It’s probably the Canadian in me talking, but I have always been a Jeff Francis fan. The University of British Columbia product was selected 9th overall by the Colorado Rockies in the famed Moneyball draft, and rewarded the team with 11.8 WAR in his first five big league seasons before succumbing to the injury bug. Now, the 31 year old finds himself without a job a month before spring training despite posting 4.6 WAR over the last two seasons.
Why is he unemployed? Well, the 85 MPH fastball has something to do with it, but so does the fact that he’s traditionally been a guy whose results haven’t matched his peripherals. Other than his 36.2 inning cup of coffee in 2004, only once in his career has his ERA been lower than his FIP. For his career, his FIP is a respectful 4.40, but his ERA is 4.78, a 38 point gap that helps shed light on the perception of Francis as a guy you only want on the hill if you have no other options. The last two seasons have been especially brutal, with an ERA almost a full run higher than his FIP – the main culprit being, as usual, a .317 BABIP. His inflated hit rate has been fueled in part by a significant drop in his infield fly rate, which has hovered at close to 6% over the last two years after being over 10% early in his career.
There is nothing overly unusual about a pitcher posting back-to-back seasons with BABIP’s of .320 and .316, or having a career mark of .310. These numbers are well within the expected random variation of our sample of major league pitchers. However, Francis has changed quite a bit since his early days in Colorado when he was throwing 89-92 MPH. He is now down in the 85 MPH range, and is now a full-blown pitch to contact guy. Last season, only 16.20% of Francis’ plate appearances ended in a walk or strikeout, ranking him 93rd out of 94 qualified pitchers. I decided to do some digging to see if just maybe these two factors might be related.
The first thing to do is find pitchers who had a high BABIP over an extended period of time in order to see if a high BABIP can have predictive value. For example, I looked at the top ten pitchers in BABIP from 2003-2008 with at least 1,000 innings. I then compared this to their 2009 performance. I repeated the process for the next two sets of seasons as well, with a BABIP cutoff of .308, giving me a sample of 26 player instances. In the original 1,000 inning sample, these pitchers had an average BABIP of .312 over a total of 26,612.2 innings.
Among these 26 instances, there were three pitchers who retired voluntarily, five who did not pitch in the big leagues the following season due to injury or ineffectiveness, and one pitcher who had moved to the bullpen, so they didn’t offer much help. Our remaining pitchers posted a BABIP mark of .309 in their test seasons, which totaled 2,153 innings. If you want to weight by innings pitched, that number drops to .303, but introduces a selection bias issue. Most of these pitchers are borderline MLB pitchers, due in part to their BABIP struggles, and those that continue to post high BABIP’s are going to lose their job in the starting rotation. Conversely, those that have a BABIP closer to league average are more likely to keep pitching, barring injury of course. Since this means that weighting by innings pitched isn’t really an option, I just made sure there weren’t any guys with a .450 BABIP in five innings spoiling the average.
This is by no means a slam dunk, but it does demonstrate that the league trailers in BABIP over a 1,000 innings sample have generally continued to post above average BABIPs. There is no guarantee that Francis will again have an above-average BABIP, as he is still subject to random variation, but this shows that it also isn’t fair to automatically assume that he will regress to league average. His mean may be a bit different than the population’s mean.
Another quick thing to note is Francis’ IFFB%. The BABIP difference between an IFFB% of 6% and 11% is only a few percentage points, but when we are comparing a .300 BABIP to .315, every little bit counts. As mentioned above, Francis has posted IFFB% of 5.5% and 6.3% the last two seasons. Previous research has shown that there is a certain subset of pitchers that have some ability to induce a high number of pop-ups year after year. The most common characteristic among these pitchers is a low GB%, which makes perfect sense, and is why we see fly-ball pitchers like Ted Lilly and Jered Weaver among the leaders in IFFB% each year.
When we look at the worst IFFB% performers from 2002-2011 with at least 1,000 innings pitched, we find some interesting results. As expected, there are numerous groundball fiends like Brandon Webb and Derek Lowe. However, the rest of the list is comprised largely of pitchers with profiles similar to Francis: low velocity, above average ground ball rate, and high contact rate. Matt Morris, Paul Maholm, Livan Hernandez, Nate Robertson, Carl Pavano and probably the best comparable, Zach Duke, all struggle to induce pop-ups, and all have above-average BABIP’s except Morris, who clocks in at .297.
Of course, this would be meaningless if there were a bunch of pitchers like Francis at the opposite end of the spectrum. However, we can quickly eliminate almost the entire list of high IFFB% pitchers using just GB%. Only four of the top thirty pitchers in IFFB% have a GB% over 45%. These four are Roy Halladay, Roy Oswalt, Carlos Zambrano and Jose Contreras. None of these four compare to Francis as they all throw well over 90 MPH and are certainly not pitch to contact guys.
As I alluded to earlier, there is no guarantee that Francis will post a high BABIP and underperform his FIP, as random variation could easily push his BABIP under .300. There is some minor evidence, however, to suggest that his true talent BABIP may just be above .300. We have to offset these findings with the understanding that Francis has spent most of his career pitching in Colorado, a notorious hitter’s park, but his BABIP on the road has always been problematic as well.
For a few million, Francis is still a decent enough back-end starter, capable of eating innings and sparing a team from too many disaster starts. However, unless he finds a way to start inducing some easy pop-ups again – or wisely signs with a great defensive team that could mask his issues – he may once again fail to produce the value that his peripherals would suggest.