Our Wins (and Losses)
Three of our proudest accomplishments (and a two things we’ve gotten wrong)
W: Identifying a Major Baserunning Inefficiency in MLB
We investigated send decisions rounding third and found that teams are far too cautious in sending players home, especially with two outs. Featured on Fangraphs’ Effectively Wild podcast with Ben Lindbergh and Meg Rowley!
W: Going In Extremely Early on the Stud Shortstops Quartet of Wilson, De Vries, Peña, and McGonigle
One of these guys is now an all-star and the other three are fashionable to have right at the top of prospect rankings, but it was not always so!
We have a particular attachment to Leo(dalis) De Vries who has been our top man ever since the launch of our current shiny tool in August 2024 (with the exception of a two week coup by Kevin McGonigle).
It took until mid-summer 2024 for us to fully get on the De Vries (it was his first pro season and he needed 150 at bats to qualify for our model), but we were already all in on McGonigle back at the start of June. Our first “hot take Tuesday” identified McGonigle as a top-five hitting prospect heading into 2025.
We also locked in on Jacob Wilson, putting him in our top 10 last summer despite skepticism about how his profile would translate to MLB.
W: Pioneering Research into ABS Challenge Strategy
With ABS continuing its plodding but unstoppable march towards MLB, we were the first (that we’ve seen) to get out there and figure out just how the heck teams should challenge balls and strikes, and what they’ve been doing up to this point in MiLB.
Our main findings: teams don’t challenge enough, having a good “eye” at the plate doesn’t make you good at challenging, and hitters get the best bang for the buck challenging high pitches. We also found, as the chart below illustrates, that when players are challenging to gain a good result their success rates are better than when they are challenging to avoid a negative result–fear of failure motivates greater aggressiveness.
L: The Ballad of Rayne Doncon
We will always have an affinity for Doncon, but it’s been shown pretty definitively that he should not have been in our top 50 prospects heading into 2025. We wrote at the time “He’s a guy who hits all the model’s favorite traits: a low groundball %, good power, plays SS, and good K%. We’d move him down a bit as his high infield fly ball rates and distance from the bigs cast doubt on his ability to produce, but overall, we’d still rank him far higher than other outlets.”
This older version of the model missed a couple things here, which led to changes in our modeling that have improved the reliability and predictive value of our projections. First, Doncon as we noted there was popping balls up to the infield incredibly frequently (his IFFB% in 2024 was an unseemly 41.1%). Second, despite spending the first chunk of 2024 repeating Single-A after a full season there the year before, Doncon’s strikeout rate and IFFB% both worsened. These issues have been corrected in the latest version of the model. We now use a metric we call OFFBLD% (percentage of hits that are either outfield fly balls or line drives) in our model instead of groundball rate, and use game log data to include trends in strikeout, walk, and home run rate. While it’s never fun to be wrong, the erroneous inclusion of Doncon in our top 50 did help us understand how we could further improve the model.
L: Erroneously Picking the Mariners to Beat the Blue Jays in the ALCS
Ok, so this article was just about the offenses of the two teams, and predicting series winners is notoriously hard, but this still counts as a miss and a learning opportunity.
In our ALCS preview, we wrote: “The Blue Jays outscored the Mariners against both handedness of pitching in the regular season, but our simulation suggests that the Mariners actually have the stronger playoff starting lineup against both righties and southpaws. We expect a full-strength Mariners lineup to outscore the Blue Jays reasonably handily.”
In fact, the Blue Jays outscored the Mariners 32 to 25 in seven games. That could happen by sheer chance, but in this case it was also supported by positive changes that the Jays made to their hitting approach over the course of the season. Multiple key players improved their bat speed significantly in a way that was not fully reflected in the full-season stats that our simulator was trained on.
