Utilizing Archie Bradley

Small sample size applies, to be sure, but the move of Archie Bradley to the bullpen has gone, not just as well as could be expected, but far beyond that. Let’s start with some numbers:

• Line: 9.1 IP, 5 H, 0 R, 0 ER, 3 BB, 11 SO, 0.00 ERA
• 1.54 FIP (best among ten pitchers on team with 5+ innings)
• 0.7 bWAR (best of any Diamondback, pitcher or position player)
• Win Probability: +56% (highest pitcher - #2 is Robbie Ray at +23%)

By just about any measure save fWAR (which we’ll get to in a bit), Bradley has been the Diamondbacks’ best pitcher so far this year, despite being neither a starter nor a closer. The question would be, are his skills being used in the most effective way to benefit the team? In the quick recap yesterday, Sprankton floated the idea of him moving back to the rotation, and I think it’s worth chewing over the possible roles for Bradley in more detail.

1. Back to the rotation

The main reason fWAR does not rate Bradley so highly is workload. At 9.1 innings, he has thrown not much more than half as much as the two players ahead of him. Ray and Taijuan Walker are both in the 16-17 inning range, and with Bradley having thrown 39 pitches yesterday, that’s probably going to stand as we complete a third circuit round the rotation tomorrow. If Bradley is as good a starting pitcher as he has been a reliever, it makes obvious sense to leverage this by getting as many innings from him as possible. But there’s the nasty little word, “if” there.

There’s no doubt Bradley is throwing harder this year. His average fastball (per Brooks, so less subject to the changes in measurement we’ve seen in 2017) has been 96.3 mph, almost two mph faster than in April last season (94.5). However, this may not necessarily all be due to him pitching out of the bullpen. For in March this year, when he was competing with Patrick Corbin for a job in spring training, Bradley’s velocity was still up, at 95.9 mph. There’s a counterpoint, however: those were in shorter outings, with none of his appearances that month any longer than 55 pitches.

One thing which might support moving him back to the rotation, is Bradley’s development of a new pitch, the cutter. It’s a pitch he used back when he was in the minors but was abandoned entirely once he reached the majors, Brooks not seeing it for a single pitch before this year. In 2017, Archie has thrown it about 15% of the time, giving him an extra weapon, as he said, it’s “another option, another weapon to where I'm not just having to throw heaters in there.” Here’s how it has altered his overall pitch repertoire between last year and this (major-league appearances only, for both)

• Fourseam: 2017 58% vs. 2016 63%
• Sinker: 8% vs. 6%
• Change: 0% vs 7%
• Curve: 19% vs. 24%
• Cutter: 15% vs. 0%

The change has evaporated from his repertoire so far, but one imagines that’s more likely due to his relief role, where he doesn’t have to face batters more than once. It’s probably still in Bradley’s back pocket, though it was never a great pitch. With the cutter offering an additional option, it could help improve on Archie’s career 5.18 ERA as a starting pitcher. The question would then be, who is he going to replace? Patrick Corbin would be the obvious choice, with his K:BB ratio of 7:7, and having been in the bullpen at the end of last year. But Corbin’s ERA has defied that ratio, sitting at 2.81, and you would be hard pushed to justify bumping a sub-three ERA from the rotation.

2. Closing time

The major advantage here is obvious: it gets Fernando Rodney out of the role. We’re barely two weeks into the season, and I’m already terrified every time he takes the mound. Given he has been perfect in save opportunities, going 4-for-4, that’s quite an achievement: it’s possible Rodney is simply executing the greatest trolling job in the history of Diamondbacks baseball. But I’m more inclined to suspect this house of cards is going to come crashing down sooner, rather than later.

The advantage of using Bradley as a closer is it will generally help ensure his talents are not squandered in low-leverage situations. By coincidence, Nate Silver was writing today about an alternative statistic to the save, the “Goose”, and the article included the chart below, which shows how leverage varies, depending on the inning and the margin of the score. As you can see: the traditional “save situation” does tend to have a high leverage, indicating performance here is most important - albeit less so as the margin grows.

Picture from fivethirtyeight.com

As the chart above shows, a one-run margin in the eighth inning is more important than two runs in the ninth. The idea of Bradley coming in, and closing things out over two (or even slightly more) innings is an appealing one. However, traditionally of late, closers do not work more than one inning. The last man in the National League to post a 20-save season, while averaging more than 1.2 innings per appearance (without making any starts), was Vladimir Nunez for the 2002 Marlins. And even there, only five of those saves involved him getting more than three outs. The last closer with even a dozen “long saves” was back in 2007 - future D-back J.J. Putz, then a Mariner.

The problem would be that we’d likely need a reliable alternate closer, because the longer Bradley went, the less likely he would be available the next day, should there be another long save opportunity. For if you look at long saves on zero days’ rest, no-one has managed even a handful of those in a season since 2002. That was Byung-Hyun Kim for Arizona, and he may represent a good template for Bradley’s use. He appeared 72 times that year, pitching 84 innings and notching 11 long saves - though also blew five save chances.

3. As you are

Alternatively, we can just keep using Bradley as he has been, providing a bridge in close games when our starter is unable to go deep - yesterday’s game against Los Angeles is a good example. However, this is not without its limitations. Even the closest of games in the sixth inning is relatively low-leverage, compared to the same situation in the ninth inning. There’s also the risk that you bring in Bradley for the sixth inning in a close game, looking for him to go two or three innings - and the offense then scores runs, taking the game out of being high-leverage, and effectively “wasting” Archie’s arm. There’s less certainty in the middle innings, than at the end.

But in general, you also never know how future games are going to play out. There will always be a risk that you’ll burn Bradley on Monday, only for a more important situation to appear on Tuesday. Or conversely and equally possible, you don’t use Bradley on Monday, only for Tuesday and Wednesday then to be blowouts. However, the D-backs should also try and avoid shooting themselves in the foot. His debut came in a game where Arizona was down by six when he entered, and in his other appearance against San Francisco, Bradley came in with the D-backs trailing by five. I hope we’ll see better utilization going forward.

It’s something manager Torey Lovullo seems to understand, when quizzed about the usage: “We may change the inning based on what he’s doing. We’re very well aware of what you’re saying. We know he’s had some quality outings and we want that to continue. It’s just going to be in any format possible to help us win a moment.” It’s true the bullpen is still evolving in terms of roles, with only Rodney apparently holding a specific job. All seven relievers have recorded eighth-inning outs: while Bradley has most PA there to date (15), J.J. Hoover (14) and Randall Delgado (11) aren’t far behind. Things may well settle down a bit as we go forward.