By the numbers: defence does not necessarily win championships

There is something about the love for defence when it comes to football. When your favourite team’s defenders take the field, it allows and even inspires you to stand, yell and scream, bang on drums and ring cowbells. There is no doubt that defence elicits and unleashes some of our most emotional behaviours when it comes to our fan experience in football.

Yet as we have recently explored and seen in recent discussions, the colloquial adage that “defence wins championships” is in fact not accurate as it applies to the Canadian Football League.


Untitled 2

The first graphic separates all the Grey Cup winners since 1980 and whether they were the regular season leaders in offence, defence, both categories, and lastly neither area. Applying raw or absolute numbers to this illustration, simply multiply by the number of Grey Cups played since 1980, which is equal to 36. This translates to the regular season league leading team in terms of offence having won 2 more Grey Cups than the regular season leader in defence.

The second graphic then looks at the actual game matchup or comparison of the two title game participants. Things are a little more evenly split here as you can easily see for yourself. In 18 of the last 36 Grey Cups, the team that emerged victorious had BOTH the better Offensive and Defensive team during the regular season coming into the game. And exactly six times each has the winning team had either the better offence or better Defensive Team. Which is also fairly interesting when you compare that to the same number of times, six once again, on which there has been an upset where the team that led in neither claimed the nation’s ultimate professional football prize.

What makes a good or great defence?

In the original research I measured defence simply by points against. This is a fairly obvious observation and way of measurement since the less points you allow your opponent to score, the better your defence and thus the better your chance at winning, as long as you score more points than them of course. (We will explore that statement further at a later date of course!)

In order to look at what specifically makes a defence good and great, we will have to pay a visit and delve into the world of our old friend; correlation of coefficients. As a quick review, correlations are a measure of the strength and direction of the linear relationship between two or more variables. Not only is the direction important, the range or strength of that direction is perhaps even more important.

Untitled 3

The following are the 10 factors that I looked at in order to gauge which are most important when measuring defences. I have also included a quick key as far are what specifically dictates the calculation as well as what outcome is most desired.

Untitled 4

Without further adieu, the following are all the correlation coefficients for all 10 defensive factors once normalized. The bar graph illustrates the total correlation coefficients for the 11-year period. The second chart details the correlation in each individual year. And finally the third chart shows the percentage variance for each individual year versus the 11-year total or aggregate correlation.

Untitled 5

Untitled 6

Key Takeaways:

  • defensive points allowed is the most influential factor. It has a higher, or more negative correlation that yards against. This would confirm the familiar phrase and strategy of “bend but don’t break” defence.
  • Giving up touchdowns is obviously worse than giving up field goals. Again underscoring the “bend don’t break” parable.
  • Pass defence is more important than run defence with yards per pass attempt being more important than total passing yards surrendered.
  • Interceptions are slightly more important than quarterback sacks, yet both are only moderately correlated to winning and losing in the CFL.
  • The standard deviation over the years is narrowest when it comes to points allowed and touchdown’s allowed which allows us to consider it a more reliable result over other factors such as total yards against or quarterback sacks and interceptions.
  • One very interesting outlier as you can see quite clearly is the 2010 regular season. The measurements are often either quite lower if not actually inverted versus the totals for every other year. This makes sense when you go back and look at that season in particular. The 2010 average winning score was 36.48 and represents the highest it has been since the early to mid 1990s.

While all these numbers are interesting at their base level, taking a look at the variances between the factors will give us a better indication as to their importance and overall significance when it comes to the “true” measure of defence.

Untitled 7

The above chart calculates the percentage variance from each factor to that of the overall most important factor: Defensive Points Allowed. To calculate differences between any two factors, simply subtract one from the other.

Key Takeaways:

  • Interceptions and quarterback sacks have 43 and 46 per cent less importance as far as correlation to points allowed.
  • Yards per pass attempt are 27 per cent more impactful than yards per rush when it comes to a corollary impact to winning and losing in the Canadian Football League. This would make sense since the CFL possesses a passing oriented game and would fly in the face of those that say you need to stop the run in the CFL. However, this will require closer examination, since a study upon down and distance would be much more illuminating.

Team by Team breakdown

Now we will take a quick look at each team’s individual correlations in terms of what has been most important and influential to their specific defence during the 11-year period from 2005 through 2015.

Untitled 8

Untitled 9

British Columbia: It’s All About The Sack

The Leos possess the most significant correlation effect when it comes to sacking the oppositions’ quarterbacks. This is not surprising when you consider past players like Brent Johnson, Cameron Wake and Rickey Foley patrolling the defensive front along with current line-backing stalwarts like Elimimian and Bighill continuing to exert pressure on opposing pivots. The interesting thing about the Lions’ defensive numbers is that the amount of touchdowns allowed has not had as much correlation to their wins and losses as field goals allowed. Quite simply this means that while they’ve been good at keeping opponents out of the end zone, they have given up too many field goals.

Calgary: Points and Pickoffs

Surprisingly, the Stampeders’ wins and losses are more sensitive to points allowed than most every other team in the league. This came as a surprise to me since we often think of the fellas with the horse on the side of their helmets as an offensive juggernaut. What this is telling me that it is their defence that is much more important to their results in the win and loss columns than what most would suspect. When it comes to what makes the Stamps defence work, their numbers indicate a high correlation and reliance upon interceptions. Quite simply, when the Stamps are picking off the ball, they are winning games.

Edmonton: Yards Per

Edmonton has struggled offensively during much of the 11-year period analyzed so it should not be a surprise to many that limiting their opponents points wise is of critical importance to the Eskimos when it comes to winning and losing. While we may think of the Green and Gold as being an aggressive and opportunistic defence that sacks the quarterback and takes the ball away, it is actually much more mundane and simple when it comes to the Esks’ defence. Yards per pass attempt and yards per rush have been the two most important factors when it comes to their defensive signature.

Hamilton: Sacks vs. Picks

The Tiger-Cats enjoy the second most significant correlation when it comes to sacking the quarterback only to the Lions. When Hamilton gets to and takes down the opposing QB, their winning ways increase significantly. What was surprising though was that interceptions is of little importance to the Ticats success in the win column. In fact the correlation of interceptions to winning runs counter to what would be expected. I would imagine this to change going forward especially if the boys in Black and Gold can continue their 2015 results in the “picks” category whilst continuing to win.

Montreal: Getting It Done Differently

Quite frankly, analyzing the Alouettes’ defence was very difficult. Their numbers are different than most all other teams as well as overall league correlations in almost every category. This leads me to conclude that Montreal is getting it done in different ways from season to season as well as differently that most all other teams in the league. For instance, when it comes to opponent scoring, it appears that the Als’ achilles heel comes from the front part of their opponents’ foot more than them reaching the end zone. Yet while their interception numbers compared to other teams are not any higher, it is very clear that picking the ball off is important to their chances of winning much more than other teams.

Saskatchewan: Yardage, Yardage, and Yardage

The Riders give up quite a bit of yardage out on the prairies, but somewhat surprisingly, this doesn’t have as much of an impact to the number of wins they’ve accumulated over the years. Yet the type of yardage conceded is worth noting. While being stout against the run, this has not translated into anything significant in the win and loss columns. On the opposite side of this are their somewhat troubling pass defence numbers. This has a much more significant impact upon their successes. Quite simply, when the Green and White are watching the ball fly over their heads, it translates to a tough offseason in the Queen City.

Toronto: Bend Don’t Break

Giving up yardage is not much of a concern for the Double Blue. They have one of the more insignificant correlations when it comes to yards yielded to winning and losing. This is the perfect example of a “bend but don’t break” philosophy. This is not really surprising when you go through the defensive coordinators that have patrolled the Argos’ sidelines for much of the last 11 years with Rich Stubler and Chris Jones. Outside of that though, when Toronto has difficulty stopping the run, their success as far as winning percentage suffers.

Winnipeg: QB’s Padding Their Stats

The most impactful aspects to the Bomber defence are in the areas of opponents’ pass completion percentage and total passing yards. Both of these had very strong correlations to their winning and losing percentages. And when you consider their record over the last 11 years, this has been troubling to say the least. It came as a surprise to me though that a team thought of having had so many excellent pass rushers that sacking the quarterback actually had very little correlation to the Bombers’ results. The same can be said for interceptions, as there was a very low correlation to winning and losing for the Blue and Gold when they pick off the ball.

As always I hope you have enjoyed our little project and I welcome any and all questions, comments and insights as we continue to look at the CFL By The Numbers. Next week we will look at the other side of the ball and take a look at what makes offences tick.

Must Read