Are bargain-basement closers worth it?

For each of the last three seasons, since Mike Hazen became General Manager of the Diamondbacks, we have seen the same thing happen with regard to the team’s roster construction. The closer’s spot has been filled by a free-agent: someone who did not end the previous year as an established closer, but someone who did have prior experience of the role. The aim is clear: get someone who has had some success as a closer, without paying the premium cost associated with the position. That brought us Fernando Rodney in 2017, Brad Boxberger in 2018, and Greg Holland in 2019. But is this tactic working?

All three pitchers had their share of struggles. Rodney was awful at the beginning of the season, with a 12.60 ERA in April. But thereafter, he was solid, posting a 2.38 ERA the rest of the way. The following season, Boxberger had the reverse pattern. He gave us a 1.80 ERA through the first two months, but then had a 6.21 ERA, from there, until the point in September where he was removed from the closer’s job entirely, in favor of Yoshihisa Hirano. This year, Holland was similar to Boxberger. He didn’t allow his first run until May 3. But over the last month he has been awful, with an 8.59 ERA, more walks than innings pitched and an 0-2 record, with three blown saves.

It’s safe to say the results have been mixed. But then again, flawless closers are the unicorns of baseball. Virtually everybody blows multiple saves over the course of a season, and those are the games we tend to remember. So I thought it might be interesting to take a more objective look at how these three closers have performed, by comparing them to others in that role, across a variety of metrics. To find how average closers perform, I pulled the stats for the top 30 pitchers by saves recorded for 2017 + 2018, giving me a pool of sixty seasons of data. These were then ranked. The table below shows the metrics, and the number needed to qualify for each quartile.

For example, in raw Saves, the #15 of those sixty pitchers was Edwin Diaz, who saved 34 games for the 2017 Mariners. #30 was Ken Giles, who had 26 saves last year. #45 would be Seranthony Dominguez and his 16 saves in 2018. So, based on the numbers of saves put up by our closers we can see whether they were in the top 25% (well above average), 26-50% range (above average), 51-75% range (below average) or 76-100% (well below average), in each of the statistics.

It is interesting on its own, to see that the difference between a top-tier and bottom-tier closer in save percentage is less than 10%. Unlike in hitting, where failure is expected most of the time, you won’t be a closer for long if you convert fewer then four out of every five opportunities. This makes logical sense. A typical save situation involves working the ninth with a two-run lead. Even the most mediocre of relievers should generally be able to throw an inning without conceding two runs, which is what it takes to convert that opportunity, But it’s what happens in those fifth outings, which determines where you end up in the rankings: save the majority of them, and you’ll be up among the elite.

We can use these numbers to create a composite of the “average” closer. He is someone who would sit between the 30th and 31st-ranked pitcher in each category. This is what a season from Mr. Ave Closer would look like:

Closer: 63.2 IP, 26 Sv, 4 BS (86.7%), 3.19 ERA, 3.38 FIP, 10.74 K/9, 3.16 BB/9, 3.25 K/BB, +131.5% WP, 29 SD, 9 MD, 1.2 bWAR

If the guy you have does better than that across the majority of the categories, then you’ve got a good closer. If not...

Now, let’s see where the trio of Hazen-era closers sit in these standings. Both Rodney and Boxberger were part of the set of sixty, so we can see their true rankings, which are shown in brackets for each stat. For Holland, I’ve listed the position he would occupy, pro-rating counting stats e.g. saves for a full season, based on his numbers through the first game against the Yankees, conveniently the two-thirds point of the season. Those categories are marked with a * in the chart below. I’ve color-coded the cells based on which quartile they fall into. Top 25% = green background; 26-50% = blue; 51-75% = yellow; 76-100% = red.

Ouch. This more or less confirms what happened, with only Rodney making it through the season and finishing as closer. His numbers weren’t too bad: of the 13 categories, he was in the top half in seven, and given his low base cost ($2.75 million), you’d be hard pushed to argue he wasn’t good value at that price. Boxberger ($1.85m base), however, managed that in only two fields, and Holland ($3.25m) was below average across the entire spectrum. What stands out is the poor walk-rates provided by all three pitchers, none of them ranking higher than 49th out of the sixty players under consideration. The trio also finished in the bottom quartile across the board in strikeout to walk ratio and ERA.

This doesn’t necessarily mean it’s a bad strategy, at least theoretically. For there are plenty of more high-priced closers who turned out to be bigger (or, at least, more expensive) busts than Boxberger and Holland. We wrote five years ago about the mirage of the proven closer, and that remains largely true. Over the last three completed seasons (2016-18), only two men - Craig Kimbrel and Kenley Jansen - saved thirty games in each year. The volatility of relievers means it still doesn’t make a great deal of sense to throw money at one, purely for shiny save purposes. Not least because sometimes, the game needs to be saved before the ninth and so having your best reliever in a more flexible fireman role has logic to it

But the failures of the last two years do suggest either a change in approach - perhaps recruiting internally? - is needed. Or maybe it’s just that the team need to sign better candidates.