It’s seven weeks into the 2023 season and the CFL and Genius Sports are still struggling to launch the league’s new stats system. As problems persist with both live game and historical statistics, fans have been left longing for the reliability of the past, while asking how this could have been so badly flubbed in-season.
Speaking to the media last month, commissioner Randy Ambrosie told reporters that the system rollout was “not a failure,” saying that some issues are to be expected during a “complete technology overhaul.” Those comments were later called into question by the CFL’s former software partner, New Jersey-based firm SharpHat, in a widely-shared tweet, which claimed their rollout of the previous stat system in 2016 had been “flawless.”
The tweet caused a stir in the CFL community, generating nearly 21,000 views for the corporate Twitter account with just 837 followers. The response surprised SharpHat executives, who were unaware of the league’s ongoing stats problem.
“We understand the CFL’s business decision to go with Genius’ software but we disagree with the way the previous stat system was described in their statements,” the company said in an official release.
“Specifically, the stat system we wrote does not use an ‘old closed architecture’ that was not upgradable. It was built with modern technologies and approaches for the time in which it was written (2014-2015). It is remarkably extensible and upgradable and would lend itself very well to adding new data and features.”
In an interview with 3DownNation this week, SharpHat president and CEO Nick Maou praised their previous working relationship with the CFL. He stressed that his response to Ambrosie’s comments was not meant as an attack on the commissioner or the league, but simply an attempt to set the record straight on a piece of software about which his team remains extremely proud.
“I tweeted that it was flawless because it was. Honestly, in seven years, we never got a call ever that something wasn’t working, during games or anything,” Maou said. “Very minor things like maybe they didn’t set up officials properly at the beginning of the season or stupid things like that, but we’ve basically been on a support contract with them for the last few years and we’ve hardly even had to use it, to be honest with you.”
When addressing reporters in June, Ambrosie suggested changing the previous stats system was a top priority following their partnership with Genius, describing it as “almost an analog type of system.”
“We can’t do what we want to do with that platform,” he later said. “We couldn’t upgrade that platform.”
That simply isn’t true according to Cy Lee, SharpHat’s chief technology officer.
“During the seven seasons in which we were in production with them, they did have occasional small tweaks that they made, certain statistics that they wanted to extend or change. We did do small changes like that for them,” he explained. “I think that the amount of change was more limited by budget than it was by technology and so we really feel those statements were mischaracterizing the nature of our software.”
SharpHat was first contracted by the CFL to custom design a new stats software in 2014, with the system going public ahead of the 2016 season. As with all projects the company undertakes, that software was then fully owned and operated by the CFL after being specifically tailored to their needs.
Maou and his team were made aware in mid-January that the league was moving on from their software, though they suspected the move was coming as soon as Genius Sports struck an equity partnership in December 2021. This was not a major change given that their role had been strictly on an as-needed basis following the initial launch.
While the company advertises their work with the CFL on their website, many fans learned of SharpHat’s existence for the first time due to the social media exchange. That’s a testament to how seamless the transition was the last time the league changed its stats system, a monumental undertaking that mostly flew under the radar.
“You’re talking about over 6,000 person-hours of development. That’s with the initial discovery part, getting the requirements of the system, the development, the testing, and then supporting that first season,” Maou recalled. “It was a lot and we had a good six-person team on this project. It was a huge investment from (the CFL) and they wanted to buy into their own system that they owned and that could grow for years.”
SharpHat’s software was designed for stats crews to input the raw data of any given play, and then calculate the desired statistics. This could be altered and scaled yearly to the league’s preferences, with the addition of new input fields creating new statistics like yards after catch.
However, before they could generate any statistics, the designers had to first grasp the nuances of the Canadian rulebook and prepare for any eventuality. That involved countless hours in conversation with CFL head statistician Steve Daniel discussing potential outcomes and diving into old gamebooks. Even something as infrequently used as the ability to punt after a pass had to be accounted for, so it would not cause errors in the software as it would for an American system.
According to sources, a key factor in the CFL’s current stats struggles is that Genius underestimated that very challenge, not realizing how much alteration would be needed to their pre-built software to accommodate Canadian rules. That’s a challenge that Lee sympathizes with.
“When we did the CFL setup, we had the benefit of writing it from scratch. The hard part for us was just understanding the differences at first,” he noted.
“I think that taking a system like Genius’, which is probably focused on American football rules for other leagues, there would be challenges in adapting to the CFL rules. Just the three downs, that’s a pretty significant difference. As with anything, it’s doable. It just is a matter of getting it right and testing the crap out of it, identifying any issues and then fixing them.”
For SharpHat, that portion of the process came long before fans were seeing any of their statistics. Throughout the 2015 CFL season, developers ran their new software in parallel to the pre-existing stats system, comparing results to ensure functionality and accuracy.
It was only after a full year of data was collected that they felt comfortable taking the product public. That cautious approach avoided the high-profile errors that the league and Genius have experienced this time around.
“We knew that you have to pilot like crazy. You have to enter tons of games. You have to test real-world situations, in real-time enter games like crazy, make sure the software can keep up, make sure the stats are accurate,” Maou said. “That whole process, they probably underestimated that part of it.”
Last month, Ambrosie said that the transition wasn’t left to the last minute, with “a lot of testing” taking place during the offseason. Nevertheless, the league opted not to run their existing system simultaneously with the new one due to the limited number of available stats crews. He did not offer an explanation as to why Genius, which controls a 10 percent stake in the league’s revenue arm, CFL Ventures, was not performing in-game testing itself, as SharpHat did in 2015.
If asked, Maou said he would have been happy to provide advice and best practices for the transition to their former partner and Genius. Most importantly, he would have strongly recommended running parallel systems in the early stages based on past experience.
Neither the CFL nor Genius Sports responded to interview requests for this article. The league’s data and technology partner has yet to comment publicly on the rollout of the new system.
Meanwhile, the goalposts for a return to normalcy continue to move. The league originally promised to have issues fixed in time for the regular season, then mid-June, then early July, with moderate improvements doing little to calm the criticism lobbed in their direction.
As SharpHat proved less than a decade ago, a smooth technological transition is possible. It simply required time, proper investment, and forethought.
