Jered Weaver broke one of the most advanced pitching stats

Jered Weaver is having a season to remember for all the wrong reasons. His 5.71 RA9 is among the worst in baseball. His 5.78 FIP is the second worst behind James Shields. The 33 HR he's given up is the third worst behind Shields and Josh Tomlin with 35. He is second to last among qualified players in strikeout rate with a paltry 12.2 percent, with only Martín Pérez being worse. On the bright side his walk rate is 6.4 percent, which is actually quite good.

You're all probably aware of why Weaver is performing so poorly. His fourseamer averages only 84 MPH and put into context, per Baseball Savant only knuckleballers Steven Wright and R.A. Dickey have a slower fourseam fastball. His twoseamer is over a MPH slower and has the lowest velocity among pitchers with at least 100 results. Weaver's lack of velocity has inspired many jokes, some of the best of which come from Jeff Sullivan of FanGraphs:

Weaver is still in the rotation because the Angels don't have any better options, and because his stuff obviously wouldn't play much better in the bullpen. The only way he gets people out at all is by his excellent command and tremendous deception, not unlike Mark Buehrle. Buehrle however, is left-handed and was also excellent at controlling the running game. Also, believe it or not, he was a historically good fielding pitcher. Weaver, on the other hand, is a below average defender who does a mediocre job at controlling runners. Among the 108 pitchers who have pitched at least 200 innings since 2015, Weaver ranks in the middle of the pack at stolen base percentage, according to the Baseball Reference Play Index.

The two most advanced pitching stats out there, Contextual Fielding Independent Pitching (cFIP) and Deserved Run Average (DRA), have some shocking things to say about Weaver’s season. His horrendous 151 cFIP is the third worst all time, behind John D'acquisto in 1976 (159 cFIP) and Tommy Byrne in 1951 (171 cFIP). As bad as that is, it is believable. Weaver is extremely homer prone and can't strike anybody out, a dangerous combination. He pitches in a pitcher-friendly stadium, where he also happened to have had more starts.

What is truly shocking is that Weaver has an incomprehensibly bad 8.10 DRA! An 8.10 DRA! As recently as last week it was over a run higher than that! Before that huge improvement, Weaver was on track to having the worst DRA ever for a single season among pitchers with at least 100 IP. He is currently the 19th-worst in MLB history.

One of the reasons for this, in addition to what was mentioned regarding cFIP, is the Angels actually employ a pretty good defense. They rank seventh in the league by DRS and ninth by UZR. Remember, Weaver has Andrelton Simmons playing behind him, whose defense has been worth 15 DRS despite having played in only 102 games.

The biggest contributor to raising Weaver’s RA9 by 2.39 runs is how hard he gets hit. One of the components of DRA is something called Hit Runs. It adjusts a pitcher’s RA9 depending on the quality of contact against him. Not only are his 27.6 Hit Runs the worst in the league, it’s almost twice as high as the runner-up, Aníbal Sánchez. The worst ever since 1950 is Larry Dierker in 1969 with -19.2 runs. Weaver is beating that by roughly 50 percent. That seems awfully suspicious. To claim that Weaver has been that “lucky” in terms of how much he’s been knocked around sends up a red flag.

It comes as no surprise that he has been knocked around as well. Weaver leads the league with 191 hits given up. I already covered the home runs, but he’s given up 38 doubles, which is the seventh worst in baseball, as is his 35.5 percent hard-hit rate. The 73 extra-base hits he’s given up is second only to James Shields (it appears Hit Runs is docking him for all that big-game time).

One aspect of Hit Runs is that it favors pitchers who pitch a lot of innings and get a lot of outs. Weaver has done the opposite of that. Due to his struggles, he gets pulled early from games after facing a significant number of batters: in other words, he doesn’t pitch a lot of innings for the number of times that he has started. Furthermore, having faced 673 batters in only 154.1 innings means that he’s not getting a lot of outs.

The Baseball Prospectus version of WAR, known more commonly as “WARP,” uses DRA. As a result, Weaver has a staggering -5.0 WARP! Before his big DRA improvement that I mentioned earlier, it was at -7.1! For the sake of comparison, Weaver has -1.1 bWAR and -0.3 fWAR. Currently, the all-time worst season by WARP is Nick Martinez with -7.5 in 2014.

Let me give those numbers some perspective. The all-time worst season by fWAR, which uses FIP, is Phil Ortega in 1965 with -1.9 fWAR. The all-time worst season by bWAR, which uses RA9, is Steve Blass in 1973 with -4 bWAR.

[You’ve probably heard of Blass because of something known as “Steve Blass disease.” It’s not actually an illness, just a slang term for a pitcher who all of the sudden has 20-grade control. In 1972, Blass had a great 2.88 RA9 and was second in Cy Young voting. The following year he had a 9.95 RA in 88 IP and was basically done after that. He went from an 8.2 percent walk rate to 18.5 percent in one year. For Red Sox fans out there who remember Daniel Bard, he was said to have developed Steve Blass disease.]

The worst ever season by bWAR for a position player is Jerry Royster in 1977 with -4 WAR. That season he hit .216/.278/.288 for a 45 wRC+, and he was a disaster defensively. More recently, Adam Dunn had a -2.9 bWAR season in 2011 as a DH who hit like a pitcher.

All of this is to say that Weaver’s DRA and WARP don’t look right, nor do the similar numbers of other pitchers historically. Weaver has many similarities with James Shields this season, and Shields has “only” a 6.26 DRA with a -1.6 WARP.

Jonathan Judge of Baseball Prospectus is one of the main inventors of DRA and cFIP. He rolled them out last year and has made improvements to DRA this year. My first thought about why Weaver’s DRA is so bad was because of his 28 percent groundball rate, which is the lowest in the league. However, Judge addressed the groundball bias in DRA a little over a month ago. So that’s not it.

Failing to find any answers, I reached out to Judge via email. He confirmed that Weaver was so bad that he was indeed breaking the DRA model. He said that he was aware of the problem and was working on adjustments. I suspect the big one week improvements in DRA and WARP were those adjustments, but they’re still really bad.

Looking at all the facts, it’s possible that Weaver is an 8.10 DRA pitcher. It could be that he and pitchers like him are actually so much worse than we used to believe. On the other hand, there are only fourteen examples of a pitcher posting a season of -4 WARP or worse. It could be that they’re all outliers that are straining the limits of what DRA can teach us.

Models are never perfect, and that includes any model you can find in major sciences such as physics and chemistry. That doesn’t mean you can’t learn from them. It especially doesn’t mean you should discard a validated model because of some weird results. I’ll be interested in seeing if Judge addresses extreme cases such as Weaver’s.

Best case scenario for Weaver probably still leaves him headed into retirement. He’ll be a free agent at the end of the season, and considering his ineffectiveness, it is hard to believe there will be any takers for even the minimum. He’d be walking away having turned in some great seasons for the Angels, and nothing can take that away from him.