Rowdy Tellez: Bullish or bearish?

This short series will examine several Blue Jays with shorter track records around whose production there is significant uncertainty and the reasons to be both bullish or bearish on what they will do in 2021. Having exmained Teoscar Hernandez, today we turn to Rowdy Tellez.

Projection Systems:

• Steamer — .255/.325/.468 in 417 PA, 107 wRC+ and 0.7 WAR
• ZiPS — .261/.324/.500 in 497 PA, 112 wRC+ and 1.6 WAR
• Marcel — .249/.319/.465 in 417 PA, ~104 wRC+
• The BAT X — .251/.322/.469 in 434 PA, 109 wRC+

Expos45 already did a very give into Tellez (linked above) and his 2020 breakout, and I commend it in its entirety. The main takeaway is that the improvement in strikeout rate look largely repeatable, and he has consistently made excellent contact so that looks real too. I will further elaborate on this from a few different angles.

I’ll start with the same approach as with Tesoscar, to disaggregate total production into two buckets of batted balls and plate outcomes (K, BB, HBP):

This vividly illustrates the absurdity and the unsustainability of Rowdy’s raucous 2018 debut. With a .391 BABIP and .300 ISO, his batted balls were literally off the charts. Conversely, with a 3% walk rate and 29% strikeout, his plate outcomes were off the charts on the opposite end. (The percentile ranks refer to the 244 players with at least 750 PA from 2018 to 2020).

Not surprisingly, in a more extended run in 2019, his batted balls regressed significantly, albeit still to a very strong level, well above the average for quality MLB regulars. Likewise, his plate outcomes improves significantly, but still ended up quite poor.

Thus coming into last year, in almost a full season equivalent of 482 PA, Tellez had posted a 100 wRC+. Anywhere else on the infield, that would be very good, but it’s not going to cut it for a plodding first baseman. Despite his BABIP entirely normalizing to .287, his batted ball outcomes were still outstanding, around the 80th percentile of regulars. But a 6% walk rate against a 28% strikeout rate suggested a “AAAA slugger” type whose holes could be exposed by major league pitching.

And then came the 2020 breakout, a 133 wRC+ albeit in just 127 plate appearances as an injury further truncated an already shortened season. The natural instinct is to dismiss such a thing as a small sample size outlier, but those are typically driven by unsustainable batted ball outcomes. But unlike the 2018 run, what sticks out is that in 2020, Rowdy’s batted outcomes—while very good—were actually a little below his 2018-19 baseline. Even if we were entirely dismiss September 2018, it’s still only a slight bump on 2019’s batted balls. This is not your typical breakout that merits inherent skepticism.

We can see this in the above “waterfall” chart that bridges his 2018-19 baseline to his 2020 production. Has he maintained the same rate of production on balls in play, his wRC+ would have been six points better. The improvement in plate outcomes added a whopping 27 points, as he went from near the bottom to well above average. And then finally, he got another 12 points from a more favourable mix: with lower value plate outcomes making up only ~25% rather than 35%, it created room for more higher value batted balls.

The upshot of all this is that the 2020 Tellez breakout comes down entirely to improving plate outcomes, and barring an unforeseen decline or collapse in batted ball production, the extent to which he can be a regular or better MLB first baseman thus turns upon his ability to maintain this ability.

Despite is coming in a small sample, Expo45’s survey suggested it appeared replicable. In addition to the data he went over, I would add the survey I did four years when Justin Smoak had a similar contact driven turnaround in plate outcomes. There’s some regression, but more of it tends to stick than not, and that’s all Tellez needs to be a productive regular (or at least everyday strong side of a platoon if he has trouble with lefties).

Beyond the contact, I wanted to see if there was any evidence of a change in approach at the plate, or anything that might look unsustainable, so I pulled the pitch level Statcast data.

The first thing that stuck out to me is that Tellez is a reasonably aggressive hitter, with a 49% swing rate in the baseline 2018-19 period, compared to about 46% overall. That’s mostly the case early and ahead in the count, less so as the count progresses and he’s closer to league averages. In terms of balls and called strikes, he appears to have roughly an average eye, his ratio of the two right around the league mark. As noted, he struggled with contact, missing on 31% of his swings.

In 2020, Tellez’s swing rate went up a point to 50%, but there’s nothing remarkable there. What did standout is that when he took pitches, relatively few were called strikes. His balls/called strikes improved from about 34%/17% to 35.5%/14%, a net 4% increase. Given that he swings quite a bit, and there wasn’t a change in batted ball results so pitchers shouldn’t have been avoiding him more, it seems that either he developed a really good eye...or maybe perhaps benefitted from some calls going his way. This is very plausible as an area where some backsliding occurs.

Finally, while his contract rate took a big jump across the board to almost 77%, one area in particular had a big spike. With no strikes, he had a 73% contact rate (vs. 70% in 2018-19). With one strike, 74% (compared to 67%). But with two strikes, it soared to 83% from 71%. That saved Tellez a lot of strikeouts. Perhaps he simply developed an excellent two strike approach. But the small sample and outlying nature of compared to other hitters (overall, there’s almost no difference in those splits league wide) suggests he was fortunate and due for some regression here too.

Conclusion: Bullish. The strikeout rate is likely regress, but it’s buttressed by the fundamental improvement in contact rate. Perversely, there might even be a little room for improvement in batted balls, so I’ll take the over on the projections and put down a .265/.325/.510 median baseline.