Full decade of Grey Cups ranked using win probability and GIFs

Photo courtesy: Johany Jutras/CFL.ca

By now we are all more than aware that the CFL has no Grey Cup game this weekend.

You also might be tired of the endless filler content that has been created to keep the CFL in the news this week despite the lack of its marquee event. This article may very well fall into that category of yet another ranking for you (and I’m sorry ¯\_(ツ)_/¯ ).

However, if not, then the goal of this article is to provide a unique numbers-based view on the past decade of Grey Cups.

If you follow a lot of NFL you may have become used to the idea of win probability. If not, I’ll give a quick explanation before using win probability calculations to rank last decade of Grey Cups.

Win probability

There is no agreed way to calculate this statistic for football or any sport. The general idea is to determine numerical values that represent the state of the game.

My calculations include: a kick-off/one-point conversion/two-point conversion/regular play; is it OT, distance from end zone, time remaining, down, yard to go, score differential; and a score differential relative to time remaining calculation.

The CFL does not produce this statistic themselves, unlike the NFL. As a result, the following win probabilities are the result of me downloading and processing the last decade of CFL play-by-play data to determine the statistics.

Once you have these values you determine for each combination of game states, how many games were played, and how many games were won. This gives you a win percentage.

Now this sounds simple, but things get messy from there. First of all, some game states are not very common, which means your sample size can get small. Small sample sizes lead to unreliable statistical models.

As a result, there are many methods that are used to attempt to smooth out these issues. The following win probability numbers I have produced are based on all CFL games played from 2009-2019. (There are some big limitations on the quality of CFL play-by-play data available before that point.)

I have trained something known as a Random Forest model. A Random Forest model is basically the combined result of numerous Decision Trees. A Decision Tree is a single model where you put in your inputs (What is the game right now) and the model moves through a series of questions that determine which path it reaches, at the end of the path is a win (100%) or a loss (0%). A Random Forest creates and sets up a large number of these decision trees of different shapes and combines all of their trained guesses about whether or not a team would win and produces a single win percentage.

This is only one of many ways to determine a win probability. Like any statistic of this type, win probability is not predictive, only descriptive. It tells us about past similar events, but doesn’t know the future. It also doesn’t know anything we don’t tell it, like who is hurt or what special play a coach has stuffed up his sleeve.

How To Rank Based on Win Probability?

How do I rank a team based on win probability? There are a couple of ways we can use this method to quantify a whole game, rather than just a single moment.

To explore this I am going to use a chart of the win probability from the 2009 Grey Cup. Montreal (MTL) won this game 28-27. Montreal is in red and Saskatchewan (SSK) is in green.

When a team is favoured to win its colour rises above the 50 percent middle line and is shaded. Looking at this game we can see that SSK controlled most of the game and was above 90 percent odds to win the game in the fourth quarter (Q4).

2009 Grey Cup WP

How does this chart let you judge the overall game, rather than just a teams win probability at a moment in time? In a general sense we are interested in the shape of the chart.

There are a couple of arguments a fan of football will use to argue a game was good.

  1. Did their team win (or their rival lose)?
  2. Was the game in doubt throughout?
  3. Was there a lot of scoring?
  4. Was there a big comeback?

We can attempt to cover all but the first of these with win probability and score over time.

Was the game in doubt throughout? This is known as the Game Excitement Index (GEI). Basically, by how much did the win probability change throughout the game. The more, and the more extreme, the larger this value. A chart with a lot of oscillation would have a high GEI.

Was there are a lot scoring? I call this the Game Shoutout Index (GSI). Basically, take GEI but combine it with the score in the game. An exciting game with more scoring, is an exciting classic CFL shootout.

Was there a big comeback? I call this Comeback Factor (CBF). This is a bit easier than the prior two. We look at the winning team, and then look at the winning team’s lowest odds to win the game. The lower their odds at any point, the bigger the comeback. (The challenge with this concept of comebacks is that teams that come close to a comeback and fail don’t necessarily get credit, and there is no differentiation between comebacks in Q1 vs Q4.)

Instead of reporting these as some number I’ve also added some context to compare to all games from 2009-2019.

The GEI for the 2009 Grey Cup was 9.67, which doesn’t mean a lot. But in context this degree of game excitement is the 76.4 percentile, which means it was more exciting than 76.4 percent of games from 2009-2019.

The GSI was 81.6 percent and the CBF was 98.5 percent. (Only 1.5% of games from 2009-2019 had a bigger comeback than Montreal did to win the 2009 Grey Cup.)


To determine the ranking I have used the Game Shootout Index values as this seems the most CFL of all the measures to me. A great CFL game has offence, but is also back and forth. One of the hallmarks of CFL rules is that late in a game it is hard to kill the game clock. This results in more opportunities for the opponent to win the game late.

To some degree, GEI is as valid and maybe more-so if you value defence equally as much as offence. The danger is that GEI can value games in which neither team really threatens to score, but the score is low, and within one touchdown, so this game is judged as ‘exciting.’ As you may guess, GEI is a NFL-developed method. GSI is a variant to highlight offence.

(Quick note: 2009 Grey Cup would have been ranked No. 1 by GSI if it had been included. It would have also been No. 1 is CBF and No. 4 in GEI.)

Now for the rankings!

Number 10: 2014 Grey Cup CGY 20 over HAM 16

GSI 13.6% | GEI 25.8% | CBF 25%

2013 Grey Cup WP

To start us off, we have the 2014 Grey Cup. This game was low-scoring and Calgary (CGY) had advantage from middle of the first quarter (Q1). CGY featured a first-year starting quarterback and soon-to-be star in Bo Levi Mitchell and a Hall of Fame running back Jon Cornish who fought concussion issues late in season to start this game.

CGY would lean on ball control after gaining early lead. There was a late return by Brandon Banks called back for a block in the back that might have changed things. Particularly, it would helped the game out of being in low 25th percentile for Comeback Factor.

A highlight of this game was this catch by eventual Hall of Fame slotback Nik Lewis to set-up one of Drew Tate’s two short yardage touchdowns in the game.

Many also remember this game for the (now illegal) crackback hit on punter Rob Maver that removed him from the game and forced placekicker Rene Paredes to kick to Brandon Banks late in the game.

Number 9: 2019 Grey Cup WPG 33 over HAM 12

GSI 22.3% | GEI 23.6% | CBF 23.4%

2019 Grey Cup WP

Next up is the most recent Grey Cup and another game that featured a Canadian Hall of Fame running back. Not too surprising to see one of the larger winning margins featured in ninth place.

Winnipeg (WPG) and Hamilton (HAM) appeared to be prepared to provide some excitement in the first quarter, but the rookie passer Dane Evans, with only a half season under his belt, began to see too much pressure and no options to throw.

WPG took the lead and used their strength along the offensive line to control the rest of the game. The main reason this game takes ninth versus tenth is that the total points scored was more than that in 2014.

The game is often remembered for this special teams hit by German-born global player Thiadric Hansen.

Number 8: 2013 Grey Cup SSK 45 over HAM 23

GSI 35.9% | GEI 11.0% | CBF 35.3%

2013 Grey Cup WP

What is memorable about this game? Unless you’re a Rider fan, mostly the halftime show.

The GEI index is by far the lowest in the last decade, only the sheer quantity of points scored by Saskatchewan (SSK) rescued the GSI and the only reason this isn’t ranked ninth or tenth.

This game was in control early for SSK and many likely even missed the iconic halftime show if they flipped by something else more interesting on TV.

I will always remember this as the moment where it seemed to be fate that SSK was going to win this game.

Other great highlight moments include the Henry fingers.

As well as the snap that Burris wasn’t ready for.

Number 7: 2012 Grey Cup TOR 35 over CGY 22

GSI 39.1% | GEI 22.9% | CBF 30.2%

2012 Grey Cup WP

This game barely finds itself in seventh. The game had less points by the winning team, but the score was closer — just enough to pass 2013 for interest.

Ricky Ray was traded in a shocker that year to the Argos — the team went 9-9 but hosted the East Semi-Final and upset MTL to make the the Grey Cup at home.

Ricky Ray had a fantastic moustache.

The Stampeders started Kevin Glenn at quarterback but it became clear early that Toronto defensive coordinator (DC) Chris Jones was more than prepared for his old team in the Grey Cup.

Toronto (TOR) was fined for tampering in the process that led him to end up in TOR after being the DC for CGY in 2011. Nik Lewis was in his prime in 2012 but not much else went right for the Stampeders.

This game actually featured Bo Levi Mitchell’s first Grey Cup touchdown pass in garbage time.

Number 6: 2018 Grey Cup CGY 27 over OTT 16

GSI 46.4% | GEI 57.7% | CBF 29.3%

2018 Grey Cup WP

The Ice District game. Hosting the World Cup prior to Grey Cup meant Commonwealth Stadium featured shorter-than-usual turf, and after leaving the field snow-covered leading up to the Grey Cup, the field had developed a layer of ice embedded in the rubber footing. This led to the slipperiest Grey Cup in memory without any snow on the field.

The most memorable moment of the game was the second quarter (Q2) Ottawa (OTT) touchdown to make it a three-point game being followed soon after by a Terry Williams punt return touchdown before the end of the first half.

Number 5: 2010 Grey Cup MTL 21 over SSK 18

GSI 51.8% | GEI 77.6% | CBF 58.6%

2010 Grey Cup WP

This game is actually the second-most exciting game but gets dropped to fifth in shootout index because it was lower scoring than others. If I was adjusting this ranking for personal feelings, this game would move up the list to at least No. 4. The late control of game by MTL holds it back.

MTL missed a field goal late to ice the game. SSK took possession and seemed to be driving to right the disappointment from 2009. SSK ran deep routes on the play and the protection fell apart.

Durant makes a Herculean effort to get free and attempts to throw the ball out of bounds to avoid the loss of yards but the ball ends up short and intercepted to end the game.

Number 4: 2011 Grey Cup BC 34 over WPG 23

GSI 57.9% | GEI 38.9% | CBF 13.3%

2011 Grey Cup WP

The GSI for 2011 isn’t much larger than 2010, but the higher scoring helps it out for the shootout factor. The 0-5 Lions turned around their season to be the home team at the first Grey Cup hosted in renovated BC Place.

WPG never managed to take control in this game, which is why the GEI index is only 38.9 percent. I would rank this below 2010 and even 2018 as both those games had bigger moments for the other team having a change in the game. But the total score does help this game out for fans who just wanted to see points on the board.

Memorable moments include this Kierrie Johnson strut touchdown from Travis Lulay.

As well as this final quarter Odell Willis knockdown which could have been intercepted and may have made the game interesting late.

Number 3: 2015 EDM 26 over OTT 20

GSI 64.9% | GEI 76.5% | CBF 78.2%

2015 Grey Cup WP

We’ve now entered the top three and all of the top three are highly-deserving of their rankings. At this point your arguments between games may be more taste-based for which of the three you like.

First up, based on GSI, is 2015. Chris Jones and Mike Reilly finally overcame the Stampeders on the way to the Grey Cup after being awarded the host of the West Final by tiebreak procedures.

Oddly enough, I don’t have many GIFs from this game. Partially due to it being a lower-scoring, tightly-fought game. Ottawa (OTT) had an early lead 13-0 but after that it was less than a touchdown difference for the rest of the game. OTT would punt with two minutes left and never see the ball again.

Number 2: 2017 TOR 27 over CGY 24

GSI 78.7% | GEI 81.0% | CBF 97.3%

2017 Grey Cup WP

This is the top comeback factor game of the decade, barely short of MTL in 2009 just before the decade at 97.3%.

In the surprising blizzard that developed headed into the day of the Grey Cup, the 9-9 Argos — who had just hosted the East Final and held off SSK with a late TD drive — would overcome the favoured Stampeders. The game is well-known for Shania Twain arriving on a dogsled.

Hall of Fame QB Ricky Ray would complete the longest Grey Cup TD pass early in the game and then start the third quarter with a Wilder-finished TD drive.

It took until a Q4 fumble near the goal line by Kamar Jorden returned by TOR to tie the game. That TD was the longest turnover TD in Grey Cup history.

With that life, Ricky Ray would drive for a late FG in the fourth quarter to take over the lead and forced Bo Levi Mitchell to play from behind for the first time. The Stampeders pivot would throw an interception throwing into the end zone from the Toronto 30-yard line to end the game.

Number 1: 2016 OTT 39 over CGY 33 in OT

GSI 79.0% | GEI 38.8% | CBF 19.1%

2016 Grey Cup WP

This game is a bit of a fun one to judge. Overall, this game is low in comeback factor at 19.1%, although Stampeders were yards away from making it the biggest comeback in the decade by far.

The game is also low in GEI as for much of the game OTT had a large enough margin to be in control. However, for GSI the game was high enough scoring to overcome this. A total of 72 points were scored between the teams, and there was overtime! The only OT Grey Cup game in the decade.

This game also featured the 8-9-1 Redblacks who had hosted the East Final and beaten Edmonton in the snow. Henry Burris would appear to injury his knee in warmup but return to start the game.

CGY was down in the game after not seemingly being prepared on both offensive and defensive side of the ball for the Redblacks. One of the few high notes for the team was Canadian QB Andrew Buckley scoring a touchdown.

However, the Stampeders would get back in the game and have a late chance to win or tie after recovering an onside kick.

Abdul Kanneh would save a TD on that drive with a PI call that prevented OTT from going down 4 with a under a minute left.

Stampeders would fail on short yardage and settle for a tie by a field goal and go to OT in the Grey Cup.

In OT, Burris would hit Ernest Jackson with the game-winning pass on OTT’s possession.

Bo Levi Mitchell would attempt to answer but couldn’t find the same magic with this pass falling through Kamar Jorden’s hands in the end zone.


The final ranking for GSI is 16, 17, 15, 11, 10, 18, 12, 13, 19, 14.

The final ranking for GEI is 17, 10, 15, 18, 11, 16, 14, 19, 12, 13.

The final ranking for CBF is 17, 15, 10, 13, 12, 18, 14, 19, 16, 11.

Hudson is a Ph.D. graduate and instructor in computer science at the University of Calgary. He is a fan of football in all its forms