How many “on base machines” do the Phillies have?

During the gamethread for the final Phillies/Dodgers game of the year, an embryonic conversation started to develop around how nice it is that the Phillies may have an “on base machine” on their hands, or maybe even more than one.

The first question I had was, “What is an ‘on-base machine’?”

Using Baseball-Reference.com’s 2017 league hitting numbers as a starting point may help answer this. MLB average for on-base percentage (OBP) is .325 this year. Batting average is at .255 and walks can be back-of-the-enveloped at 8.5% from 3.26 walks in 38.19 plate appearances per game. Also, Fangraphs tells me BB% is 8.5% here.

This will obviously vary from position to position. Average for a shortstop is .260/.314/6.9% (batting average, OBP, and BB% - this convention will continue throughout). First base looks like this: .261/.343/10.4%. Second base is: .263/.328/8.0%. Outfielders vary a bit by position, but not as much as I expected. As a whole: .260/.332/8.9%.

All that is too confusing to remember without a chart, so here’s a chart with some made up stuff at the bottom. I’ll get to that in a minute.

Some goon at my house who has been eating mescaline

At an MLB-wide level, to be an on-base ‘machine’ (“OBM”) the threshold requirement appears to be superior performance in getting on base. Round numbers make great metrics because of their roundness. Round numbers are awesome because of the Golden Ratio or the Mother Goddess or something. I picked (right out of my ass) an on-base threshold of .350. Should we ignore components? The god of OBP does not care how you do it, right? Just don’t make outs. Except an insane BABIP year could give crapola results and mess up this whole exercise, so should batting average just be completely ignored in favor of walks? Conversely, if someone walks a lot but can’t hit a lick, their OBP won’t be that great, so walks alone are not enough.

I decided unilaterally on .235/.350/9.5% as MLB baselines.. You can be a bad hitter (for average) as long as you get on at a .350 clip. And your walks have to be above average. I’m ignoring hit batters and such, and I know my numbers don’t add up. It’s not internal consistency we seek, after all, it is an objective measure of my subjective soft, fuzzy feeling about the rule of thumb goodness of the things I actually want to measure, which is getting on base goodly.

The made up crap on the chart contains my effort to take the baseline of each position and make it match the made-up “goodly” numbers I created out of whole cloth for the MLB average. Basically, they are purely arithmetic adjustments, nothing more or fancier.

Now let’s test this out a bit.

Rhys Hoskins: .293/.425/17.8% vs 1B OBM baseline of .241/.369/11.62% — Winner!

Cesar Hernandez: .291/.363/9.4% vs 2B OBM of .243/.353/8.94% — Winner!

J.P. Crawford: .255/.379/17.8% vs SS OBM* (SSS) of .240/.338/7.71% — Winner!

Aaron Altherr: .281/.351/7.8% vs LF OBM of .241/.355/9.61% — Sad Trombone

Cameron Rupp: .217/.299/10.3% vs C OBM of .225/.338/8.94% — Sad Trombone

Andrew Knapp: .250/.356/14.4% vs C OBM of .225/.356/8.94 — Winner!

And so forth.

Rupp would qualify except he doesn’t hit enough. Altherr would qualify except he doesn’t quite walk enough. Herrera likewise needs to walk a bit more. Galvis needs to hit and walk more, even at the reduced target for shortstops. Franco...well...no. Just...no.

Does this answer any questions about anything important? Not really, I don’t think. It’s just trying to think about a term we sometimes toss around (“on base machine”) and trying to figure out what it might look like to codify it into quantifiable rules. The arbitrary ones I picked confirm what we already know from looking at OBP, which is Hoskins, Hernandez, and Crawford get on base well both generally and relatively within their respective positions.

One of the fun things I learned about while trying to define OBM was the variability by position of various offensive statistics. I always kinda sorta knew, but seeing the table was fun. Also, learning that MLB has had 4,981 plate appearances of total dreck by pitchers (wRC+ of -18) may have made me re-think my anti-DH position. I was also a little surprised that designated hitters weren’t collectively a little better than they were.

Anyway, it was kind of fun to think about on a night when we already ate all our baseball during the day and have to go wanting at night.

*Crawford is a superior on-base performer in the small sample size we have for him. His average is not superior among shortstops but his walk rate is crazy high, though, more than offsetting for a minor sub-par component. Let’s take our thumb and smear those colors together to get an accurate picture shown by the OBP. I also know that you know that I know that we all know that batting outcome sample sizes, including walk rate, do not stabilize at 58 plate appearances. Reference.