A New Season and a New Team
As far as I’m concerned, 2022 could not be more welcomed. I have much to cover here, some 2021 stuff to wrap up, and some changes to explain, but all in all I am so very excited to be back to this! A quick recap, in case you’re just joining, this idea did not come to me overnight, but has been a work in progress over the last decade plus. What started as a fun with spreadsheets kind of a thing, eventually became enriched with ever evolving algorithms and my going on and on about parabolas and coaching changes and the late night game between Wyoming and Hawaii with my friends. I’m super fun to hang out with! Eventually, I started to not just win, but consistently at about the same rate every year, 56-58%, then with some well tested changes, up to 58-62% for a few seasons. I knew this could be extremely profitable for anyone actually betting heavily on it, but what I never realized was that professional gamblers don’t generally average much better than that. I had been mistaken in thinking that I needed to win like 75% of my games to have accomplished anything with this, and once I realized my error in judgment, I’ve felt ever since that I’ve been on to something.
Now, I’ve tried to be measured about the effects of Covid on my life versus others. It was scary professionally, but my job survived, even began to thrive in the new environment. Health wise, I’ve never, to the best of my knowledge, actually caught the damn thing. I probably have and didn’t know it, but I’ve kept up with the vaccines and may even be among the lucky asymptomatic types in the first place, so given the damage it’s caused worldwide, I don’t have much to complain about. But if there is collateral damage in my life from all of this, the butterfly effect in full motion, it’s most noticeable with this little project. Covid took the system down. The disaster season of 2020 dropped me to a rather depressing 48.4%, but had rebounded in the second half to even get that number up to where it was, perhaps nothing more than regression to the mean, but it certainly looked as if the system was working the way it’s supposed to, adjusting to the new environment. The problem was that the season soured the data, data which, good or bad, the system cannot even operate without. So the system began 2021 just as it had in the previous season, by losing, but with more games to lose this time! I had targeted 2020 as the time to start writing about this, unfortunate timing to be sure, and by the end of the first month of last season I was seeing very few signs of getting back on track, so I shut it down. I was still tracking of course, but I don’t know everyone that might be following this and wouldn’t want people betting on what appeared to be no better odds than a coin flip, at best. Plus losing makes it difficult to continuously write about, as I’m sure you can imagine.
As 2021 rolled on, the system once again adjusted, this time with better results than the progress shown the year before. There was even a 12-2 week in there, and by the time the bowl games and championship wrapped up, I was sitting at 51.4%, a little better than a coin flip. Now, if you have ever placed a bet on point spreads, you know there is juice, a 10% take for the house every time you lose a game, and 51% does not make up for that. So let’s say you bet $200 on every game the system told you take. In 2019 you would have turned a profit of $4820, nice but also about what you had become accustomed to if you had been betting the system for the 4-5 seasons before it. Then in 2020, you do the same and, ouch you lose $2120, and then last year, even with the winning record, you would have lost $700 after the juice. Remember, the sample size is large here, very important to insure statistical significance, but definitely not good if the juice is taking a sizeable toll. A savvy gambler might have stopped betting as the system was flailing around, showing little progress, and then resumed once it was clearly demonstrating some consistency, in which case a profit wasn’t really that far off on the horizon, but for the most part, 2021 and it’s soured data from the debacle of a season before it, only served to offer optimism if you consider that we begin 2022 with the soured data one more year removed from us.
And of course, there is a twist in this narrative. As I’ve mentioned before, I’m always testing something, most of the time only to throw concepts out or rethink them. I try the best I can to stick to the scientific method, demanding that subjective thought get hammered away at by objective requirements, but always looking to improve. As I’ve mentioned, the success of my system was dependent on putting the teams to scale based on productivity, with outliers on the top and bottom, gradually building to the bulk of the teams somewhere in the middle, a perfect parabola. To find value versus the spreads, essentially the math blacks out all games that fall underneath the resulting graph, and says “go” when the point differential exceeds what my expectations are. But as I’ve done this, there have been things that troubled me. One, it’s definitely favorite heavy, which is fine, but at what point would an underdog be getting so many points that there is actually value in the team that might not be the desired team to bet on? Two, occasionally betting against a really good team or betting on a really bad team seemed somewhat apathetic, simply because of the opponent, which is how gambling usually works, but shouldn’t more consideration be given to outliers due to their potential or lack there of? Three, at times when one outlier at the top is playing another at the bottom, the point spread can be so high that a coach might lay off the gas well before it ever gets there. Should Alabama beat New Mexico State by 65 points? Well sure, but that doesn’t mean that Nick Saban actually wants to do that, plus the margin for error is such that one pick six by NMSU throws the chances into extreme improbability land. Four, how close to the margin is too close, and for what teams? Can there be a line where I should consider one bet better than another? Well, my choice of the word “line” there was specifically chosen, because once again linear thinking provided no luck until I set out to design another scale, simply put, a second parabola that addresses all 4 of these concerns.
I’m only offering concepts here and not precise numbers as I would like to maintain some ownership of my efforts, but envision one parabola directly above another on a graph. Now black out the space in between the two and call it the no bet zone, thus giving you 2 numbers instead of the one I was using before. That’s simplistic, and not precisely how the math works once matchups and betting lines are introduced, but regardless, instead of saying this is the team I want to bet on in this matchup and how much I’d be willing to give up to take them, I now have a number I’d be willing to give up and also a number I’d be willing to see value in the team I don’t even like. There is always too much and too little, such is life. Also, if you draw out the 2 parabolas, you’ll notice that the space in between narrows at the top and then widens as it approaches the outliers, thus addressing some of the concerns I had with my good team/bad team scenarios. I call it the second scale.
So does the second scale work and if so, why haven’t I been using it already? Well I’ve certainly been testing it for long enough and the addition of more underdogs, fewer favorites, resulted in a slightly lesser sample size, albeit nothing significant, but it was always lingering just below my numbers without it. For example, in 2019 I finished at 58.42%, up $4820 if one had put $200 on every game, and the second scale produced 58.37%, up $4660 had the same amount been placed on every game. And the preceding years had gone about the same, with only one season favoring the second scale, but always very close. So why make the change? Well, the tides turned in the second scale’s favor post Covid. In the seemingly impossible to win year of 2020, the second scale went 66-58, and if placing $200 on every game, a rather modest winning percentage of 52.23% would have been enough to overcome the juice and provide a $440 profit, the only bright spot in every concept I was testing for that season. And last year was slightly better, this time at 53.42%, with a larger sample size than the shortened season before it, now offering any $200 player a $480 profit. That’s a lot of effort for such modest gains, but at least it’s a winner in an environment otherwise providing me a whole lot of losing. If you’re going to put money on something, better to go with the one that remains undefeated, even against Covid, and since the differential between the two was never that great in the first place, I think it’s a no brainer; this year I’m adding the second scale to my system of record.
So that’s it then, no other concerns? Wrong. College football has all kinds of curveballs to consider. One is the transfer portal and now with money getting tossed around, officially as opposed to allegedly, the opportunity for upward and downward mobility seems greater than ever before. This could be a problem leading into every season from here out. Remember that, as you consider putting your family inheritance on one group of 20 year old boys to beat another in their 1st month back to school. But it could also prove useful because early season spreads are defined by people anticipating said mobility, and if programs remain even close to more consistent than anticipated across the full spectrum of all FBS teams, the system will be here to collect on it. Call me cautiously optimistic.
The only other consideration before I move onto my picks is a new addition to FBS football, as James Madison begins their journey into the big time. I produce zero data on FCS teams, so just like I did before with Coastal Carolina and Liberty, along with the teams that sat out 2020 all together, I simply put the new coach tag on them, giving me 2 years for them to settle into my scales before I’m ready to suggest any plays on one of their games. A starting point is the hard part, as it’s totally arbitrary from where I’m sitting, so I figure Vegas is the best resource I have available here, even if my goal is to eventually find value on James Madison Dukes games, against the spreads. This weekend, James Madison is a 5.5 point favorite at home against Middle Tennessee, so I’ll use Middle Tennessee as my launching point, and start the Dukes slightly above them on my scales, and then do the same with Sam Houston State and Jacksonville State with their respective first games when they join next season.
OK then! I think I’ve covered what I need to with my first post in awhile and it’s time to move onto my first picks of the new season! I actually had two already last night, as I mentioned on Facebook and Twitter, and I’m off to a 1-1 start. Thanks for nothing, Purdue. Here are my picks:
Thursday
Central Michigan +21.5 at Oklahoma State WINNER
Purdue +3.5 at home vs Penn State LOSER
Friday
Western Michigan +23 Road at Michigan State
Saturday
Appalachian State +1.5 at home vs North Carolina
East Carolina +11.5 at home vs North Carolina State
South Florida +12 at home vs BYU
Memphis +16 at Mississippi State
Louisville -4.5 at Syracuse
Monday
Clemson -22 at Georgia Tech
I acknowledge that you may not always be able to get the same spread I locked into, but that could go the other way as well and you could find a better line from time to time. For this week, there is one particularly close consideration, as the system’s second scale eliminates Louisville from playable if that spread gets to -5. The others look solidly locked in as value picks, even with a little movement. I’m also always tracking close call spreads as they approach game day, and I’ll post any newly likeable picks on social media should they happen late in the process. The closest this week was actually Pittsburgh last night, which I was waiting for a -6.5 line that never happened. Wish it had! UCLA and Houston are other possibilities this week, but meh, I doubt we see enough coming there. Happy football, it’s good to be back!
