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 –2% when weighting by confidence (1 for low, 2 for medium, 3 for high) across 7,910 published picks, not including pending futures; and an average return on investment, per pick, of +1% across 2,397 completed high and medium-confidence picks. The low confidence picks are the problem. Most of our picks are in the low confidence category.
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’re afraid you might have a gambling problem, seek help.
Odds for these come from the better option between Bovada and BetOnline. We used to use the Vegas Consensus, but it’s no longer consistently available in an accurate form online. We rely heavily on FanGraphs for all our MLB action. We rely heavily on Nate Silver’s presidential election model for our election futures. For college football bets, we primarily use our own model. For NFL bets, we lean on ESPN’s FPI.
Active markets: MLB moneylines, MLB futures, single-game NFL. Here’s the context on each:
MLB moneylines: Last year, we finished the season up about 8%. We’re 170–152–4 so far this year, down 13.21 units. It’s been a bad showing, and we are grateful it will soon be over.
MLB futures: In four of the five years we’ve done these, we’ve profited, twice by large margins. We began this season with 750 units in our portfolio. We place two medium confidence bets most weekdays throughout the regular season. Now that division futures have begun paying out, we’re doing it on weekends as well. These are still tracking well—of the expected twelve playoff teams, three would be bad World Series winners for our portfolio, four would be great, and the other five would be in between—but there’s a long way to go.
Single-game NFL bets: We published these for the first time last year, and they went terribly. So far this year, though, we’re 7–4 and up 2.47 units.
Kansas City @ Atlanta
Our “bet against the perceived hottest team” test has not gone well (down almost four units in less than two weeks), but we are committed to the test. It doesn’t make much sense for the Royals to win today, but that’s kind of the point. This is baseball, and we still suspect there might be some value in that perception.
Pick: Kansas City to win +183. Low confidence. (Marsh and Morton must start.)
World Series
We love the FanGraphs Playoff Odds model. It does a great job. Our one complaint is that it gets weird these four days of the year when the rotation balance for the rest of the season is so different from what it will be in October. We really wish they’d build in a postseason depth chart projection and not just continue the one for the rest of the regular season through the playoffs in their simulations, but it’s their model, not ours. Thankfully, the problem will resolve itself either tomorrow or Tuesday.
In the meantime, this is flashing positive EV and we think that’s real. The Yankees clinched home-field advantage on the AL side, which is partially important because it helps them avoid the Astros in the Division Series. With these guys slipping into unprofitable status for us again, doubling down on them helps get that scenario back above even while we wait for the NL Wild Card picture to play out.
Pick: New York (AL) to win +450. Medium confidence. x2
Buffalo @ Baltimore
To recap our NFL strategy: We’ve been siding with FPI, with probably too generous an interpretation of home-field advantage. Because we publish one pick a day on days there are games, this has led us to bet more on Thursday and Monday games than on Sundays. We haven’t had as many choices on Thursdays and Mondays, and in our extremely small sample, that’s gone better for us. So, we’re sticking with it and going with the Sunday Night Football game.
The issue with this is that because the Bills have played on a Thursday and played on a Monday, we’ve bet on the Bills multiple times now. Take away the Bills, and we’re only 5–4 on the year, which is even less meaningful than 7–4.
Is our strategy to trust FPI? Is it to trust FPI in night games? Is it to bet the Bills? Is it even working in the first place, or is eleven picks too small a sample for us to glean anything at all from this exercise?
The strategy is to trust FPI, and we’re limiting it to night games to try to take judgment calls out of it as much as we can. Today, that leads us to the Bills. We’ll handle tomorrow tomorrow.
Pick: Buffalo +2.5 (+100). Low confidence.
