Are Umpires Worth Considering When Handicapping MLB?
Around a decade ago, umpires were a significant part of my MLB handicapping process. Put John Hirschbeck or Doug Eddings behind the plate and it was an automatic bet on the under. And on the rare occasion one of those two was calling a “getaway day” game, my bet size would reach well beyond the normal 1-2 unit size. But like all good things in the betting world, it eventually fizzled out. Hirschbeck retired at the end of the 2016 season and while Eddings continues to have a wider than normal strike zone he somehow recently trended over the total three straight seasons (2014-16). The biggest impact on umpires these days is the technology introduced to track how they perform. Most bettors will tell you that today’s umpires simply don’t have much in the way of an identity. And who can blame them as the call for an automated strike zone grows louder and louder. But until the robots completely take over, we’re talking about humans making countless amounts of judgement calls. Take a look at a MLB pitch chart and you’ll notice a majority of pitches are within a baseball-sized margin of the strike zone box. Of those pitches, a number could be called both ways and if you’ve ever bet baseball you’re well aware that literally one pitch can sway the outcome of a game. I do believe that certain umpires are more prone than others at calling that outside pitch three inches off the plate but in looking at some of the data — strike call rate and K-to-BB ratio — there appears to be some randomness when trying to correlate it to over/under results. Below we took every full-time umpire who has trended over or under the total by 55% or more since 2015. For most of our qualifiers, that gave us around a 100-game sample size.
Manny Gonzalez: 73-43-6 (62.9%)
Jim Wolf: 70-45-14 (60.9%)
Dana DeMuth: 48-31-3 (60.8%)
Adrian Johnson: 68-45-8 (60.2%)
Brian Knight: 62-42-2 (59.6%)
David Rackley: 64-47-5 (57.7%)
Bill Welke: 68-50-8 (57.6%)
Alfonso Marquez: 70-53-4 (56.9%)
Paul Emmel: 51-39-4 (56.7%)
Mike Winters: 67-52-3 (56.3%)
Tim Timmons: 69-55-3 (55.6%)
Mark Wegner: 67-54-8 (55.4%)
Jeff Kellogg: 58-47-3 (55.2%)
Chad Whitson: 43-35-2 (55.1%)
Lance Barksdale: 47-70-11 (59.8%)
Bruce Dreckman: 35-52-4 (59.8%)
Mike Estabrook: 46-68-7 (59.6%)
Dan Bellino: 45-66-13 (59.5%)
Cory Blaser: 50-71-2 (58.7%)
Vic Carapazza: 49-69-7 (58.5%)
Mike Everitt: 43-60-3 (58.3%)
Hunter Wendelstedt: 46-63-13 (57.8%)
CB Bucknor: 45-61-9 (57.5%)
Gerry Davis: 52-70-5 (57.4%)
Rob Drake: 43-56-7 (56.6%)
Phil Cuzzi: 50-65-2 (56.5%)
Eric Cooper: 51-63-2 (55.3%)
Brian Gorman: 47-38-4 (55.3%)
Small Sample Size Over
Jansen Visconti: 17-13-1 (56.7%)
Small Sample Size Under
Jeremie Rehak: 9-16-2 (64.0%)
Ryan Additon: 12-19-1 (61.3%)
What I found was pretty interesting. Since 2015, Manny Gonzalez has trended over the total at nearly a 63% rate — the highest mark in MLB — while Lance Barksdale has been the strongest “under” ump at just shy of 60%. Take a look at their statistics during that span…
Manny Gonzalez (10.3 rpg, 15.8 kpg, 2.35 K/BB ratio, 63.1% strikes)
Lance Barksdale (8.7 rpg, 15.8 kpg, 2.37 K/BB ratio, 63.1% strikes)
Nearly identical with the one glaring difference; a whopping 1.6 runs per game!
Adding to the confusion is the aforementioned Eddings. Eddings remains one of the more pitcher friendly umps in the game based on the stats but those stats sure haven’t translated into unders as a 121-game sample shows 53.1% of his games have gone over the total. Note that Eddings has been around since 1998 and despite trending towards the over, the scores of his games (8.5 rpg since 2015) were below league average (8.92 rpg since 2015). He’s one of maybe a handful of veteran umps who I feel oddsmakers and the betting markets still account for in how a game is priced and bet.
Breaking it down even further, here are the year-by-year rankings of Gonzales, Barksdale, and Eddings based on K-to-BB ratio.
2019: 67th out of 85
2018: 83rd out of 89
2017: 56th out of 92
2016: 82nd out of 90
2015: 67th out of 92
2019: 55th out of 85
2018: 72nd out of 89
2017: 60th out of 92
2016: 61st out of 90
2015: 87th out of 92
2019: 63rd out of 85
2018: 19th out of 89
2017: 2nd out of 92
2016: 5th out of 90
2015: 5th out of 92
Gonzalez’s K-to-BB rates do actually line up with that of an “over” umpire but Barksdale’s are very similar and here he is, cranking out 60% unders. And per usual, Eddings ranks near the top but he’s been more of less “neutral” for going on four-plus years.
In the end, nearly half of the league’s umpire roster has “trended” over or under the total based on our 55% parameter. The problem however is there just doesn’t seem to be a clear cut path to accurately project whether or not any of them will continue trending in such a strong manner. Note that three of Gonzalez’s four games in 2019 have gone under the total which based on the time and effort to put this piece together, had me saying, “that sounds about right.”
Before we conclude this study, let’s go back to the influence of technology. Umpires get reports after each game on how they performed. Let’s say an umpire called five strikes on the outer half that according to Statcast were balls. Not only is the umpire made aware of this, but teams and fans are too. Nearly every night on Baseball Twitter you’ll see a screen clip of an umpire who botched a bunch of calls. Years ago, no one would have batted an eye because, without the technology, it was literally a judgement call. Nowadays, when an ump screws up, a correction is likely made. As a result, I feel that there’s a lot more uniformity. In a perfect world, MLB wants every umpire to have the same strike zone and Statcast is there to help assist in achieving that. That doesn’t mean that umpires still don’t favor certain portions of the plate or want to hit the links after a Thursday day game. But overall, while the results say we should consider betting totals based on some of these umpires, the statistics to a large extent do not.
The best advice I can offer if you are determined to make money betting umpires is at minimum make sure their statistics line up with their results. Or perhaps the better approach is to be proactive and isolate umpires whose stats DON’T currently line up with their results. A good example of this would be Barksdale who statistically should be more “over” or at least “neutral.” One would think that his under rate will eventually regress towards the mean. Of course Barksdale has already cashed on three of four unders this season because MLB is notorious for making a bunch of sense.