Wagglelytics.ca
CFL stats in motion. Updated through the Division Finals of 2026.
Site news · By Caleb · October 28, 2026

The EPA model now covers three seasons

The whole 2024 season is on the site now, playoffs included. It has its own tab on the Players, Teams, Games and 3rd Down pages.

How it got here

A Reddit user messaged me with a link to every 2024 game. I’d wanted to add another season for a while, and that made it possible. So thanks again for that.

I ran all 86 games through the same checks as everything else on the site. The score in every quarter matches the official box score, and so do the passing, rushing, picks and sacks. The standings line up too, ties and all. 2024 had two of them: Montreal and Calgary went 19–19, and Saskatchewan and Ottawa went 22–22.

Why it matters for EPA

EPA asks what an average offence would expect from a given down, distance and spot on the field. The model learns that from past drives. Before this, the 2026 numbers leaned on 2026 alone, about 10,700 plays. Now it learns from all three seasons, 33,280 plays in total.

That helps most in spots that don’t come up much. Think 3rd and long from your own 10. The model now has three times as many of those to learn from. Learn the stats

So what changed for 2026?

Not the rankings. Every quarterback is in the same spot he was, and every team is in the same place in the power rankings. Special teams is the one exception: Winnipeg moves past Edmonton into 1st.

The numbers themselves did go up a bit, though:

2026 EPA per dropback, before and now (starters with 450+ dropbacks)

QBDropbacksBeforeNow
MTLDavis Alexander540+0.29+0.30
BCNathan Rourke463+0.22+0.24
SSKTrevor Harris625+0.13+0.15
TORChad Kelly540+0.08+0.10
CGYVernon Adams Jr.539+0.06+0.09
EDMCody Fajardo568+0.05+0.08

Why? 2026 has been a high-scoring year. Teams are combining for about 57 points a game, up from 53 in 2025 and 52 in 2024. Measure this year’s offences against a three-year normal instead of just this year, and they look a little better. The 2025 numbers went up by about the same amount.

Update, October 31: the model now also knows which rules each season was played under (2026 took away the single on kicks that leave the end zone untouched). That means 2026 is measured against its own higher scoring again, so the 2026 numbers are back close to the before column: Alexander +0.29, Rourke +0.23, Harris +0.14, Kelly +0.08, Adams +0.07, Fajardo +0.06. The rankings still haven’t changed, and Winnipeg is still 1st in special teams. Win probability now uses all three seasons too. Numbers in this article are from October 28; the Players page has the current ones.

A first look at 2024

Trevor Harris was the best quarterback in the league by EPA in 2024, at +0.17 per dropback. Chad Kelly was next, on 317 dropbacks.

2024, top six of the nine starters with 300+ dropbacks (leader in bold)

QBDropbacksEPA per dropbackSuccess rateCompletion
SSKTrevor Harris453+0.1752%71%
TORChad Kelly317+0.1251%68%
HAMBo Levi Mitchell629+0.0849%69%
BCVernon Adams Jr.359+0.0748%65%
EDMMcLeod Bethel-Thompson483+0.0447%68%
MTLCody Fajardo420+0.0348%73%

My favourite part was seeing where some of this year’s best were two years ago. Davis Alexander was at +0.03 on 165 dropbacks in 2024. He’s at +0.30 now. Nathan Rourke was at −0.18 on 224 dropbacks, and he’s been at +0.22 and +0.24 since.

Numbers are through Week 21 of 2026. Kneels and plays wiped out by penalties are left out, and overtime isn’t part of the EPA model.