Can Marco Estrada suppress $26-million of contact?

It's funny the difference a year makes. Just over one year ago the Blue Jays acquired Marco Estrada with a projected salary under $5-million in his last year before free agency, a transaction that overall was not well received by Blue Jays fans. Fast forward to this offseason, and the consensus had swung 180 degrees to Jays needing to re-sign him, even if it took a multi-year deal at over $10-million annually. And that's exactly what the Jays did, first giving him a qualifying offer before agreeing to a two year deal for $26-million.

The difference of course was between a disastrous 2014 campaign in which Estrada gave up 27 home runs in 107 innings, and a brilliant 2015 campaign in which he posted the 6th best ERA (3.13) among qualified starters in the American League. While his results were better than ever, as he was along the best in baseball at managing contact, there were some red flags in terms of metrics that try to measure underlying performance. Estrada posted career worsts in xFIP (4.93) and SIERA (4.64), which are suggestive of a very poor to replacement level starting pitcher. And while some pitchers can sustainably outperform these metrics, coming into this year Estrada had actually significantly underperformed them.

This wisdom of this deal thus comes down to Estrada's ability to continue to manage contact well. He doesn't have to replicate 2015, if over the next two years he turns in 350 innings with an ERA around the league average of ~4.00, it should work out okay. That's the question I want to examine today: how much contact can we reasonably expect Marco Estrada to suppress?

More and more reliant on contact

Estrada's strong historical strikeout and walk ratios were one of the major selling points for me last winter, but instead of regressing towards those stronger numbers, he instead extended the decline (note that BB* is non intentional walks and hit by pitches, and stats are as a starter only):

Estrada has gone from a pitcher who added a lot of value on these two outcomes, to one who is essentially league average (actually slightly below). The difference between 2012 and 2015 in terms of expected run prevention is 1.08 runs per nine innings. Even using an average of 2012-14 instead of the 2012 peak year as the comparison results in an expected difference of almost 0.75 runs (4.43 v. 3.69 FIP).

The focus today is contact so I don't want to stray too far into it, but simply put, that's a massive change that requires a lot of contact management to offset. In my view, most analysis of Estrada has more or less glossed over this, essentially giving him a free pass on it. If there's no further decline over the next couple years, it can be worked around. But if the decline continues, I'm not sure all the contact suppression in the world will make this contract work.

There's a secondary knock-on effect as well: it leaves him more exposed to outcomes on contact, since ~75% of hitters are putting the ball in play rather than ~70%. That's fine if he's managing contact like he did in 2015, but a major issue if he's giving up a pile of home runs like in 2014.

The paradox of Estrada and managing contact

This brings me to the first issue: while Estrada was brilliant at managing contact in 2015, previously he had been well below average at doing do. In 2015, batters hit .251 and slugged .450 when putting the ball in play against Estrada, compared to a league average of around .330 and .525. But from 2012-14, the other years when Estrada was primarily a starter, batters hit about .312 off him while slugging .554 when putting the ball in play. He succeeded in suppressing ~20 points of batting average, but coughed up 45-50 points of isolated power.

It is not surprising that Estrada would suppress batting average, since fly ball pitchers and Estrada particularly induce more popups which are akin to strikeouts as basically automatic outs (see here for detail). But It's worth noting that historically that gap was 20 points of average, not 80 points like in 2015 - and that 20 points is more in line with Estrada's pop-up differential of 2-3% compared to league average. That's really about it in terms of managing contact - as a group they don't allow fewer line drives, or fewer home runs compared to their total of fly balls that aren't popped up.

So we've got a historically poor contact manager, especially when it comes to power, who had one brilliant season, and we know that contact outcomes tend to be very volatile from season-to-season to begin with. Absent strong proof of fundamental change, investing $26-million on a career year is a very bad idea, which is why I was very strongly in favour of letting Estrada walk as a free agent for most of the year.

How much contact can Marco suppress, if Marco can suppress contact?

There is reason to be optimistic about positive changes being a catalyst for sustainable improvement, which were outlined well in an article at Blue Jays Plus a couple months ago. In 2015, Estrada pitched up more often, which combined with longer running transition to a higher release point to generate more spin, fly balls and ultimately weak contact. Critically, that's the same formula used by Chris Young, who has excelled at managing contact to the tune of a career .247 BABIP. This would seem to bode well for Estrada going forward.

To get a sense of how this could carry over to Estrada, let's take a look at exactly what batted ball outcomes Young has succeeded in preventing. This data is from 2006-08 and 2012-present when he was healthy and an established contact manager:

Across MLB, when balls are in play, roughly 22% go for singles, 7% for doubles or triples (grouped together since the difference is mostly about runner speed and parks, not pitchers), and 4% for home runs. Stunningly, Young has allowed singles at a rate about 25% below league average. This is a function of his obscene 8.4% career popup rate, which is almost 5% above the MLB average, and not co-incidently, Young's gives up about 5.5% fewer hits on balls in play. As we saw above, Estrada has benefitted from this same effect, although not to the same degree since he doesn't get as many popups. But he should be able to maintain something in the 18% range.

But notice that his contact management does not carry over to the power outcomes, and in particular the extra bases hits that stay in the park. Young has given up doubles and triples at almost exactly the league average rate. If there is any ability to allow fewer hard hit balls (in the air), it's entirely offset by allowing more of them. Estrada's career has been interesting in this regard: overall, he's a little better than average at 6.7%, but with big fluctuations. 8.5% in 2012, down to 5.2% in 2013, up to 7.3% in 2014, back down to 5.4% in 2015. And that's with Carlos Gomez and Kevin Pillar anchoring the middle of the outfield and tracking down balls in the gap. So there's little reason to think Estrada will be able to add value here over the long term, and just regressing towards his average can be expected to add about 0.45 to his 2015 ERA.

The home runs are more complex. Young has given up about 13% more home runs than the MLB average, but that's actually a very good outcome for a pitcher who allows over 50% more fly balls. His FB/HR rate of about 8% compares very favourably to the MLB average which tends to be 10-11%. This has been worth about half a run to his ERA over his career. A large part of Estrada's success in 2015 was a HR/FB of 8.3%, and if could sustain that it would be a gamechanger.

But Young has pitched in very friendly home parks: Petco, Citi Field, Kauffman, Safeco. At home, his HR/FB ratio is 6.9%, whereas on the road it's been a more pedestrian 8.8%. Jered Weaver is another extreme-fly ball pitcher with a Young-type contact management profile, and he too has a huge home/road split: 6.4% HB/FB at home compared to 9.7% on the road. Matt Cain is another. When you go down the list and look at pitchers preventing home runs over time, it seems ballpark is a fundamental element.

Moreover, Estrada had a reverse split in 2015: he allowed 7.8% HR/FB in 2015 at home, versus a more normal 9.5% on the road. Over the past 5 seasons, Blue Jays pitching has averaged 1% higher at home on the road.  Given that the best contact managers are at or above 9% HR/FB on the road, as is Estrada's 2015 road performance, the 2015 home number looks like an outlier. Even discounting Estrada's pre-Toronto/pre-transition struggles, my estimate would still be around 10% going forward (10.5% at home, 9.5% on the road).

Taking a stab at a projection

So let's put this all together and make an estimate for 2016, assuming Estrada's K and BB stabilize around 2015 levels, as Steamer does and the batted ball outcomes from above:

I've calculated an expected ERA based on the run value of all outcomes (xERA- is like wRC+), and included FIP as a point of comparison. This base case would have Estrada as almost a perfectly league average pitcher, despite his FIP slipping even further. As long as he stays healthy, his contract is decent value in this market.

If you want to make adjustments for yourself, a 1% change in a metric has approximately the following ERA effect: HR/FB = 0.23 in ERA; 1B% = 0.19, 2B/3B% = 0.28; K or BB = 0.10. So if Estrada can prevent home runs like he did in 2015, there's strong potential for an above average ERA in the high 3.00s - not 2015 good, but quite good. But if it was a mirage, the contract could look pretty ugly.

* * *

The next two years will be something of a test: can an extreme fly ball pitcher consistently manage home runs in a hitter friendly park and division. In an ideal world, I wouldn't want the Jays to be committing $26-million to be finding out (and I still think Estrada is best suited to playing in KC, Anaheim, Seattle, San Diego etc), but with the cost of free agent pitching being what it is, it's not the massive overpay that it looked like it might be three or four months ago.