Cole Hamels is the Unluckiest Man on the Face of the Earth

A friend of mine has convinced both himself and me that he is the unluckiest person on the planet. He takes the poker game Texas Hold’em pretty seriously and has been playing at various card rooms including the casinos in Atlantic City. When I talk to him after a session, I usually get a report about how he was card dead for six hours or some fish donked off all his chips only to catch a two-outer on the river (translation: a bad player bet away all of his money due to poor decision-making only to be rewarded by catching one of the few cards in the deck that would help him).

I saw the misfortune in person when he and I, along with a mutual friend, made the 90-minute drive down to the Borgata in A.C. All three of us were seated at the same table by the floor manager and immediately dealt in. Both myself and my other friend folded our hands. The friend in question looks down at pocket aces, the strongest possible starting hand in Texas Hold’em. It’s been a few months, so I’m fuzzy on the exact details of the hand, but suffice it to say that a lot of money was put into the middle of the table before any of the community cards were dealt.

After the first three community cards were turned over (the flop), both my friend and his opponent’s chips were in the middle. My buddy flipped over his aces while the other guy revealed pocket jacks, a much inferior hand that will win only once every five times against pocket aces and only once every 20 times when the Jacks don’t connect with any of the first four community cards (flop and turn). The jacks, as expected, don’t hit one of the two remaining jacks to make a better hand on the flop or the turn. But, lo and behold, he hit his miracle two-outer on the river, the fifth and final community card. Mind you, this is the first hand my friend saw after sitting down at the table following a 90-minute drive. Suffice it to say, I then bought into his complaints of bad luck. (I realize that it isn’t even close to being on the list of worst bad beats of all-time; it’s just a case of confirmation bias.)

There is, however, one person on the planet whose bad luck could rival that of my friend.

Colbert Michael Hamels.

The 2008 World Series MVP, the lefty who had a 3.08 ERA during the ’08 regular season, is certifiably unlucky. It is said that Eskimos have a bunch of different words they use to say “snow” (which is not true); in Philadelphia, you need only one word for “unlucky” — Hamels. “Ah, poor John, he had no idea his house was built on top of an Indian burial ground. He got Hamels’d.”

Hamels’ unluckiness has, unfortunately, been perceived as a lack of or a decline in pitching skill. However, smart baseball people have found out what pitchers do and do not have control over on the baseball field. For habitual readers of the blog, you have seen me talk about this ad nauseam, but for those newer to Sabermetrics, let me briefly go over this again.

Pitchers have a lot of control over:

  • Their strikeout rate (K/9)
  • Their walk rate (BB/9)
  • Their ground ball and fly ball rates (GB% and FB%), including infield flies (IFFB%)

Pitchers have little control over:

  • The rate of batted balls are converted into hits and outs (BABIP)
  • The rate at which fly balls land beyond outfield fences (HR/FB%)

Let’s take a look at those stats. First, a chart of his strikeouts and walks per nine innings.

Note: K/9 and BB/9 averages are based on 2007-09 data from HEATER Magazine.

Essentially the same in 2008 and ’09. So far this year, he has had a large increase in strikeouts but also in walks. His strikeout-to-walk ratio is at its lowest in his Major League career (2.7; previous low was 3.0 in ’06). While the increase in walks is concerning, he is walking about as many batters as the average pitcher. The increase in strikeouts is encouraging since pitchers with high strikeouts tend to have a lower BABIP (see: Ryan, Nolan).

The following chart displays Hamels’ batted ball rates since 2006.

More encouraging signs: more ground balls and more infield flies. More ground balls means less fly balls which means less net home runs. More infield flies means weaker contact which means less hits (and less home runs).

As the above charts have shown, Hamels has either pitched similarly in 2009-10 as he did in ’08 or he has improved. Yet the results would not lead one to believe this. Have a look at his BABIP:

That is incredibly unlucky! Since pitchers have little control over BABIP, we expect Hamels’ to rest around .300 but it has been nowhere near that level since 2007. In ’09-10, it has been much, much higher than what we would expect despite his showing characteristics that would lead one to believe his BABIP would be lower than normal.

Hamels’ HR/FB% had been relatively normal, just a percentage point or two higher than the average. This year, however, he has been a bit unlucky on fly balls turning into home runs. It could be due to the small sample of 50 innings, but unlucky is unlucky.

Exactly how unlucky has Hamels been?

In 2008, when Phillies fans thought he was the next Steve Carlton, he actually out-performed his SIERA by more than four-tenths of a run. Last year, he under-performed his SIERA by nearly eight-tenths of a run, and has the same gap through 50 innings this year. All that, despite being essentially the same pitcher with the same stuff throughout most of his Major League career.

The verdict is in: Cole Hamels is the unluckiest man on the face of the Earth. Even my friend would agree.

As always, a gentlemanly doff of the cap to the great FanGraphs for the data.

Leave a Reply


This site uses Akismet to reduce spam. Learn how your comment data is processed.


  1. hk

    May 18, 2010 04:43 PM


    If you want to use the smalle sample of one game, the answer is easy…Cole was lucky on Sunday night. His BABIP on Sunday night was somewhere around .200, which helped him produce an ERA (again using the small sample of just Sunday night’s game) of 2.70.

    While I think you were using Corey Hart’s fly ball in an effort to refute Bill Baer’s point, you actually reinforce his point. If Cole can have some more good luck like he did Sunday night and produce more games where his ERA is 2.70, the good luck will help even out the bad and produce an ERA for the entire season closer to the expected 3.70 level.

  2. Ryan

    May 19, 2010 07:04 AM

    Doodler –

    Way to take a comment completely out of the context of the original article, and discussion within the comments.

    No one is ragging on Howard, what was being discussed was that he rolled a groundball that wasn’t by any means crushed, and essentially, as you said, he fought off a good pitch, and was lucky that the ball was just out of the reach of the ONE fielder on the left side of the infield.

    Again – there, the pitcher makes a good pitch, hits his spot, appeared to catch Howard looking for something else, and pitched into the strength of his defense (inside hoping Howard pulled the ball into the shift IF he made contact) and what is the result? A weakly hit groundball, that was inside outed that just eluded the fielder.

    Is it a good idea by Howard? Sure… did he have control over if that ball was AT LaRoche or just past him? Nope… all he was trying to do was get a piece of that ball because, as you said, he appeared to be looking for something else.

    Fortunately for him, it found a hole. Unfortunately for the pitcher, it found a hole. No one is ragging on Howard. It was a decent approach – obviously better than a K, but it still required some luck because he didn’t exactly drive that ball.

  3. Chareth

    May 19, 2010 08:03 AM

    hk hit the nail on the head. I think we’re all measuring Hamels’ luck on an incorrect scale. A scale that was calibrated on the back of his 2008 regular and post-season performance, which I think we can all agree was buoyed at least slightly by luck.

    Right now he’s giving up about, what, half a run above what we’d expect given his peripherals? Calling him the “unluckiest” based on that seems somewhat exaggerated.

    Bill, as far as your comment about even 3 seasons not being enough sample size, aren’t you moving the goal posts a bit to suit your own purposes? If 3 seasons worth of data is too small, what does that say about any analytics you produce with less data than that?

  4. Bill Baer

    May 19, 2010 08:31 AM

    I haven’t argued about any goalposts. I was challenging Heather’s presumption that something fishy is definitely going on if Hamels’ bad luck continues.

    Here’s when statistics become reliable:

    50 PA: Swing %
    100 PA: Contact Rate
    150 PA: Strikeout Rate, Line Drive Rate, Pitches/PA
    200 PA: Walk Rate, Groundball Rate, GB/FB
    250 PA: Flyball Rate
    300 PA: Home Run Rate, HR/FB
    500 PA: OBP, SLG, OPS, 1B Rate, Popup Rate
    550 PA: ISO

    And yes, “unluckiest man on the face of the Earth” was intentional exaggeration.

  5. Chareth

    May 19, 2010 09:39 AM

    1) Heather’s question was perfectly logical and valid — assuming Hamels’ numbers to continue to skew for the remainder of the season; asking if there might be another contributing factor to consider does not mean she thinks anything is “fishy” with your analysis. Under normal circumstances, one would think 2 seasons worth of data should be more than enough to smooth out any luck factors. Which brings us to…

    2) Those fangraphs numbers were specifically derived for hitters, so in a sense you didn’t really answer my question, unless you are implying that they are compatible for pitcher regression as well. (The original study with derives these numbers no longer exists as referenced by your link.)

    So let me ask the same question more directly. How much data do you feel you need for a pitcher to smooth out luck factors? Earlier, you implied it could be as high as — or even higher than — 3 seasons.

  6. James Ammon

    May 19, 2010 12:59 PM

    There is an article in the 2010 Bill James’ Gold Mine that may have a bearing on Hamel’s declining success. It wasn’t about Hamels per se but was about how a pitcher’s success rate declines over time if he throws alot of change-ups. That when there is a great K/BB rate, low batting avg., great WHIP but a declining success rate with alot of HRs, this is indicative, according to Mr. James, of a change-up pitcher. Hamels throws more change-ups than any other pitcher in baseball, 34% in 2008, 30% in 2009, 25% so far this year. Although maybe his best pitch, if hitters know it is coming, they are sitting on this pitch more even though he is throwing it less. Less in his case is still too many. The two things that most often happens when a hitter swings at a change-up- he either misses it completely when he guesses wrong or hits it hard when he guesses right. What may be needed is more curves and cutters. A change-up is only effective when it comes as a surprise. Just a thought.

  7. Michael E

    May 19, 2010 02:26 PM

    Apologies if I should have learned this in SBMTS 101…

    I’m relatively new to this blog and my knowledge of Sabermetrics is limited – but I’m working on it! The Hamels debate, however, made me wonder something that might add something in terms of assessing his situation re: luck vs. numbers.

    What, I wonder, would a statistical analysis of a pitcher reveal if BABIP was evaluated just for those at-bats immediately after they give up a hit? Might this be a way to evaluate their tendency to avoid “big innings” or exhibit “mental toughness” statistically? (I’m perhaps working too hard to straddle the worlds of numbers and feelings, but that does seem the flavor of the comments on this particular post, right?) Because it seems that a lower post-hit BABIP would suggest a pitcher who “corrects” himself, whereas a higher one might suggest a player who gets into trouble. (I imagine there would be some sort of mean or avg that indicates what is a major league avg in such a situation.) Also, I wonder if this post-hit BABIP could be ratio-ed with standard BABIP to determine, well, something.

    But maybe the counter here is that this sort of thing is unnecessary because it is essentially contained within the standard BABIP.

    But I’m curious anyway. Anything to this? Or is my question itself a redundancy?


  8. Ryan

    May 20, 2010 09:24 AM


    Really good post – and I think there are some measures to what you are talking about.

    Situational BABIP gets tough because of sample size. If you start to really break down given situations – babip after a walk, or babip after a hit – you start to limit the amount of data you have and start to limit the meaningfulness of that data.

    Now – there are numbers that you can look at to say “ok, how is he doing with runners on”. One of the more common is the strand rate or the LOB% as fangraphs labels it.

    One of the BIG arguments about JA Happ last year was that his success (low era) was due largely to the fact that he stranded a HUGE number of runners. In 2009 Happ stranded 85%, the league average for strand rate is 72%.

    This takes us right back to Luck vs controllable outcome. Does Happ throw better from the stretch? Does he really focus with runners on base? Or did he just get a higher % of balls in play converted to outs in those situations.

    That brings me back to Hamels and the question of “how is avoiding the big inning” or “how is he doing after a hit/walk”. The babip numbers in that situation may tell some of the story – but here is his strand rate for 2010 – 81%. His career rate IS better than the league average at 75%, but that number indicates that Cole is actually doing better right now with runners on base.

    The number that stands out to me in terms of the culprit of the big inning is the HR rate. He’s giving up hr’s (both per 9 innings and per fly ball) at a MUCH higher rate. Now we’ve argued before if that is a product of poor pitches, or poor luck.

    The sabr community says pretty clearly that most pitchers give up hr’s at a pretty similar rate per fly ball. Meaning that most pitchers give up one home run per every 10 or 11 fly balls they surrender (Hallady hr per fly ball – 10.1%, eaton 11.5%, hamels 12% you get the idea).

    Only this year – Cole is giving up a MUCH higher % – 17% of the fly balls he gives up are ending up over the fence, which has lead to him surrendering almost half a home run per game more than his career average.

    I know this post was long – but here’s my take.

    In terms of the big innings, and the reaction to a hit or a walk – I think there is some validity in taking a look at babip after an event (hit, bb, hr, whatever) but you have to be careful, because the more you narrow it down, the wilder the variation may be in babip – and the less meaningful the data may be.

    But if you look at other indicators, he appears to be throwing better than the league average with runners on (very high strand rate vs. league average – may be having good luck there). Instead the focus appears to be that MANY MANY more of his fly balls are landing over the fence – many more than his career numbers, and many more than the league average numbers.

    IF Cole has some outings where the ball stays in the park – despite his goofy BABIP, his numbers are going to continue to improve.

  9. hk

    May 23, 2010 08:50 AM

    For those of you who are still following, Cole has been luckier in the small sample size of two games since Bill posted this piece. His BABIP has dropped from .348 to .330 and, not coincidentally, his ERA has dropped from 4.29 to 3.92. If you are in the camp that thinks “Cole has turned it around,” re-read the article. You will learn that Hamels’s pitching has not been much different from 2008 to 2009 to the first 1/4 of 2010, only the results (as measured by ERA) have.

Next ArticleChecking in with Brett Myers