Editor’s Note: Since November 2018, Joe has been publishing picks here and back at All Things NIT, our former site. Overall, the results have been mixed, with an average return on investment, per pick, of -5% when weighting by confidence (1 for low, 2 for medium, 3 for high) across 3,418 published picks, not including pending futures, and an average return on investment, per pick, of 0.1% across 773 completed high and medium-confidence picks (low confidence picks, these days, are in experimental markets for us). Our ETA to get the overall number back profitable is currently November 5th, the date of a potential World Series Game 7.
Use these picks at your own risk. Only you are responsible for any money you lose, and you should not bet more than you can afford to lose. If you have a gambling problem, get help.
Lines for these come from the Vegas consensus or the closest approximation available at the time picks are written, unless otherwise noted. For futures bets (in both sports and politics) and motorsports bets, odds are taken from the better option between Bovada and BetOnline as our best approximation of the Vegas consensus, which isn’t currently/accurately available online. FanGraphs is heavily used in making MLB picks. FiveThirtyEight’s SPI is heavily used in making soccer picks. ESPN’s SP+ is often used in making college football picks against the spread. ESPN’s FPI is heavily used in making NFL picks.
College football, MLB, and soccer futures today. For unit context: We started the season with 520 units in our MLB futures fund (20 units per week, 4 units per weekday) and an additional 520 in reserve for potential hedging. We started each soccer season with 50 units in our respective futures portfolio (1 unit per week, once a week) plus an extra 50 in reserve for each for hedges. We’re starting the college football season (admittedly, a bit late) with 60 units in our futures portfolio (4 units per week, once a week) plus an extra 60 in reserve for hedges.
CFP National Championship
We soft-launched our college football model today, and while we’re very confident in its projections of who makes the playoff given various scenarios, we’re less confident in its simulations of the games to get there. Its evaluations of various teams have a strong track record over sixteen years of backtesting, but we don’t want to be overconfident in that, especially since we tailored the formula behind those evaluations to optimize its performance over said sixteen years.
So, we’re going with things that make sense on three fronts: first, our model; second, our own internal understanding of the college football apparatus at present; and third, ESPN’s models, specifically Bill Connelly’s SP+ for comparison to ours linearly (we have Oklahoma State much higher than he does, so we’re laying off the Pokes) and FPI for a gut-check on conference championship probabilities (it served us well last year). Importantly, we are not trusting FPI on playoff projections. We do not endorse any ESPN model on playoff projections, and we want to be loudly clear about this.
Now, Georgia.
They’re the best team in the country. At least 40% likely to win the SEC, plenty of playoff paths beyond that, an absolutely dominant outfit who may have gotten better after winning it all last year. We’re in on a repeat.
Pick: Georgia to win +210. Low confidence.
Pac-12 Championship
In the Pac-12, fear of USC is legitimate. They’re an unknown. Even with that, though, the Pac-12 switching to a top-two conference championship instead of a division-winners conference championship puts value behind Utah. They don’t necessarily even need to be better than USC in the regular season for this to hold value for us. Though, to be clear, we think they’re better than USC, and disagreement on that where it does exist is either overly simplistic or describes the Utah/USC gap as a coin toss.
Pick: Utah to win +250. Low confidence.
Big Ten Championship
I’m unconvinced that Ohio State’s better than Michigan, especially after Marshall had an even easier time stopping Notre Dame’s offense than the Buckeyes did. We won’t learn much about the Wolverines until October (barring something cataclysmic happening for them), but we like these odds.
Pick: Michigan to win +400. Low confidence.
ACC Championship
And finally, Pitt. FPI and SP+ are a little lower on these guys than we are, but they aren’t so low as to make this appear outright foolish. Pitt’s a good football team, their division’s wide open, and while they have to go to Miami, Miami has to go to Clemson the week before. In some ways, this is buying us the possibility of hedging against the Panthers if they do make the conference title game, or get close enough for it to be a possibility. We’ll happily make that purchase.
Pick: Pittsburgh to win +1600. Low confidence.
World Series
Doubling up on this one today, giving us a little stronger angle on one of the favorites. Last night’s win was expected but big for the Mets, pulling them a full game ahead of Atlanta. They’re likely to hit 100 wins, and they’ve got the best pitcher in the world anchoring their staff. There is a ton to like here, and the narrative has lost sight of that.
Pick: New York (NL) to win +650. Medium confidence.
Pick: New York (NL) to win +650. Medium confidence.
Premier League
Our portfolio is longshot heavy, and this is one of the few safer plays that’s showing value. So, it’s a good fit for us today. It’s hard to see these guys staying up.
Pick: Nottingham Forest to be relegated -125. Low confidence.
EFL Championship
There’s probably better value out there, but we want to capitalize on the value we’re getting from holding Norwich +500 to win the EPL, and with the possibility either they or Sheffield United runs away with this, we’re taking this still-positive-value but not sensational play. So far, this pair has been a cut above the rest of the league, and we expect it to more or less stay that way. If Burnley or West Brom starts to get it together (we’re already in on Watford), we can adjust, but we want this in our pocket.
Pick: Sheffield United to win +225. Low confidence.
