SAN FRANCISCO — Pirates general managers, radio shows and projections haven’t always been the greatest match. Remember the Neal Huntington/10,000 simulations debacle from August 2019?
What Ben Cherington said on 93.7 FM The Fan Sunday wasn’t that. However, his comments that the Pirates’ internal projections didn’t have the team playing above .500 through 51 games seemed to upset a sizable portion of the fan base.
As the Pirates have one more chance to actually win a series this month — an example of the back-to-earth crash they’ve experienced in May — let’s work through what Cherington said and try to separate fact from fiction.
The setup
Host Greg Brown started the exchange with a fairly simple question: by asking if the Pirates’ internal projections had them over .500 through 51 games.
“No,” Cherington said. “And I don’t think many others did either. I don’t mind saying that.”
Let’s pause right there, because this is what seemed to irritate people.
Here we have the general manager of a professional sports franchise openly saying it expected to lose more than it won nearly a third into the season — in a year where there’s been ample talk about the need to improve and actually win games following back-to-back 100-loss seasons and one more (2020) played at a 100-loss pace.
That … doesn’t seem like the kind of thing you’d want to say publicly. But Cherington continued and began to discuss some important caveats with projections.
‘They’re still imperfect’
Projecting team and player performance is very much a science. It’s not folks sitting in a room, acting on gut feelings. There’s mathematics and leveraging mounds of data to build intelligence and attempt to predict outcomes.
If anyone thinks this stuff is done in a haphazard way, attend the next Carnegie Mellon University sports analytics conference. There’s fascinating and thorough work done by extremely smart people.
Every team does its own projections. External websites such as FanGraphs do, as well. Flaws are everywhere.
Not even Mitch Keller’s parents could’ve predicted this for him. Doubt anyone had Oneil Cruz’s ankle injury, Jarlin Garcia unable to grip a ball or one homer from Ji-Man Choi on their Bingo cards. You also can’t predict luck.
“Projection systems can’t know any of that stuff, so they’re not precise,” Cherington said. “They’re just more precise than I would be if I was doing it on my own, which is why we use them. They’re so valuable because it gives us an anchor. A clear-eyed, objective starting point.”
This sounds a little more reasonable. Although, again, if the Pirates didn’t predict the Pirates to play .500 ball with everyone healthy, what do those projections look like now? We might not want to see.
‘Prove them wrong’
One of the more palatable things Cherington offered during this portion of his radio show involved how the Pirates think about projections. Working to disprove projections is a huge part of the job, Cherington said.
Projection models are rarely kind to teams like the Pirates because they rely so much on year-to-year changes and also seeing value where other teams or external analysts do not.
“We talk about that as a staff all the time,” Cherington said. “The job we have is to basically take the projections and beat them. That’s fundamentally what we’re being paid to do. That’s kind of the fun part of it. Whatever they’re telling us, let’s prove them wrong.”
The uncomfortable truth
It might hurt to hear, but Cherington really only said what we already know out loud.
Think about it: Few sane people before the season picked the Pirates to win 82 or more games. So why are we shocked now when the GM says their own projection models didn’t have them going 26-25 before another loss on Sunday dropped them back to .500?
Sure, maybe it would’ve been nice to play above .500 for a third of the season, even for around a third, then fall below .500 for the final third when injuries, trades or a lack of depth take their toll. Maybe randomization allows the Pirates to flirt with winning 82. That’s also not reality here.
Vegas set the over/under for Pirates wins at 67. The optimistic crowd went low-to-mid 70s. That’s still within range, and there are a variety of unpredictable factors to consider: Keller being ace-level good; team pitching that has been better than expected; an offensive nosedive in May; the aforementioned injuries; Andrew McCutchen exceeding expectations, among other small items.
The only disappointing thing is probably that Cherington, in his own way, admitted that the Pirates’ desire to win more games isn’t really a desire to compete. It’s merely to stink less than they have in recent years.