With their season, for all intents and purposes, over, eight games out of first place, on the last day of a 9-21 August, the Boston Red Sox decided to let struggling reliever Julian Tavarez start in what was essentially a meaningless game against the Blue Jays. The Red Sox won 6-4, but Tavarez was unimpressive: three innings, three runs, three strikeouts, three walks. Nevertheless, the Sox kept him in the rotation and the results over the rest of the season were impressive. In September, Tavarez made five starts, going 3-0 with a 3.52 ERA. Have the Red Sox stumbled upon a hidden gem? Could Tavarez, who was not very effective as a reliever, make for a good starter?
That’s the very question that was raised a few weeks ago across Red Sox Nation, and on the Sons of Sam Horn message board. One poster, “Noah,” gushed, “Tavarez is ridiculously durable and has crazy movement on all his pitches.” Eric Van, a statistical consultant for the Red Sox and a frequent SoSH poster, wrote:
It’s very unusual for a guy to pitch better as a starter than as a reliever, so just how good a starter Tavarez will be in the long run is still unclear after this small sample. But I think the following conclusions are probably correct:
1) Compared to the average pitcher, his perfromance [sic] as a starter is better than you’d expect based on his performance as a reliever.
2) Given how much more valuable starters are, he should be in some team’s rotation next year (i.e., top 6 starting candidates) rather than in someone’s pen.
What is unclear is just how good he is. At the extremes, he may be good enough to be a fifth starter on a top contender, or he may be only good enough to be a sixth starter type (making 20-25 starts and pitching long relief the rest of the time) on a team with a lesser playoff aspirations.
Is it possible for a mediocre reliever to become a good starter? It doesn’t really make sense. And indeed, in 79 career starts prior to 2006 Tavarez had a 5.12 ERA, a run-and-a-third higher than his ERA as a reliever. According to research by Tom Tango in The Book, a pitcher should be about .80 to 1.00 runs better in relief than he is as a starter. Given all that we know, is there really even any point to looking into whether or not Tavarez might be a special case based on a 30-inning sample? Since there’s more article below, the obvious answer is yes, there is.
Are all pitchers affected by the starter/reliever difference in the same way?
Our own Steve Treder looked into this question with lots of detail. To summarize, he found little difference between different types of pitchers going from starting to relieving, except for one group: pitchers who don’t allow many home runs. To quote Steve, pitchers who allow home runs at a low rate “show significantly less change in their walk and strikeout rates when relieving than the total group shows.”
Tavarez is an extreme groundball pitcher, only allowing .71 home runs per nine innings in his career. It is possible that this would make him well-suited for starting.
First of all, Tavarez has never had a particularly good strikeout-to-walk ratio, but that problem is magnified in relief, when it’s often necessary to get a strikeout in a tough spot and to keep runners off the bases. Starters can make up for their mistakes more.
As a starter, Tavarez had more of an opportunity to work out his mistakes and he didn’t press as much. In fact, after allowing eight home runs in 58 innings prior to September, he allowed just two in 30.2 innings in the last month of the season. His ground ball-to-fly ball ratio jumped from 1.82 to 3.82. Could it be that Tavarez was more willing to trust his stuff when allowing a hit or two didn’t mean that he had just allowed a run or was going to be replaced?
It’s possible. Of course, it’s also possible I’m reading way too much into a very small sample. Steve’s research was good, but not definitive, and again, we’re looking at just over 30 innings to draw our conclusions about Tavarez. If this was all the evidence we had, I wouldn’t be writing this article.
Two decades ago, Bill James invented a nice little tool for comparing players which measures how “similar” any two players are based on their statistics. The system has since been modified to improve projections. But similarity scores, both types, rest on statistics, and in some case, numbers just don’t tell the whole story. If we were to look at pitchers whose stat lines read most similar to Tavarez we wouldn’t find anything interesting—just a bunch of mediocre relievers.
But what if instead we were to isolate pitchers who throw like Tavarez instead of those that perform like him? Much of the optimism around his potential as a starter centers around Tavarez’s nasty sinking fastball and decent secondary stuff. How do guys like that do? Using Scouts Inc. scouting reports gathered prior to the 2005 seasons, we can take a look.
Scouts Inc. tells us that Tavarez is a right-handed pitcher with a “6” two-seam fastball, a “5” changeup, a “4” curveball, a “5” slider, and “6” control (all on a nine-point scale, where nine is the best and one is the worst—most categories average around a six). So what I did was look for right-handers who exclusively throw a two-seam fastball rated between five and seven, a changeup between four and six, and whose best breaking ball was rated between a four and a six, with five to seven control. There were 21 such pitchers in 2006, other than Tavarez.
Of those pitchers, nine were starters, a pretty healthy percentage (suggesting that guys with Tavarez-type stuff are indeed pretty good). The starters and the relievers both had ERAs 1% better than average (just about Tavarez’s career average, by the way, which he equaled in 2006), lending more credence to the idea that Tavarez might experience less of a penalty moving into the starting rotation (since these two groups of pitchers are supposed to be similar in ability).
If Tavarez could be average as a starter, well, that would be worth quite a bit. It would certainly be good enough for a fifth starter, and worth double-digit wins pitching for the Red Sox (only two Boston starters had that many this season).
But again, we’re looking at a pretty small sample. So I’m going to do what’s rational and cut it down to one.
And here comes Lowe
Of those 21 pitchers with similar repertoires to our hero, one immediately stands out as a great comp for Tavarez. He too pitched for the Red Sox, and he too was a reliever, and he too was converted back to a starter in the last games of a lost season (on a side note, I was there for his first start as a Red Sox), and he flourished. His name is Derek Lowe.
Lowe had come to Boston with Jason Varitek in 1997, both traded for Heathliff Slocumb. After making some starts in 1998, he settled in as a reliever and a very good one at that. In a hectic 1999 season, Lowe stabilized the closer’s role and took it over permanently in 2000. That year, he racked up 42 saves with a 2.56 ERA and an All-Star appearance to boot. The next season, he started to fall apart mid-year, and after blowing three out of six save opportunities in the month of August, he was deemed too fragile to close (every Red Sox fan is familiar with the “Derek Lowe face”—a look of total helplessness and fear punctuated by a certain disgust and pain known only to long-suffering Red Sox fans and players). So in the last month of the season, the Sox gave him a chance to start. And in three starts, Lowe pitched 16 innings with a 1.13 ERA.
The next season, the Red Sox put him in the rotation for good. After one month, Lowe had gone 4-1 with a 2.04 ERA and one no-hitter. At the end of the season, he was the owner of a 21-8 record and a 2.58 ERA in 219.2 innings pitched.
Derek Lowe also happens to be one Tavarez’s best comps, if not the best. Tavarez has a “6” two-seamer, Lowe has a “6.5.” Tavarez has a “5” changeup; Lowe’s is rated a “6.” Tavarez has a “5” breaking ball, and so does Lowe. Tavarez gets a “6” for control, while Lowe gets a “5.”
If Lowe did so well, couldn’t Tavarez? The biggest difference between the two, despite what the scouting reports tell us, is that Tavarez does not have Lowe’s control. Derek Lowe has walked 2.6 batters a game in his career; Tavarez 3.5. But Lowe is a very good starter; maybe Tavarez could simply be okay? That would make Red Sox fans happy enough.
There is no definitive proof that Tavarez will be a successful starter. Reason and statistics tell us that he won’t, frankly. But there is much circumstantial evidence that Tavarez could be as good a starter as he was a reliever. Of course, a starter with a 4.50 ERA is worth a hell of a lot more than is a similar middle reliever. So much so that Red Sox fans should indeed be intrigued by his potential performance in the rotation, and that the Red Sox front office would be smart to let him battle for a slot in the rotation in Spring Training. Who knows?
They might just have their diamond in the rough.