Trevor Hardy
CFL.ca
What do Leif Thorsen, Ben Wearing, Farwan Zabedi and Lukas Shaver have in common? Well, they were each selected in the 2001 CFL Canadian Draft. More than that, they were each selected prior to three-time CFL All-Star and Grey Cup Champion Kevin Eiben.
Eiben was selected in the fourth round of the 2001 CFL Canadian Draft, 26th overall. This year, the annual CFL Canadian Draft will take place on May 8, 2011. On this day, 48 Players will be selected across six rounds. This upcoming draft got me wondering: how likely is it that a Club will find a “Kevin Eiben” so late in the draft? And does finding a “hidden gem” so late in the draft translate to success on the field?
Quantifying Success at the Draft Table
There are a number of different ways to define a “successful draft”. For the purposes of this analysis, I have decided to quantify a Club’s success in the Canadian Draft in the following ways:
1. the number of Players selected in the Canadian Draft during the period 2001 to 2010 that made a Club’s roster for at least one game;
2. the percentage of all games played during the period 2001 to 2010 by the drafted Player. For example, a Player drafted in 2001, who played the entire season in 2001 but never played another game would have played 10% of all games played (18 games ÷ 10 years x 18 games); and
3. the number of times a Player selected in the Canadian Draft during the period 2001 to 2010 has been selected an All-Star.
Making a Club’s Roster
A Player selected in the first round is approximately three times more likely to make a Club’s roster than a Player selected in the sixth round:
| Round | # Selections | # on Roster | % on Roster |
| 1 | 83 | 73 | 87.95% |
| 2 | 84 | 69 | 82.14% |
| 3 | 83 | 58 | 69.88% |
| 4 | 84 | 50 | 59.52% |
| 5 | 83 | 45 | 54.22% |
| 6 | 83 | 26 | 31.33% |
| Total | 500 | 321 | 64.20% |
Intuitively it makes sense that the likelihood of making a roster decreases the further down a Player is drafted.
Using this measurement, Calgary has had the most success of all Clubs with approximately 74% of its draft picks making a roster at some point. Ottawa had the least success in this regard, with approximately 54% of Players it drafted making a roster:
| Rd | BC | CGY | EDM | HAM | MTL | OTT | SSK | TOR | WPG |
| 1 | 88.24% | 93.33% | 71.43% | 90.91% | 90.00% | 100.00% | 100.00% | 66.67% | 100.00% |
| 2 | 66.67% | 92.86% | 100.00% | 88.89% | 75.00% | 28.57% | 87.50% | 100.00% | 100.00% |
| 3 | 85.71% | 60.00% | 72.73% | 60.00% | 66.67% | 25.00% | 88.89% | 75.00% | 66.67% |
| 4 | 100.00% | 72.73% | 66.67% | 50.00% | 50.00% | 50.00% | 25.00% | 61.54% | 60.00% |
| 5 | 40.00% | 66.67% | 50.00% | 75.00% | 50.00% | 75.00% | 30.00% | 50.00% | 55.56% |
| 6 | 0.00% | 40.00% | 14.29% | 11.11% | 35.71% | 50.00% | 16.67% | 56.25% | 37.50% |
| Tot | 66.67% | 73.61% | 65.38% | 64.41% | 60.81% | 53.85% | 59.18% | 64.52% | 60.00% |
Percentage of All Games Played
This metric will recognize the difference between a Player who cracks the roster just once versus a grizzled veteran of many battles.
The evidence suggests that Players selected in earlier rounds have longer careers. For example, a Player selected in the first round, on average, plays in about four times as many games as a Player selected in the sixth round:
| Rd | # Games Played | % of all Games Played |
| 1 | 4342 | 52.96% |
| 2 | 3439 | 44.53% |
| 3 | 2287 | 32.31% |
| 4 | 2008 | 23.11% |
| 5 | 1489 | 19.40% |
| 6 | 1039 | 14.05% |
Using this measurement, Winnipeg has had the greatest success (98.15% of all games played) with Players selected in the first round of the Canadian Draft (although, in fairness to the other Clubs, Winnipeg only had one first round draft pick – Brendan Labatte – in the period 2001 to 2010), whereas Edmonton (26.74%) has had the least success:
| Rd | BC | CGY | EDM | HAM | MTL | OTT | SSK | TOR | WPG |
| 1 | 58.26% | 67.42% | 26.74% | 55.86% | 53.94% | 50.36% | 51.80% | 32.03% | 98.15% |
| 2 | 44.38% | 46.98% | 57.80% | 38.22% | 38.83% | 15.43% | 44.80% | 59.51% | 65.74% |
| 3 | 34.16% | 18.98% | 38.67% | 25.37% | 33.30% | 17.88% | 48.69% | 40.97% | 29.08% |
| 4 | 46.70% | 11.07% | 23.91% | 26.80% | 16.97% | 13.89% | 15.07% | 28.45% | 27.34% |
| 5 | 9.82% | 21.33% | 4.38% | 27.83% | 34.82% | 49.21% | 4.08% | 25.57% | 14.88% |
| 6 | 0.00% | 15.56% | 12.70% | 0.46% | 21.07% | 2.84% | 11.20% | 23.20% | 21.59% |
Number of All-Star Selections
One final measurement we’ll consider is the number of times a Player drafted was named a Divisional or CFL All-Star.
During the period 2001 to 2010, there were 32 occasions when a Player drafted during the corresponding period had been named a Divisional All-Star, and nine occasions when he was named a CFL All-Star. Using this measurement, Edmonton has had the least success, and Toronto the most – in no small part due to Eiben’s prowess on the football field:
|
|
Does Success at the Draft Table Lead to Success on the Field?
To answer this question, I decided to test all of the Clubs’ performances at the draft table – using the metrics discussed above – against all of the Clubs’ performances on the football field (as measured by their winning percentages) during the period 2001 to 2010.
To help set this up, I’ve calculated each Club’s winning percentages during the period 2001 to 2010 as follows:
| Team | Winning % |
| BC | 60.28% |
| CGY | 48.89% |
| EDM | 51.94% |
| HAM | 34.17% |
| MTL | 63.89% |
| OTT | 31.94% |
| SSK | 53.61% |
| TOR | 48.31% |
| WPG | 48.61% |
I then calculated the “correlation” between these winning percentages and each of our metrics discussed above. It would be very difficult to describe how a correlation coefficient is calculated in this space. It’s quite complicatedand I rely on a spreadsheet program, statistical software or calculator to perform the legwork for me.
A correlation is measured between negative 1 and positive 1. A correlation very close to positive one is said to be “highly positively correlated”, whereas a correlation very close to negative one is said to be “highly negatively correlated”. A correlation of zero indicates no correlation whatsoever.
What correlation is measuring is the degree to which two variables act in the same way. For example, if you’re the President of a football team you may be interested in looking at the relationship between the game time temperature and the sale of lemonade. We would expect that there would be a very high positive correlation between these two variables (as the temperature goes up, the sale of lemonade goes up). If you decide instead to test the game time temperature to the sale of fleece hoodies, you might expect a very high negative correlation (as the temperature goes up, the sale of fleece hoodies goes down).
Getting back to the task at hand, before I started this analysis, I expected a very high correlation between a Club’s winning percentage and its draft success. However, what I found was that there were no definitive results on any rounds except for the middle two rounds. What I found in rounds three and four was that there was a very high positive correlation between winning percentage and the number of selections in the third and fourth rounds, as well as a high positive correlation between winning percentage and the number of drafted Players who made a Club’s roster.
Conclusion
Successful CFL Clubs have been built any number of ways: through free agency, or trades, or negotiation list rights. But most CFL General Managers will agree that strong Canadian content is key to a winning team. And, obviously, the Canadian Draft is integral in accumulating that Canadian content.
Based on my analysis, Clubs can most greatly impact their probability of success on the football field by being especially prudent with their selections in the middle rounds. And who knows – come the fourth round on May 8, 2011, a Club may select the next Kevin Eiben!
