A happy new year to one and all. We hope you had some good, safe fun last night, and that the holidays have still held some charm (but if they didn’t, that’s ok too—you don’t have to have the perfect year every year).
We’re hoping to make 2021 a big year for this website. With the help of so many of you, our growth was abundant in 2020, with our pageviews multiplying by a factor of about four and a half. We’re hoping to repeat that performance this year, if not improve upon it (have you signed up yet for our daily email newsletter?), and that’s not the only goal we have. I carry the least weight around here in terms of pageviews, so I asked Joe and Stu for their Stu Year’s Resolutions. Here they are:
NIT Stu, in his words:
- Make biscuits twice (set a goal of making them once in 2020, achieved it on December 20th)
- Get a dog. Name it Fargo. Turn Fargo into a content machine.
- Bring back MilkTime. Not that it ever left (in the spiritual sense).
- Will the NIT to happen in its full form, including a 32-team field, a final four at Madison Square Garden, myself attending that final four, and nobody (aside from the NCAA as a non-human entity that it’s therefore ok to think about killing off) fearing for their life because of the NIT.
- Wave at Shaka Smart at least one time and get Shaka Smart to wave back exactly one time (if I overshoot this I’m gonna try to get him going on the hang loose sign)
- Get Joe Kelly a Cy Young vote
Joe Stunardi, in his words:
Our models have been a big piece of this site’s growth. NIT Bracketology alone accounted for just over half our pageviews in 2020. So we need to both build upon that growth and spin out into parallel channels that can drive comparable results.
With college basketball, it’s all about getting a bracketology model that’s competitive accuracy-wise with the best straight bracketologists (i.e., folks who do brackets but don’t do probabilities like we do). Hopefully we don’t have to wait too long to get the probabilities running again, and hopefully we can dive in on those down to the team page level, where people can see what various outcomes would do to different teams’ chances, but some of that is out of my hands due to the virus. So in that arena, the priorities are 1) getting the model as accurate as possible and 2) building out the capabilities to offer a more in-depth look into the outputs of the thing, which is something that could improve fanhood for a significant sample of college basketball fans while also increasing our pageviews significantly.
With college football, the groundwork’s there. The playoff selection process isn’t that hard to model (believe us, it isn’t), and there’s the option to bring back the FCS model as well. We’ll probably have to build our own rating system for individual games: There isn’t good back-data on FPI or FEI, and Bill Connelly’s SP+ changes with some frequency and is a little too subjective for us to use reliably as the backbone of a model. We could also incorporate Las Vegas odds, though, which are probably the most consistently predictive metric of individual game outcomes, so perhaps we can figure out a way to balance those things out. The bottom line is that we want to build this into a major traffic-driver, and to do that we want to build it into something that provides legitimately interesting insights to a broad cross-section of fans. We want to do full bowl projections, not just playoff projections. We want to give fans a channel through which they can see the outcomes in various scenarios. If we can’t get the team-level pages up during college basketball season, we should really focus on them over the summer so they can be an option for college football. That’s going to require some major advances in the area of technological know-how, but hopefully it isn’t unattainable within eight months.
We’re also looking into modeling the NHL and NASCAR, two relatively major sports that go unmodeled by FiveThirtyEight, an organization that, obviously, we hold great respect for and are always emulating to some degree. Those seasons each start in the next few weeks, so they’re an immediate focus alongside basketball, and we’re keeping an eye on the golf and tennis calendars.
Finally, we’ve spoken before about building out models for more college sports: hockey, volleyball, softball, lacrosse, etc. That’s still on the radar, but with the seasons so odd this year due to the pandemic we probably won’t add any models until fall at the earliest for collegiate sports. We’ll see what happens with baseball and softball, but I’d guess at this point that professional golf is more likely than college baseball or softball.
And with all of this, I’d like to really shore up our transparency and how statistically sound we are. We aren’t unsound statistically, but improving that area will improve the models, and both of these things seem important from a legitimacy perspective. While you can throw a bunch of shit out there and get people to interact with it (something a certain media organization did this college football season), we’d rather do what we do in a responsible manner that tells people exactly how confident they should be in our projections. Basically, we want to be humble so that we can earn some pride. Depending how this weekend goes, I should have a post up tomorrow about some of this. I’m hoping to make it a weekly focus, for the sake of keeping myself accountable both to accountability and to growth.
Outside of models, I’d like to figure out the best way to blog about Iowa State and the Cubs, and whether they’re worth blogging about. Our Cubs blogging has driven some traffic before, but that market’s saturated to the level that it’s hard to be a tangential Cubs blogger and still have a following. There’s a long discussion about team-focused blogging that could be had here, but to keep this from going too long, the general focus is trying to figure out the proper level of engagement with specific teams, and whether that’s something that belongs on The Barking Crow (out of me—I think Stu’s Texas and Burnley blogging makes a lot of sense regardless of what makes sense for my blogging).
And lastly, the published bets. They’re close to falling into the red, and as I’ve often told myself, there’s no reason to publish these if you’re publishing losers. I’d like to keep them going, but we’ll see. At the moment, I’m hoping to get the college basketball model good enough that it can be used to build a futures portfolio. From there, I’m hopeful that NCAA Tournament and NIT daily bets can lift us in March, and that MLB daily bets can lift us from April through July. We have a good track record in all of those months with all of those bets.
We’ll see where it goes. There’s plenty of space to grow, and in a way, it’s nice that the biggest constraint is time.
