Saturday SOC: The Full-Time Tyler O’Neill

Greetings from the land of sleet and snow. This morning, Friday, I’m coming to you from my “man cave” basement room surrounded by my wall-o-Cardinals jerseys, my guitar collection, and other visual displays of my extreme nerddom.

That means you get a stream of consciousness post while I mentally debate the merits of shoveling 2 inches of sleet off my driveway.

This week I was thinking about Tyler O’Neill. Something I highly recommend because my golly gosh is he worth thinking about. I’ll probably do a deep dive on his stats later this offseason (which looks like it’s going to go on well into spring). For now, though, I am returning to an old argument: Tyler O’Neill is much better as a full-time player than he was as a part-time player.

That debate was settled once and for all this past season when O’Neill busted out with a near-MVP caliber 5.4 fWAR, rooted in his .384 wOBA. Yes, his K’s stayed up but some of his walks came back as the season progressed and his power never left. His .274 ISO while playing half his games in Busch Stadium is the stuff of demi-gods.

That kind of production over 537 PAs did positive work on his career numbers. His overall slash line is now .260/.324/.497 with a .348 wOBA and a 121 wRC+. That’s a very good career line.

But could it be even better?

For the first two seasons of O’Neill’s career, he was a part-time player, getting a handful of sporadic plate appearances whenever the already-forgotten Marcel Ozuna and Dexter Fowler were out of commission. I and others noted that when O’Neill received regular playing time – say, a week or more of starts in a row – he was able to settle in and produce at a pretty high level. When he was asked to be primarily a pinch hitter and defensive replacement, his K rate rose and his production stats went all over the place.

That got me wondering – if we remove all of those volatile spot-start, pinch-hit appearances from O’Neill’s career, what happens to his overall numbers?

It’s kind of a pointless exercise since he’s firmly entrenched as a starting player. However, if I was to be able to show that O’Neill has been a very productive player throughout his career as a full-time starter, wouldn’t that help us feel more comfortable about, say, offering him a contract extension? Or projecting a high level of performance going forward? Sure, it would.

Let’s do that.

What I have below is every section of games where O’Neill received 3 or more plate appearances in at least a few consecutive games for each season of his career. Some of this involves a little subjective selection since he frequently came off the bench for a game or two even during stretches where he was the de facto “starter”. In the end, though, there are pretty obvious stretches where O’Neill was starting and when he was not. In the data below I am including his first day back in the lineup as a starter and all the data through the end of that run of playing time, including the occasional pinch-hit appearance that’s bookended by regular starts.

I have a wide selection of statistics available for this kind of split, so I’ll go ahead and use a full array: PAs, BA/OBP/SLUG slash line, wOBA, and wRC+. Thankfully, Fangraphs has a nice sum function for game logs to make my job easier, and I took an Excel class once upon a time.

Here’s an image of the spreadsheet for your viewing pleasure:

Subjectively scratching out all of O’Neill’s pinch-hit appearances in 2018 does a wonder on his overall stats.

2018 actual: .254/.303/.500, .340 wOBA, 116 wRC+
2018 starter: .283/.312/.556, .364 wOBA, 132 wRC+

Pinch-hitting stole 16 points of wRC+ from O’Neill’s performance. Of course, it also erases 46 plate appearances. So, we need to keep in mind the small sample sizes we are dealing with. Still, sample size acknowledged, there’s an obvious distinction in his performance by playing time. It’s clear he was pretty great when he started games consistently in 2018 and pretty rotten when he came off the bench.

The same was true in 2019:

2018 actual: .262/.311/.411, .308 wOBA, 92 wRC+
2018 starter: .289/.333/.453, .334 wOBA, 109 wRC+

The difference here is remarkably consistent. In 2018 there was a .024 point difference in his wOBA full season vs. as a starter and a 16 point difference in wRC+. In 2019 that wOBA difference was .026. His wRC+ difference was 17. Nearly identical.

You can see why the Cardinals were willing to roll with O’Neill as their full-time starter in 2020. They had this data. Their math was probably righter than mine. Based on when O’Neill was in the lineup and not riding the pine, the club had every reason to believe that he would be well-above-average, even for a corner outfielder. His wRC+ during his stretches as a starter would be around 120 with massive power numbers.

Then 2020 happened. Notice that I’m not in any way playing with his 2020 data. He was a full-time starter all season. He was pretty terrible. We’ve gone into the reasons why. Most of it is a BABIP-collapse, but considering what happened with his BB and K numbers (BB’s rose and K’s fell…) it might have been at least partially self-inflicted. Normally cutting K’s and increasing BB’s has a positive effect on a batter’s hitting line. Not for O’Neill. We can get into more of that later. For now, just know that I didn’t cut a single PA from 2020 in the chart above or his career numbers.

2021 was a breakout season but in the context of his split stats as a starter it’s not as out-of-left-field. A 132 wRC+ when he was a consistent starter at age 23 does (sort of) portend a possible 145 wRC+ as a 26-year-old, 4-year veteran having a career season.

Add it all up and if we cut out his pinch-hitting appearances, O’Neill’s career line ticks up enough to be interesting:

O’Neill Career: 987 PA, .260/.324/.497, .348 wOBA, 121 wRC+
O’Neill Starter: 886 PA, .266/.333/.517, .360 wOBA, 128 wRC+

What to do with this data? Here’s one of the things that I love about ZiPS’ projection systems. Dan Szymborki’s system looks at all kinds of things to create his models for the future. I don’t know that pinch-hitting vs. starting is part of the equation. I do know, though, that he is very thorough.

O’Neill’s ZiPS projection for 2022? .266/.334/.535, .366 wOBA, 133 wRC+.

ZiPS’ projection is within spitting distance of what O’Neill has done throughout his career when he’s been a regular starter. Szymborski’s computer is .006 higher in wOBA and 5 points in wRC+ – all well within the margin of error for this kind of projection.

Compare that agreement to Steamer. Steamer Projections has O’Neill dropping back to a .251/.318/.494 slash line with a .344 wOBA and a 119 wRC+. That’s much closer to O’Neill’s actual career line, which includes all those failed pinch-hit appearances early in his career.

Based on the game log data I present above, who should you believe? ZiPS’ line, which matches O’Neill’s historical performance as a consistent starter? Or Steamer’s line, which matches O’Neill’s overall career performance regardless of playing time environment?

I’ve tipped my hand already. I don’t fear a massive dropoff in production from O’Neill this season. Yes, his 2022 was an outlier compared to his career stats and he likely won’t replicate that. But if you start cutting out PAs that don’t matter – pinch-hit appearances 4 years ago – it doesn’t look so far off that he won’t remain a very, very productive corner outfielder.

This is not a “Four Stats that Matter for 2022” article, but if I was going to put a prediction on O’Neill’s upcoming season, I think I would stick exactly with what ZiPS has. O’Neill should remain very good and there’s plenty of material in his past to support that kind of optimism.

Happy sledding! Or shoveling, maybe? No… I’m going sledding.