Let’s be reasonable about Joakim Soria

On Tuesday evening, Joakim Soria blew another lead. I don’t know, there’s no nice way to say something like that (not that I have to say something nice - when have I ever?). He blew another lead. To date, he’s blown seven saves this season. Since last year, only Santiago Casilla (who coincidentally also blew a lead/save on Tuesday night) has blown more saves than Soria.

Soria has 14 blown saves. The next closest Royals pitcher is Kelvin Herrera with six. Joakim the the Royals career leader in blown saves with 34, fifteen more than Aaron Crow. Now for what it’s worth, Soria is sixth all-time in innings pitcher by a Royals reliever, and by the end of this season he’ll likely be in 4th (needs just 11 more innings). Before his contract is due up, he’ll likely pass Doug Bird for third all-time as well. Blown saves are effectively an accrual stat (you can’t get two in one game), so naturally a guy with more innings has more chances to blow a save.

Since FanGraphs began tracking blown saves in 2002, the “leader” is Francisco Rodriguez, and he’s one of the greatest closers of all time. Even Mariano Rivera is in the top 11. With 42 blown saves in his career, Soria is 19th since 2002. With save opportunities come blown saves. I’m not saying anything new here, but saves are a bad stat to begin with.

As Keith Law wrote in his book Smart Baseball, saves essentially ruined an optimal bullpen. Back in the 60’s, Jerome Holtzman unknowingly shaped how managers dictate their bullpen strategy when he created the save stat. Instead of using their best relievers at optimal times, they save their best reliever for the last three outs, or the Zach Britton rule. That’s just bad journalism right there...

Saves and blown saves aren’t the worst stat around, though they are bad, since saves/blown saves are decent proxies for higher leverage situations. And pitchers who blow saves or collect a save, often are blowing the lead late in the game. That’s something Soria has done this year but it’s also been something he kinda hasn’t too.

In high leverage situations, situations where the leverage is twice as much or more as average, Soria has been both good and bad, depending on the metric. By ERA he’s been bad, with a 9.20 ERA in 14.2 innings pitched, but by FIP he’s been dominant with a 1.84 FIP. That FIP is pretty close to being the equivalent of peak Wade Davis in those spots. That’s one of the best relievers in the league in high leverage.

You can see the disconnect here. Soria, by FIP, is not only one of the better relievers in high leverage, but one of the better relievers this year overall. He has the fourth best FIP of 2017 on that leaderboard, but also the worse ERA by some distance. On a defensive independent measure, Soria has done his job as a pitcher looking for events he can control, but when you include batted balls and his defense, Soria has been terrible.

There is the disconnect too that’s been sparking arguments against those saber-friendly folks like myself, and those still clinging to the old metrics of the game (like the save). What’s a pitchers job when he pitches? It’s obviously to get outs, and through not getting outs, not allow runs. We know that. But we also know that pitchers have basically no control of what happens to a ball when it gets put into play (though they do have control roughly of where the pitch is thrown; pitches down the middle usually go for hits). So which do you judge a pitcher by? I chose to use FIP because everything about DIPS Theory just makes sense to me to evaluate a pitcher by, and by DIPS Theory, Soria has been good not in high leverage situations, but all year too.

Reliever ERA is also shady to begin with. First from a formula standpoint, ERA of course uses earned runs extrapolated over nine innings. Relievers of course don’t pitch nine innings. They hardly ever pitcher more than one inning. Some of them only face one batters at times. Yet, ERA is cited to make them even with their starting pitcher counterparts, which doesn’t seem like a fair comparison. League average starter ERA is always going to be higher than the league average reliever ERA.

Of course, another inherent flaw of ERA is the first letter: earned. Take Tuesday night, Joakim Soria was charged with four runs, but he wasn’t on the mound for three of them when they came around, and Mike Minor was charged with the blown save to boot. Minor doesn’t get dinged for Soria’s runs (though he does for one of his own he allowed) despite letting them all in on one swing. That seems reasonable I suppose, but it’s still a silly system, and that’s without going on in length about how a pitcher can commit an error and not get hit for an earned run from an error he committed.

This is a long-winded way of saying that ERA isn’t particularly great for relievers, and Soria is absolutely getting bashed for his ERA, with no credit from what I’ve seen, for his pretty dang good FIP. You may not believe it, but he has the 7th best FIP among qualified relievers this year and the 12th highest fWAR. He’s actually putting together a nice year from a peripheral standpoint but... yeah the ERA, if you want to put more stock in that, isn’t great.

The overall problem I think is Soria’s usage. From a runs allowed standpoint, he hasn’t been good, and the situations he had been used in has been probably not the most ideal. Take it from a pure ERA standpoint: 3.96. The league average for relievers is 4.09, so he’s just slightly better than the average. Right next to him is Mike Dunn at 3.95. Dunn is on a playoff contender, he’s not just deadweight soaking up innings for a losing team. Dunn has been sort of a jack of all trades for the Rockies relief corp, pitching in the fifth, sixth, seventh, eighth, and ninth for Colorado. He’s allowed the same ERA as Soria, but he’s not someone that Rockies fans are calling for pitchforks each time he allows a run. He’s a cromulent reliever (from a runs allowed standpoint) and he’s not being thrusted into high leverage spots or creating them when he allows a runner to reach (due to the inning he’s pitching in; two on in the sixth is different than two on in the 8th/9th).

When Dunn enters an inning on average, the leverage index has been 0.81, just slightly under an average situation. When Soria enters begins an inning: 1.38 LI - 38% higher than an average situation. When Dunn exits the game: 1.14. When Soria exits the game: 1.63. Now maybe Soria is creating those higher leverages by allowing runners, but again baserunners in the 6th/7th aren’t equal to baserunners in the 8th/9th.

Soria by runs allowed is a fine enough middle innings guy, but not necessarily someone you want in high leverage situations, because when he does allow a run, it’s like the worst time to do it. I mean, there are no good times to allow a run, but a run allowed in a low leverage situation is different than one allowed in a high leverage situation, but ERA counts them the same.

Maybe it’s not fair to blame a pitcher for blowing a situation they are put in. If you put Chris Young in the 9th inning with runners on, yeah, he’s probably gonna give up some runs. That’s not on Young fully because he shouldn’t be put in those situations. That’s on the manager for putting someone in a spot they don’t belong (remember Cheslor Cuthbert helping costing the Royals a game because of him being put in at second?). No, I don’t believe that fully, but I do think that a manager should shoulder some of the blame.

Still, I think Soria is having a good year because I think FIP is the best evaluation of a pitcher we have. Let’s say you want to go with the 50/50 method of WAR and combine both runs allowed and FIP, giving the pitcher half blame for balls in play. That would give him a 1.1 WAR, still the second best on the team, just 0.3 behind Mike Minor using the same method.

Soria needs to take a break from these innings and it looks like he’s due to possibly take a break overall since catching the intercostal strain bug that’s been going around. So when comes back, who better should the Royals put in in those high leverage spots? That was a question that should have been answered by July 31st...