We like things that are easy. It’s why we add together OBP and SLG to get OPS, which we use as a shorthand for offensive production. In mathematical reality, you never add things with different denominators. But adding together OBP and SLG gets you very close to the “right” answer so we do it anyway. If you’re at a computer and can look up FanGraphs, you’re better using wRC+ – but who likes typing mixed cases and a special symbol? Save that garbage for sites that think it adds to your online security.
The other thing about OPS is that it undervalues OBP. The math majors have determined that it really should be OBP*1.8 + SLG to get a more accurate number. For the majority of players, this doesn’t make much difference. But when you get guys who excel at OBP, you might see it as a problem properly valuing their contribution. Enter Brandon Nimmo into the discussion.
Lifetime, Nimmo has a .387 OBP and in 2018, he reached base at a .404 clip. In that season, he ended up with an .886 OPS and a 148 wRC+. Bryce Harper finished that year with an .889 OPS and a 134 wRC+ while Xander Bogaerts had an .886 OPS and a 133 OPS+. Why did the players who finished directly above and below Nimmo in OPS have significantly inferior wRC+ numbers? Some of it is ballpark-related, as wRC+ (like all “+” stats do) adjusts for park. The rest is properly valuing OBP.
Let’s look at a player on the 2019 Mets to take away most of the ballpark adjustments. Pete Alonso had a SLG-heavy .941 OPS, a 55-point edge on 2018 Nimmo. Yet Alonso’s wRC+ was 143, five points below what Nimmo produced the year before. We still have another factor at play here, as wRC+ adjusts for league context and 2019 was a more hitter-friendly season than 2018. If we just look at September of 2019, Nimmo had a .995 OPS and a 159 wRC+.
If we did OPS using the 1.8 OBP valuation we get:
Nimmo – .404 *1.8 + .483 = 1.210
Alonso – .358 *1.8 + .583 = 1.227
The 1.8 OPS valuation cuts the OPS deficit between the two from 56 points to 17. It doesn’t put Nimmo ahead, like wRC+ does, but it cuts away roughly 2/3 of the difference. So, our easy OPS formula shows a big advantage for Alonso comparing the two seasons, our slightly more complex 1.8 OBP shows a reduced advantage for Alonso and our most accurate and complex offensive measurement shows a slight advantage for Nimmo. Here’s the formula for wRC+ if you want to see how complex it really is.
If given a choice to pick a player strictly for their offensive performance, 2018 Nimmo and 2019 Alonso are very similar if you properly weight all of their outcomes in the batter’s box and adjust for league and park context. The average person is going to see Alonso’s 53 HR and conclude he’s better and then invent all kinds of narrative to explain why. If given a choice between numerical proof or narrative, you should opt for the former.
This is different from the usual projection post because most of what usually goes into these pieces was covered back in November. So, let’s take a look now at the computer models and see what they forecasted for Nimmo in a 162-game season:
ATC – 499 PA, .238/.369/.418, 15 HR, 50 RBIs
Marcel – 381 PA, .252/.379/.457, 13 HR, 40 RBIs
Steamer – 495 PA, .232/.358/.400, 14 HR, 49 RBIs
THE BAT – 496 PA, .240/.358/.413, 15 HR, 50 RBIs
ZiPS – 431 PA, .237/.365/.420, 13 HR, 45 RBIs
The computer forecasts have more variability in these forecasts than for most Mets we’ve seen so far in this series. A lot of that has to do with the missed playing time in 2019. That’s not a surprise. What is curious to me is that they expect more playing time and a worse wRC+ than what Nimmo produced a year ago. Since Nimmo’s 2017 and 2019 production was so similar – and 2018 significantly better – my assumption was that the computer models would have forecasted a slight increase in production. My expectation from these computer forecasts was a wRC+ in the 120s.
When we started this projection series back in 2013, the hope was that by using the computer models as a starting point and then adding in things we could see as fans that the computer models did not take into account – that we could come up with a better forecast. The computer models don’t know that Nimmo’s 2019 season was ruined by a neck injury, one that he played with for about a month when he should have been on the IL. Instead, all they know is that he played in just 69 games and had a drop of 45 points of wRC+. They don’t know how he came back in September and performed even better than he did in 2018.
So, here’s my completely biased forecast for Nimmo, based on a 162-game season:
You’ll have more credibility if you chime in now with what you think Nimmo will do this year. Next up to undergo the forecast microscope will be Michael Wacha.