This season was full of ups and downs for me. At one point in the season I was down over 25 Units, which is a quarter of my betting bankroll. That was difficult to deal with after a massively successful 2020 season and a belief that my process truly works. But I pressed on, knowing that extreme negative short-term variance was the primary cause of my setback. I was rewarded, finishing the season with 7 out of 8 winning weeks, ending 111-96-5 (53.6%) for -2.01 Units. Posting a public loss on the year is obviously disappointing. But in reflecting on the season, I am as optimistic as ever that I will be successful doing this in the long run. I am convinced my process truly works. Here’s why I believe that:
1. The Accuracy of My Macro Team Assessments
Starting in 2020, I’ve now watched every snap of every game for two seasons. I have graded every performance for its overall effectiveness and created a data point for each team’s offensive and defensive performance that goes beyond the numbers and the results. My 2021 season win total projections were independently crafted based on my Effectiveness Ratings from 2020, adjusted for roster moves, coaching changes, injury recoveries, and true strength of schedule (again, using my own independently derived numbers). I did not adjust them at all based on what odds Vegas gave out.
Unlike a lot of “pro” handicappers, my win total projections were fairly off-market. In total, I projected a difference of over half a win above or below the team’s Vegas win total for 15 of the 32 NFL teams. This is fairly significant. But these off-market projections ended up extremely sharp. If a bettor had just blindly bet my off-market numbers against the sports books’ offerings, without even shopping around for the best price, they would have gone 10-3-2 against the number, for a 77% win rate.
| Team | Vegas Win Total | SC Projection | Difference | Record | Result |
|---|---|---|---|---|---|
| BAL | 11 | 9.35 | -1.65 | (8-9) | Win |
| MIA | 9.5 | 8.20 | -1.30 | (9-8) | Win |
| NE | 9.5 | 8.22 | -1.28 | (10-7) | Loss |
| PIT | 8.5 | 9.73 | +1.23 | (9-7-1) | Win |
| ARI | 8.5 | 7.38 | -1.12 | (11-6) | Loss |
| DAL | 9 | 7.99 | -1.01 | (12-5) | Loss |
| CHI | 7.5 | 6.59 | -0.91 | (6-11) | Win |
| NYJ | 6 | 5.16 | -0.84 | (4-13) | Win |
| LAR | 10.5 | 11.20 | +0.70 | (12-5) | Win |
| CIN | 6.5 | 7.19 | +0.69 | (10-7) | Win |
| GB | 10 | 10.67 | +0.67 | (13-4) | Win |
| PHI | 6.5 | 7.14 | +0.64 | (9-8) | Win |
| IND | 9 | 9.62 | +0.62 | (9-8) | Push |
| BUF | 11 | 11.58 | +0.58 | (11-6) | Push |
| WAS | 8.5 | 7.95 | -0.55 | (7-10) | Win |
Most importantly, this shows that I was seeing most teams clearly from the opening bell. I was wrong on the Patriots, but when I made that projection I had Cam Newton starting and understood that Mac Jones brought tremendous variance. Notably, despite being over an entire win under the Vegas win total, I did not bet their Under due to this variance. I also did not bet against the Cowboys, because I knew their offense had upside. Looking back, I did not upgrade them enough for a projected full season with Prescott, a massive upgrade over the quarterbacks they had for most of the year in 2020. And their defense rose from 29th to 3rd in one season, an unprecedented improvement. It turns out defensive performance is much less sticky than offensive performance – but more on that later in the offseason. My one really big miss was projecting the Cardinals to struggle. I’m going to do a deep dive into every team during the off-season, and self-diagnose my mistake there.
My views on teams were also strong during the season. Before Week 18 I placed a bet on the Bengals to win the AFC at +900. When I projected out all the wildcard matchups I picked all six winners, and only lost against the spread with LV +6.5 (they lost by 7, driving to score late). I projected the 49ers to upset the Packers and predicted the exact score of the Super Bowl. The teams that had lower Effectiveness Ratings than market value (GB, ARI, PHI, TEN) all failed in the playoffs when it mattered. Ultimately, despite the bottom line on the season, my team assessments were widely accurate once again, and should provide an extremely strong foundation for 2022.
2. The True Value of My Bets
There are multiple ways to evaluate true betting performance, and none are perfect. The most obvious way is Units won or lost. When betting on roughly +110 to -120 odds consistently, record matters too. But any strictly results-based evaluation misses the nuance and variance in individual outcomes. I don’t want the false impression of being successful when I’m simply getting lucky, just as I don’t want to judge myself too harshly when I’m making good bets that don’t pan out in the short term.
One good (but certainly not perfect) measure of a bet’s value is its closing line value (CLV). That is, the price of your bet as compared to the price at close, when limits are higher and all the bets have been placed. If you consistently bet prices that move in your favor, there is a strong chance you will succeed in the long run. I have audited every spread, total, and money line bet I made this season and categorized every bet according to its CLV:
- Materially positive CLV: A spread bet where the number either moved 2+ points and/or over a key spread number (10, 7, 4, 3) in my favor, or a total bet that moved 2 or more points in my favor, or a ML bet that gained 20+ vig (for example, a +150 ML that closed at +130).
- Minor positive CLV: A bet that moved in my favor but not materially.
- No CLV: A bet where my number matched the closing line.
- Minor negative CLV: A bet that moved against me but not materially.
- Materially negative CLV: A bet that moved materially against me (using the same parameters as category 1).
Using these metrics, I secured a very strong ratio of good CLV over the course of the season:
| CLV Category | Number of Bets | % of Total Bets |
|---|---|---|
| Materially Positive | 77 | 47.5% |
| Minor Positive | 21 | 13.0% |
| No CLV | 29 | 17.9% |
| Minor Negative | 15 | 9.3% |
| Materially Negative | 20 | 12.3% |
This record is particularly impressive given that, of the six games in which the line moved materially due to unexpected Covid or injury news that came out after I placed the bet, five moved against me. Taking out those numbers, I secured positive CLV on 62.2% of my bets, including 48.7% with material CLV. This is compared to just 19.2% with negative CLV and 9.6% with materially negative CLV. Those are winning data points in the long run.
In auditing my bets, I found that 14 bets (out of 162) involved extreme end-game variance that caused an unlikely result. That’s pretty normal when betting football games and you have to allow for that. In the long run, those types of results even out. But in the short run you can get heavily impacted. Unfortunately, I went 4-10 on those bets this season. If any of those 10 had gone the other way, it swings my season into the positive.
My bottom line on the season was also heavily affected by Covid and injury news, compounded by suboptimal reactions to the news. In addition to the six games where the news materially altered the spread, I got involved in a game that ended up getting postponed (to my detriment) and cashed out of a winning bet. The context of these eight games shows how unfortunate I was in these circumstances and how much of an impact it had on my bottom line. Of the seven games that moved against me, I hedged or cashed out on four. All four initial bets would have been winners. I held the other three, and all three were losers.
That last part is hard for me to write because I am firmly on team #neverhedge. I went against my principles because I was posting plays for the first time and got scared. That was negative EV, and I learned from it. Which brings me to my final point: I am still learning.
3. Improving My Process
They say the only constant is change. I try to make sure that change is positive. So for me, the only constant is growth. I developed my methods during the 2019 season and now have two full seasons of working through my process and recording my data. This was the first year posting my bets publicly, and I had some growing pains. But every year I hone my process, and I’m still averaging over 60% overall. I’m getting stronger in my evaluations and decision-making. I have already identified some flaws in my process, such as too wide a range of Unit sizes. Over the course of the off-season I have an opportunity to implement changes and analyze my data to identify strong indicators. This is where I thrive.
If you stick with me, I’ll be starting my deep dive into every NFL team’s 2021 performance. This is my favorite time of year, and where I really thrive and build an advantage. Stay tuned for updates!