The Scale
A few Thanksgivings ago, I was invited by my childhood best friend to play touch football, and I can remember joking before hand that I would probably pull a hamstring. Then on the first play, and no kidding, it was the very first play, I actually did pull a hamstring. I limped for a few weeks, no big deal, but it was clear to me as I hobbled around my everyday life that my athletic prime was behind me. Now, what my athletic prime ever entailed in the first place is not the kind of thing that a guy revels in the past about, but I had my moments. I could hit a baseball and I could shoot a basketball, but my deficiencies were also always ever present, namely no arm, particularly for a guy most frequently stationed at third base, and I was a complete disaster in the realm of distance running. To put these abilities on a scale of human performance, you’d likely find me well above average as a basketball shooter, slightly above average as a hitter, below average on throwing things with velocity, and as for distance running, how far near the bottom would depend on how severely handicapped the people you are comparing me to were.
Yet of all the relatively modest successes or failures of my youth, it is one that was perhaps most benign of all that was most memorable, at least in its application to some activity I might be participating in now, like say, this. It was gym class in 8th grade, maybe 7th, but 8th as I remember it, and our task at the time was to run around the very big yard at my middle school. Now, I’ve told you about my distance running woes, and they are real, but what I could do from time to time was get out fast, ahead of the pack, which is exactly what I did that day. I remember a kid, his name I think I know, but I might not even have his name right at this point and it could have been another kid altogether, so I’ll just exclude it from the conversation. Anyway, this kid was right behind me, second place, and he antagonistically kept repeating, “you’re out too fast, Burkett, you can’t keep this up.” He was correct. It would end up being a first to nearly worst story, but rather than the humiliation of the defeat, it was what I witnessed in the process that has been branded in my memory for 3 decades now. As my fade began, first it was the mouthy kid that passed me, then another. The fade was so harsh by that point, I was just an obstacle to go around, but instead of doing it one by one, it was more like, 1, then 1, then 2, then 1, then 3, then 5, then 12, 10, 9, 7, 8, 4, 3, 2, 1, 2 1 and by the time I crossed the finish line, there were only a couple kids still behind me. Basically, the closer to the middle of the group one got, they were running in packs and I got to see them blow past me accordingly, with the kids in the front, which I couldn’t really see anymore around the cavalry charge and all, and the kids like me in the back, we both were outliers, some far better than the mass and some far worse, but most were closer to average than the rest of us.
So who cares about an 8th grade gym class, right? Well, that’s true in that my defeat at the time probably took less than an hour to get over, but it resonated in another way. What that moment did for my life was begin a perspective that I would apply to every aspect of human performance, a scale to surmise about every profession, every ability, every version of aptitude or ineptitude that we maintain, at every level, and the psychology of it can be useful, or a detriment, when one considers that how good they are at something is completely dependent on who they are comparing themselves to. And so eventually, I applied this concept to college football, but it took a bit.
When a person wants to play the against the spread game, it’s useful to consider a number of things, but if one wants to establish where the Vegas spread might be giving better value in some games than in others, the best way to do this is to create some process where he or she creates their own spreads for comparison. This is tricky and even though I tout an acceptable winning percentage doing this in non-Covid years, I’m never fully comfortable with it and I’m always trying to refine it where I can find the potential to. That being said, if you really want to go down this road, you’re going to find that you will have to put the teams, all 130 or so of them at the FBS level, in some sort of order, and you’re going to have to put them to scale numerically. Remember first, that whatever scale you choose to use, don’t limit yourself to integers only, as there are far more teams in college football than points scored in any game, so you’re going to want to allow the use of fractions. And what I found was that the intervals between the teams, the value I inserted from one team in the order to the next, was of upmost significance.
At first I struggled. I tried lining them up in a linear fashion, with an equal interval between 1 and 2, 15 and 16, 75 and 76, 125 and 126, and so on, all the way through the line with each team being equidistant from the team in front of them and the team behind. Failure. OK fine, so next I attempted classes of teams, dividing all of college football into 10 to 20 groups of teams, with upward and downward mobility jumping from one class of teams to another. Failed again. This all required a lot of test runs, going back years into the team schedules and imputing the projections, as well as the results, and when one does this without any positive results, it not only feels like, but very much is a massive waste of time. I put it aside and forgot about it for awhile, and by that I mean at least a full year of inactivity. One day though, I was having some conversation with someone about human performance, the outliers that would reside both atop the mean and below, and the majority residing somewhere in the middle, even harkening back to my 8th grade gym class, and then boom, the epiphany. College football. I could hardly wait to get home. I went about scaling my teams so that the intervals between both top and bottom teams allowed them to be outliers, with enough symmetry heading toward the middle of the pack that a chart of it would make for a perfectly centered parabola, with there being very little, if any difference between the median of the very large number of teams. And long story made, well, less long, it worked and has for several years now. Until now.
There are, I’m aware, so many reasons why this just isn’t working, and besides the necessity of taking my time explaining this week’s post, which I have deemed as the most significant since the beginning, I am also not adverse to presenting my picks at a time like mid-Saturday, where nobody could fathomably lose money on them. Seriously, I don’t want anyone betting on my advice right now. My hope is that any dedication as a reader will eventually pay off when the system, along with the world, rights itself, but for now, don’t do it. Besides the fact that last week I finally got my sample size, finally some games to play, and the system failed miserably, what’s most alarming as I look at it now, is how the teams are scaling out. Residing at the very middle of the pack, the teams making up the system’s median, or the top of the parabola, if you need a visual, are the following: Colorado, Colorado State, Buffalo, Ball State, Oregon State, Arizona, Stanford and Eastern Michigan. See the problem? None of those teams, not one of them, have played yet! The system now thinks that the exact average performance in college football is best accomplished by not playing at all. You win, Covid.
So that’s where we stand, with my explaining this mid-day on a Saturday, instead of after a fictional 3rd successful week in a row that I might anticipate in another season. And the worst part is that I’m not sure how long the correction might take. First the rest of the teams need to play, but the damage done could take awhile to overcome, perhaps even well into the 2021 season, and that’s assuming we have an eventual end to this deadly pandemic and manage to avert any potential of another. The possibilities are not particularly fun to comprehend without inserting a little bit of optimism, so I’m go inclined to do so, even remaining hopeful that we might see some level of success by the end of this year. Just don’t bet on it. This week’s 6 games are off to a good start, but I wouldn’t look at any results for now as indicative of a reliable model for projection. Here are the games:
Thursday
Georgia Southern -4 at home vs South Alabama
Saturday
Purdue -7.5 on the road vs Illinois
Southern Mississippi -1.5 at home vs Rice
Northwestern +2.5 on the road vs Iowa
Louisiana Tech +12.5 at home vs UAB
Oklahoma State -3.5 on the road vs Texas
We’ll just keep plugging away.
