about probabilities, what is this “New Identity”? We’ve established what happened to the individual players’ penalty frequency during the period after the Wideman incident and in the first half of 2016/17, but what about the players that have come and gone? Much of the team’s identity lies with it’s core players, true, but there have been many significant moving pieces around them. In order to isolate player movement and to compare this season to the last season more specifically, here is a comparison of the first 48 games of 15/16 (the end of the Passive period) and the first 48 games of 16/17:


And there it is. While the core group of players has certainly gained some penalties per game, the net effect of players in and out of the lineup is just as drastic. If we look at players with values greater than 0.1 (equivalent to 5 minors per 48 games), the Flames replaced 2 such players with 5. Tkachuk alone replaces the minor penalties of the 7 lesser penalized players to leave the team (Smid to Colborne). Furthermore, and to pick up on the “Average excluding some players” from before:

Excluding Frolik, Hamilton, Bennet, and the addition of Brouwer, Chiasson, and Tkachuk, the Calgary Flames are no more or less penalized today than they were before the Wideman Incident.


There are obviously many different ways of looking at this. You can present data in various ways to prove your point. It is probably worth noting that I am indeed a Calgary Flames fan, and have tried (to the best of my ability) to take an objective view on this whole issue. After looking at the facts and figures presented above, I believe that there was absolutely an immediate impact after the Wideman hit last season. In all honestly, it would be kind of ridiculous to expect some backlash. The incident itself was unfortunate, and the NHL’s handling of the investigation, Wideman’s own behavior and attitude afterwards, and the team’s philosophy of protecting their own all led to a very tense period. Referees are human after all, and they have biases too.

Looking closer at the individual players’ numbers however, I am more skeptical of any real “Wideman Effect” in the 2016/17 season. I think that Calgary is simply a much more physical team under Glen Gulutzan. They have drafted, traded, and promoted tougher players than those they let go (with Tkachuk in particular becoming a league-wide pest almost instantly). As long as Gulutzan remains behind the bench – and Burke as president of operations for that matter – I believe that this is the new reality for Calgary.


First Game Back After Road Trip

A hockey truism that I’ve heard a few times in the past is that the first home game after an extended road trip is a “trap” game, and that the longer a team is on the road the less focused they are when they finally return home. Supposedly, the players are so happy to be back home after a grueling trip that they don’t always show up for that first game back. I’ve never given much thought to this idea, but it always struck me as a bit odd.

The other night I was watching the Sportsnet pre-game show and Colby Armstrong made the peculiar statement that it was “scientifically proven” that the first game back after a road trip is a difficult one for teams to win. That kind of statement piqued my interest enough to actually grab some data and see if there’s any truth to this idea.

I looked at all NHL games played over the past 10 seasons.

The first thing to note is that home teams win 55.0% of the time. So, yes, home ice advantage is a real thing. However, it’s only the first home game back that I was interested in. Of the 11,790 games played in the past 10 seasons, only 5,423 were games where the home team was playing its first game back from the road. And how did they do? The home teams won 2,968 of those games, or a win percentage of 54.7%. A slight drop from 55%, sure, but not enough to be a “thing”.

In fact, when breaking it down by the length of the preceding road trip, I found even more interesting results that contradict the 2nd part of this truism: that longer trips mean even less success upon return home. Take a look at the chart below:

First Game Back

Surprisingly, teams back from 4, 5, and 7 game road trips perform better than average home success. In fact, with the exception of 6 game trips (which may simply be a sample size issue), 1 game trips seem to have the worst affect on the success of the following home game. This could possibly be attributed to quick turnaround and busy travel schedules.

Overall, there seems to be little support for this idea. Home games are home games, and with a large enough sample size there seems to be very little deviation from the standard 55% home win ratio.

And yes, there was one 14 game road trip… (Bonus points if you can figure out who and when).

Does first place matter?

I’ve decided to take a look at the historical champions for all 4 major sports leagues. I want to see what sort of trends emerge when plotting these teams by their final standing. Lets start with the NHL, I will explain some of my methodology as I go along.

NHL Stanley Cup Champions

NHL Champions by standingI went back to the 1926-1927 season as this was the first time the NHL used a playoff system with more than 3 teams (and also the first season with multiple divisions). I used overall record to determine standing, which is why you’ll see many seasons in which lower-ranked teams made the playoffs over higher ones. In the second chart I compare the actual total of champions by standing to the expected number assuming 50/50 win rates. This means that in a season from the 50’s with 4 playoff qualifiers, each is assumed to have a 1/4 chance of winning the cup. Similarly, in a modern season with 16 playoff teams, each is assumed to have a 1/16 chance of winning. The curve of this line is not due to higher placed team having a higher likelihood of winning, but is actually due to the fact that past seasons had fewer playoff teams, meaning that lower ranked teams may have had no chance of winning at all. This is why the top 4 teams all have equal expected numbers; the top 4 teams have always qualified to the playoffs.

So, what does the actual data show? It is pretty clear that 1st place teams have won a highly disproportionate number of championships over the years. If winning was truly random, we’d expect about 12.2 1st place champions, but in fact we have had 39 of them. Conversely, we see that every other position but 2nd has had fewer championships than expected. This seems to indicate that the best team during the regular season really does have a higher chance of winning it all. The dominance of 1st place team may stem from a rich history of dynastic teams, especially during the “original six” period. This conclusion may seem rather obvious, but lets see if it holds for the other leagues.

NBA Champions

NBA Champions by standing

The history of NBA champions seems pretty similar to the NHL’s, with one notable exception. In this league, both 1st and 2nd place teams have won many more times than expected. We’d expect to see about 6.7 championships for both 1st and 2nd place teams, but there have actually been 30 and 19, respectively. This trend can be attributed in part to the legendary rivalry between the Celtics and Lakers, which often faced off as the top two teams in the league. 12 of the Celtics’ 17 championships were won as the 1st ranked team in the regular season. 8 of the Lakers’ 16 championships were won as the 2nd ranked team. Overall, however, we still see the top teams well ahead of the expectation curve. Lets see if this holds for football.

NFL Superbowl Champions

NFL Champions by standing

This data set should probably be taken with a grain of salt, as the NFL’s short season means that there are very often teams with identical records at season’s end. As such, I had to simply round teams up or down (for example, determining which of 4 teams with a 12-4 record was technically first). Regardless, there is a similar trend once again. The top 2 teams win most consistently with the top team overall winning 21 championships, compared to an expected 5.7. This is getting pretty repetitive, isn’t it? Well, then there’s baseball.

MLB World Series Champions

MLB Champions by standing

Well, look at that. MLB champions seem to line almost perfectly with the expected line. Beginning in 1969 – the first year with a true playoff structure – baseball has had a very “random” pattern of champions. There really appears to be no consistent trend, with only 2 occurrences of the top team winning consecutive world series. This is probably due to the fact that baseball has the greatest disparity between season length and playoff length. Fewer playoff series could allow lower ranked teams to “steal” a championship more often than other, more rigorous post-season structures. Then again, the NFL has a similarly short post-season (and single elimination games to boot), and its results seem less sporadic. Perhaps this indicates that the MLB’s 162 game season allows teams to build up such leads that they win the regular season by mid-summer and enter the playoffs on a cold streak or plagued by injuries, whereas other leagues with their shorter seasons would see teams in similar situations quickly drop down the ranks.

Final thoughts – where have all the dynasties gone?

While focusing on individual leagues, I did not touch on another interesting trend they all have in common. There is a very noticeable dichotomy in the NHL, NBA, and NFL when it comes to 1st place champions. Since the 1990’s, there have been very few such teams. Since 2000, of the 51 champions across all 4 leagues, only 8 finished the regular season 1st overall (about 16%). Since 2005, its been only 3 of 31 (less than 10%). Ironically, these numbers are much closer to the expected rates with random winners (the NFL and MLB would have 12.5% per team and the NHL and NBA would have 6.25%, if winning was always 50/50). This is a clear illustration of the impact of the salary caps and luxury tax introduced over the past couple of decades.