What Were MLB Forecasters' Best (And Boldest) Picks?
Checking in at the one-third mark of the season.
Picking which MLB teams will be good or bad isn’t easy. There’s the delicate balance of weighing last season versus previous track records and/or the possibility of what players might do in the future. There are also those pesky injuries that can foul everything up. So when forecasters go out on a limb and get things right, it’s good to celebrate those wins.
In that spirit, I looked at the preseason forecasts from four different sources (as of March 23) — Clay Davenport, Baseball Prospectus, FanGraphs and Out Of The Park Baseball 24 — and calculated the season win-total projections that are on pace to finish best (according to the lowest squared error against Baseball-Reference’s forecast, a neutral arbitrator). I also wanted to highlight successful “maverick” picks: those that most bucked the overall consensus and ended up going in the right direction.
Let’s run down the forecasters in order of overall accuracy (a lower error is better) so far this season.
Out Of The Park Baseball 24
Root Mean Square Error: 7.82 wins
Most accurate pick: Arizona Diamondbacks (84.1 wins vs. B-R pace of 84.0)
Best maverick picks: Miami Marlins (under consensus of 76.8; B-R pace of 72.3), Arizona Diamondbacks (over consensus of 77.9; B-R pace of 84.0), Philadelphia Phillies (under consensus of 86.4; B-R pace of 75.0)
While OOTP’s simulations were at their boldest when calling for the Marlins win 10 fewer games than the rest of the forecasts, I’m most impressed that the game’s engine/ratings were enthused about the Diamondbacks. Arizona, led by Geraldo Perdomo, Zac Gallen and Corbin Carroll, has been one of the biggest pleasant surprises in MLB so far this season. Bonus points go to OOTP as well for thinking the Phillies would regress hard following their magical World Series run; while we can’t rule out another one of those, Philadelphia is already in a rough spot to start the year.
FanGraphs
Root Mean Square Error: 7.84 wins
Most accurate pick: New York Yankees (91.5 wins vs. B-R pace of 91.6)
Best maverick picks: Detroit Tigers (over consensus of 67.9; B-R pace of 75.2), Minnesota Twins (under consensus of 87.2; B-R pace of 84.6), Cleveland Guardians (under consensus of 87.4; B-R pace of 77.6)
Apparently betting against the AL Central favorites was the smart move before the season. While the other forecasts called for the Twins and Guardians to battle for the Central crown with somewhere between 88 and 90 wins, FanGraphs saw this division for the trash heap it has ended up being — with a collective 106-144 record (easily the worst of any division) and only one team above .500. Minnesota, that lone winning squad, remains the AL Central favorite… but the Twins will likely do battle with whoever wins the NL Central for the title of 2023’s worst division winner.
Clay Davenport
Root Mean Square Error: 7.89 wins
Most accurate pick: Cincinnati Reds (68.0 wins vs. B-R pace of 67.5)
Best maverick picks: Oakland Athletics (under consensus of 66.3; B-R pace of 49.0 😱), Kansas City Royals (under consensus of 68.3; B-R pace of 61.0), Boston Red Sox (over consensus of 82.4; B-R pace of 82.7)
It didn’t exactly take Nostradamus to think Oakland would stink this year; the A’s traded away by far the best player (Sean Murphy) on a team that had gone 60-102, and did little otherwise to improve. But Davenport’s projection was the closest to foreseeing just how historically atrocious this Athletics team would be. Oakland is the first AL or NL team to win 10 or fewer of its first 52 decisions since the 1932 Boston Red Sox also went 10-42 to start their season. They’re also currently on pace to allow the most park-adjusted runs per game relative to MLB (2.73 RPG worse than average) in the history of baseball. These A’s are so bad that they may not finish the season within 20 games of that 66.3-win consensus projection.
Baseball Prospectus
Root Mean Square Error: 8.49 wins
Most accurate pick: Toronto Blue Jays (89.2 wins vs. B-R pace of 88.4)
Best maverick picks: New York Yankees (over consensus of 90.9; B-R pace of 91.6), Kansas City Royals (under consensus of 68.3; B-R pace of 61.0), Houston Astros (over consensus of 90.8; B-R pace of 95.7)
The Yankees were an interesting team to project going into 2023: Coming off a 99-win season, their biggest offseason move was re-signing a megastar (Aaron Judge) who was probably going to take some kind of step back from a season for the ages in 2022, and their second-biggest move was signing a starter (Carlos Rodón) who hasn’t thrown a pitch yet this season because of various injuries. As a result, the forecast systems saw regression coming in the Bronx. But BP was more bullish on the Yanks than any other projection, and it looks like New York is overcoming its slow start to validate that pick. The same can also be said of the Astros, who hovered around .500 throughout April but are 13-8 in May and have won 11 of their past 15 games.
Filed under: Baseball