Neftali Feliz or Daniel Bard?

Yesterday at Baseball Nation, Jeff Sullivan came out with a pretty good response to a kind of silly article by Jon Heyman regarding the "cheapness" of the Boston Red Sox this offseason. Of course as most people know, the Red Sox sudden thrift has been largely overstated. Read Sullivan's article for the entire explanation, but the gist is that a lot of the players on the 2012 roster are holdovers who are getting paid more than they were last year.

However, there was one line of Sullivan's article that struck me as off:

Daniel Bard is supposed to take one of the rotation slots, and it's hard to see how he's a worse bet than Neftali Feliz.

Is Daniel Bard really as good as Neftali Feliz? And if so, does that make him as likely to succeed as a starter? My gut told me no (on both counts), but I figured it would be worth looking into. So, for the record, this is a response to a response. Considering the line (CBS -- Baseball Nation -- Me) the law of diminishing returns applies here. I'm going to guess several thousand people read Heyman's article, several hundred read Sullivan's response, and... well, if you're reading this, that means somebody read my response to the response. It's all good.

Ok, let's dive in:

Year	Tm	W	L	ERA	G	SV	IP	H	R	ER	HR	BB	SO	ERA+	WHIP	HR/9	BB/9	SO/9	SO/BB
Feliz	Texas	7	6	2.55	154	74	162.2	98	49	46	11	56	164	178	0.947	0.6	3.1	9.1	2.93
Bard	Boston	5	13	2.88	192	5	197	132	71	63	16	76	213	154	1.056	0.7	3.5	9.7	2.8

Bard and Feliz were called up within a couple months of each other in 2009, so their careers are nicely comparable. I considered taking W/L record and saves out of the stat line, because they are essentially white noise - Bard wasn't a closer, so of course he isn't going to get the same number of saves. Going straight to the other stats, Feliz has slight advantages is HR/9, BB/9 and ERA. However, considering that Arlington is a tougher place to pitch than Boston, Feliz actually has a pretty notable advantage in ERA+. Not that Bard is a slouch in any of these categories. His walk rate is high, but not exceptionally so, and  makes up for it with an extremely impressive K rate. Feliz has the slightly better numbers, but Sullivan is right - based on their careers so far, it's hard to say that Feliz has been so much better that you'd risk your Nolan Ryan rookie card on the fact he'll continue to be better.

Still, what stat tends to change the most when a pitcher moves from the rotation to the bullpen (and back again)? It's the strikeout rate. Since a pitcher knows he's only throwing for an inning or two, he tends to throw harder and hold back less than one who might have to pitch multiple innings. This isn't just conjecture, it's statistically verifiable. In 2011, American League relievers struck out 7.7 per nine innings, while starters struck out 6.6. In 2010, it was 7.4-6.6. In 2009, 7.6-6.5. Way back in 1993, it was 6.4-5.5. The trend holds. If we can expect a starting pitcher to strike out 10% fewer batters, that hurts Bard more, because his strikeout advantage will be slightly blunted, make the fact that he's done everything else slightly worse than Feliz even more notably. 

Next, the age factor. While Bard and Feliz were called up around the same time, Bard was allegedly born June 25, 1985, while Felix was allegedly born May 2, 1988. So Feliz is younger by two years, 10 months and 8 days. Feliz's 178 ERA+ through age 23, or Bard's 154, in a similar inning count, through age 26? At the very least, that should say something about Feliz's "upside," shouldn't it?

Then, there's the fatigue factor. Much was made about Feliz's slow start last season. Could it be that the 77 appearances in 2010 has tired him out? Possibly. It's also possible that it was a sample size issue. In the first half, through 34 innings pitched, Feliz had a 3.18 ERA, and even more distressingly, only 23 strikeouts to 18 walks. While that's ugly, it's important to remember that it was in only 34 innings. That's about 6 starts. Six! One bad month. In the second half, though? 28.1 innings, 31 strikeouts, only 13 walks, and a 2.22. Again, a small sample, but considering the rest of his career line, the first half last season is a bit of an outlier.If the problem was fatigue, he overcame it.

On the other hand, Bard's splits were in the other direction. In 26 appearances from May 27th to July 31st, he gave up zero runs. Zip. Zilch. At the end of July, Bard had compiled 51 innings pitched, giving up 29 hits, only 10 earned runs (a 1.76 ERA), 13 walks, and 49 strikeouts. On August 1, it all fell apart, giving up 3 runs in the 8th inning in a loss to Cleveland. From then on, Bard pitched 22 innings, and gave up 17 earned runs (a 6.95 ERA), walking 11, and striking out 25. The strikeout rate remained high, but everything else went in the wrong direction.

Of course, that's also a small sample size issue. Still, we're talking about which of the two is more likely to make the transition to starting pitcher. With that in mind, wouldn't you bet on the guy who pitched less well at the beginning of the year, rather than the guy who imploded after 51 innings? Particularly since this was the second year in a row where Bard's walk rate increased significantly in the second half. On the other hand, Bard had several more appearances in 2011 of over 1 inning, and he was extremely effective in them. Feliz had only one save where he retired more than three batters. In-game, Bard did not show any signs of fatigue that would lead you to believe that he necessarily couldn't handle a longer role.

So, we've gone through past performance, age and fatigue, and find that Feliz looks better than Bard. All this without getting into the most important distinction - Feliz was made a reliever because the Rangers needed a reliever, while Bard was made a reliever because he was so bad as a starter in 2007 in Single-A that the Red Sox never used him to start another game.

I don't have the starter/reliever splits, but take a look at the minor league numbers of Feliz: he made 53 starts from the time he signed with the Braves in 2006 until he was converted to the bullpen in 2009 (plus one more start on rehab in 2011 where he pitched only the first inning). In 2008, as a 20-year old, Feliz pitched 127.1 innings, with a 2.69 ERA, 153 strikeouts, 51 walks, and only three home runs allowed. This made him Baseball America's #10 prospect in all of baseball going into the 2009 season. The pitchers ahead of him were David Price, Tommy Hanson, Brett Anderson and Madison Bumgarner. He followed that with another dominant season in 2009 between Triple-A and the majors (not quite exhausting rookie status), and Baseball America made him their #9 prospect. He then won the 2010 Rookie of the Year award, as a closer.

Now take a look at Bard's minor league numbers. After being taken 28th overall in the 2006 draft out of North Carolina (where he was the #2 starter behind future Red Sox teammate Andrew Miller), the Red Sox challenged him with an aggressive placement in High A. He made five very poor starts, and they moved him down to the South Atlantic League, where things continued to go badly. How badly? In 22 starts across the two levels, he managed only 75 innings. He walked 78 and struck out just 47. His ERA ended up at 7.08. Fearful they'd drafted a bust, the Red Sox moved him into the bullpen in 2008, with outstanding results, and he's been in the pen ever since.

I don't think 22 bad starts in 2007 mean Bard will fail as a starter. In fact, I think the opposite - he's been a very good pitcher in the bullpen, and very good pitchers in the bullpen tend to be able to convert to starters or at least an acceptable level. Not always, but usually. Plus, Bard wants to start - the worst case scenario here is that he ends up going back to the bullpen. After working out as a starter, that should at least work to counteract those possible fatigue issues.

But as for the assertion that Bard is as likely to succeed as Feliz? Feliz is three years younger, has been a slightly better reliever in the majors, and was an extremely successful starter in the minors, which Bard was definitely not. Which one would you rather bet on?

Table made with TABLEIZER!