How Cole Got His Groove Back

Cole Hamels is back, ladies and gentlemen. What did he do to bounce back from a tumultuous 2009 season (and April 2010)?


Just as he did nothing to contribute to the swift descent following his impeccable 2008 in which he won the World Series MVP award, the Cole we have watched through 13 starts (12 if you exclude his rain-shortened outing in Atlanta on June 1) has similarly remained constant.

Sure, he added a cut fastball but it has been his worst pitch according to the pitch type values found on FanGraphs. He has actually reduced the use of his three other pitches in favor of the cutter and he still finds himself 11th in the National League in SIERA (among pitchers with at least 65 IP). In his latest start against the Boston Red Sox in which he allowed only one run over seven innings, 65% of his pitches were fastballs and 27% were change-ups. In fact, the fewer cutters Cole has thrown, the more successful he has been.

Hamels has beefed up his strikeout rate, averaging nearly one per inning and about one more per nine innings than in ’09 and ’08. His walk rate has hurdled above three per nine but it hasn’t meant much since he is stranding nearly 83% of base runners. The big key to his success — and this should come as no surprise if you have been reading this blog for a while — has been a regression in batting average on balls in play (BABIP).

In 2008 when he had tremendous success, his BABIP was .270, much lower than the traditional .300. In ’09 when he was painful to watch pitch, his BABIP was .325. This year, which has thus far met his ’08 and ’09 seasons somewhere in the middle, his BABIP is .308. Since pitchers don’t have much control over BABIP, the regression is simply due to Phillies’ defenders turning batted balls into outs and good ol’ fashioned luck. The extra strikeout per nine has been icing on the cake, as is his newfound dominance over left-handed hitters.

What else could have contributed to his success?

I noted after his second start in April that he may be toying around with release points but Pitch F/X guru Harry Pavlidis was helpful in reaching the conclusion that the machines were calibrated differently as the two starts were at two different ballparks. The calibration, though, seemed to only affect vertical measurement. Hamels has shifted his release point towards the left-handed batter’s box as the following graphs will illustrate:

The release point hasn’t made much of a difference though because hitters are making the same amount of contact and swinging and missing at similar rates as last year.

Season Contact% SwStr%
2006 72.3% 12.8%
2007 73.9% 13.6%
2008 76.9% 11.5%
2009 75.2% 11.9%
2010 75.6% 10.7%

Others will note the increase in velocity, as his fastball hit 94 MPH or higher 46 times in his most recent start in Boston. While it was great to see, his fastball hasn’t been like that all year as the velocity chart from FanGraphs illustrates.

For the third straight year, you can chalk up most of Hamels’ performance to BABIP luck. Just imagine how good he’ll be when his HR/FB rate (over which pitchers also have little control) regresses down from 16.5% to the more appropriate 10-11%. He has the potential to reach Nolan Ryan territory if he can kick the Twilight habit.

Leave a Reply



  1. Phillies Red

    June 16, 2010 11:59 AM

    Great stuff as always, and I’ve been looking forward to this article ’cause I’ve got some questions about this BABIP situation.

    First, according to fangraphs, Hamels’ FB velo is up about 1 mph so far this year over last year. Are we really to believe that this uptick has nothing to do with lowered BABIP? I thought I recalled something from insidethebook about how power pitchers have lower BABIPs. I don’t think that Hamels has become a power pitcher, but his velo over the last 7 or so starts (looking at the fangraphs charts) seems to be moving towards 92+ mph, and that has corresponded to lower BABIPs in May and June. Any reason for us to draw a connection here?

    Second I have a question about BABIP in general. Looking at Hamels’ BABIP splits by month over the course of the last two years, I’m surprised by how consistent his BABIP is within a given year (aside from this year, where he has been up and is now way down). In 2008, he only had one BABIP month over 300, but last year he only had one month below 310! The fact that Hamels was consistently low and then consistently high in BABIP really makes me wonder if there is something about Hamels that affects his BABIP. Velo, location, pitch selection? But my question is this: are pitchers in-year BABIPs consistent? Perhaps put another way, is a pitcher’s early season BABIP predictive of his season BABIP, or even later month BABIP? Hope that question makes sense (btw, this seems like it would be really useful to know for fantasy baseball).

    Great blog, always enjoy stopping by.

  2. Bill Baer

    June 16, 2010 12:13 PM

    Yeah, very high-K pitchers have tended to have significantly lower BABIPs. See: Ryan, Nolan.

    We wouldn’t expect Hamels, as good as he has been, to fall into that category. Even Randy Johnson had a career .302 BABIP. However, if he is increasing his fastball velocity while keeping his change-up velocity constant, he is likely to experience a lot more success as the velocity differential between the fastball and change-up is positively correlated with pitching success.

    As for the in-year BABIP, I’m not sure. I’ll forward your question to Matt Swartz and see what he has to say, since he has studied pitcher BABIP much, much more closely than I have.

  3. Phillies Red

    June 16, 2010 12:26 PM

    Thanks, Bill! So it sounds like we are left saying that Hamels’ good velocity of late (or really the FB/CH differential) might have helped with K-rate, but we really have no reason to think it impacts his BABIP.

    One final thought/question for you: Hamels is walking more guys, suggesting to me that he’s either throwing off the plate more, or a little more wild. Any thoughts on how this is impacting things like his K% (which is up, as you point out) or even his BABIP? I know these aren’t strict stats questions, but it’s an odd development for a pitcher that has been so consistent with his BB%, and I’m curious how it might be impacting his other numbers/outcomes.

  4. Bill Baer

    June 16, 2010 12:34 PM

    Overall, hitters are swinging 4% less this year than they did last year and Hamels has thrown 4% fewer pitches in the strike zone than he did last year according to FanGraphs.

    The higher walk rate may just be a function of opposing hitters. The scouting report on Hamels may be that he’s hittable in deep counts, especially if you can get ahead.

    Last year, when the batter got ahead in the count (1-0, 2-0, 3-0, 2-1, 3-1, 3-2) their triple-slash line was .344/.462/.603 compared to when Hamels was ahead (0-1, 0-2, 1-2) it was .202/.210/.303.

    At the moment, that’s my best guess at the increase in walks.

  5. Bill Baer

    June 16, 2010 12:54 PM

    Matt Swartz responds:

    I only have internet on my phone right now, but in short, I really
    doubt theres anything to that. Even with Hamels, he was evenly split
    in 2007 with 3 months of babip above .300 and 3 months below, and in
    2006 it was 2 mos. above and 3 below. So most years Hamels is up and
    down, except for his two most extreme babip years. I’m sure that
    injured pitchers might have repeated high babips and all, but healthy
    pitchers still have most of babip variation as luck– especially with
    a sample size as small as a month. This is selective sampling where
    the reader found two seasons with extreme babip for a pitcher who has
    had more extreme seasons than others and found a pattern of 2 and
    tried to make a conclusion. Even k-rate and bb-rate will be subject
    to plenty of luck in a given month.

  6. Phillies Red

    June 16, 2010 01:48 PM

    Awesome, thanks Matt and Bill! In fairness to myself, however, I never drew any conclusions from Hamels’ recent seasons. I was asking questions about how BABIP works, and how it works in-season. An observation on a (clearly very small) sample led to a question I didn’t have the technical expertise to answer, so I thought I’d ask those in the know. No need to point the dreaded “selective sampling” finger at me!

    So it seems like BABIP is really driven just by luck, except for high K pitchers, who seem to have slightly depressed BABIPs. I’m curious if the BABIP-is-luck conclusion is based on actual analyses of BABIP and how it works, or is driven more by the lack of the right type of data, data that would allow us to drill down to things like ball trajectory and velocity off bat, landing spot of the hit, defensive positioning, or defensive ineptitude (or conversely, efficiency). Sounds like these might be questions for a different day.

    Always a pleasure, and thanks for looking into this with/for me. Very interesting stuff.

  7. Bill Baer

    June 16, 2010 02:00 PM

    Matt wasn’t accusing you of being dishonest with the data, just that it was a selective sample. No malice there.

    Defense plays a big role in BABIP. Pitchers who have good defenses behind them will have lower BABIP. I threw last year’s team pitcher BABIP and defense UZR/150 into Excel and found a negative correlation of -.53. That means that as UZR/150 decreases, pitcher BABIP rises. The r-square is .28 which means that about 28% of pitcher BABIP is explained by the defense.

    I’m sure the other factors you mention — “ball trajectory and velocity off bat, landing spot of the hit, defensive positioning” — will provide further enlightenment but DIPS theory is pretty strongly supported.

  8. Chareth

    June 16, 2010 02:19 PM

    Thank god the luck is finally shifting. Now we just need this luck for the entire offense.

  9. BS

    June 16, 2010 03:14 PM

    Cole’s increasing K-rate has been really interesting I think. Actually, it’s even sort of perplexing. I wonder how sustainable it is.

    Looking at the numbers, Cole’s been in the zone a lot less this year, his swing rate and swinging strike rates are also down, and his contact rate is actually slightly up. If you add those up, shouldn’t we have expected his K-rate to decrease?

Next ArticleWhat's Eating Joe Blanton?