Position Roundtables: #4 Starter

Jeff · March 29, 2005 at 9:25 pm · Filed Under 2005 Roundtables 

Jeff: No. 4 Starter: Gil Meche

Gil Meche is younger than I am by four years. Already, he’s had
big-league success, major surgery, a long recovery process, another
full season as a starting pitcher and an almost comically up-and-down
season that saw him return to the minors, get knocked around, and
subsequently return to Seattle and a modicum of success.

To engage in a bit of hyperbole, Meche’s 2004 was a bit like the
Portland woman who was driving along just fine, then lost control of
her car and flew off of the Morrison plunging 60 feet to the river
below. No one would expect a dramatic recovery, especially after the
SUV sank 55 feet to the bottom of the Willamette. But that’s exactly what

Eventful life to this point (and overblown analogies) aside, there is
really no reason to expect 2005 will be any less turbulent for Meche.

He’s got the stuff to be successful, but the track record of pitchers
with this type of history is spotty at best. Hence, PECOTA expects
Meche to break down and slip from the rotation at some point, making
three relief appearances mixed in with 21 starts.

There is good news. At 26, Meche’s body should more resilient than
that of an older player. Contract-wise, he also makes relatively

Finally, he turned in a fairly strong post-All Star break run that saw
Meche lower his ERA from around 7.00 to 5.01, turning in a 3.96
September figure that included an 116-pitch shutout of the Red Sox.

I’m not expecting Meche to turn in 30 starts, but I do hope that he’ll
fall into the category of “good when healthy.” His September record
offers hints that this just might happen.

Dave: There’s a great scene in Ocean’s Eleven where Clooney and Roberts, now
divorced, are having a biting back and forth conversation about her
willingness to replace him with a casino boss.

Ocean: Does he make you laugh?
Tess: He doesn’t make me cry.

Gil Meche is baseball’s Danny Ocean. Clooney’s character commanded
the allegience of men and made the hearts of women swoon. While
having experienced some past successes in life, he also spent a
significant amount of time in prison, losing out on some of the things
he’d achieved at an earlier age. While being rehabbed in a
penitentary, he’s theoretically better than new. Of course, with his
new found freedom, he then proceeds to rob a casino and end up back in
jail. The entire plan was brilliant, carried out to perfection, and
was a joy to watch, and in the end, he still ended up back where he
started, in prison.

Gil Meche makes us laugh, but he also makes us cry. He’s a ball of
potential, a 94 MPH fastball, a biting curve, dominating half seasons
mixed into a career of twists and turns. At his best, he’s a good
pitcher and was a borderline all-star in the first half of 2003. At
his worst, well, he’s back in Triple-A, working on his mechanics,
wondering why his ERA is over 7.00. He’s shown the ability to throw
strikes, miss bats, and keep the ball in the park, just not at the
same time, at least not for more than a few weeks in a row.

Gil Meche is a good talent. He’s not a good pitcher, not yet. He’s
flashed his potential, but he needs to put it together for six months
in a row. Championship clubs are not made by occassional brilliance,
and right now, that’s what he’s given us; a flash of success mixed
with a dab of failure is the recipe for mediocrity. What Meche needs
is a healthy dose of consistency. If he can continue to pitch like he
did in the second half of 2004, well, we’ll finally be able to call
him a good pitcher. Until then, he’s just an ex con robbing casinos
and losing his woman to Andy Garcia.

Derek: My only comment on Meche is the same comment
I have every time we discuss him: he didn’t turn into an ace after his demotion.
He chopped his walks dramatically, and that’s really the only thing he did.
Balls put into play against him turned into outs at a far greater rate than they
did before his promotion: combined with the drop in walks, it made his season
look like night and day, and leads to all kinds of hasty conclusions and rosy
predictions for this coming year.

Cutting walks is great, and it offers Meche the chance to be a good pitcher, but
until he can sustain that (and as Dave notes, sustaining any kind of positive performance
has been trouble), it’s only a chance. His starts are pulled from a grab bag, and his
2005 season … nobody knows. In any event, this is likely his last year as a Mariner, and
I worry that it will mean Meche is more likely to drive himself hard in starts and conceal
nagging injuries to try and get his starts and wins up, while the Mariners have a manager
who likes to work his starters hard and as an organization won’t be invested in his future at all.

Jeff: We always hear about consistency when discussing an up-and-down
player. The following rant is really unrelated to Dave’s points or
Meche, but is something I’ve been needing to vent my spleen about.

Why do people assume that becoming “consistent” means that the player
will be consistently good?

Sure, there are many talented but streaky players. There are also
players (like, hopefully, Meche) who just need to put a few things
together in order to make great progress. But there are lots of
players who can’t find consistently solid performance because their
true skill level just isn’t as high as the competition’s is.

Neifi Perez is consistent. Willie Bloomquist has been very consistent
since his hot-hitting cup of coffee. That hasn’t been a good thing for

Oh, and I lost my woman to Andy Garcia once. Once.

Jason: When looking at most players on the M’s, I feel I have a general idea what
sort of performance to expect next season. Sure, there are best- and
worst-case scenarios making up a sort of spectrum of performance, players
have huge breakot years (and horrible collapse years), and of course there
are players like Ichiro about whom it’s tough to guess (Derek’s touched on
this before with regard to PECOTA).

Then there’s Gil Meche. Darned if I know what he’s going to do next season.
He could work 200+ innings with a 2.95 ERA. His arm could fall off tomorrow
and never get properly reattached. He could start 30 games, post an ERA of
5+ and then bounce around the majors the next ten years looking for that
magical change of scenery to turn his career around.

While I fully admit that there’s a little bit of this to every player —
Player X could perform to the best of his abilities, Player X could suffer a
career ending injury, Player X could become a journeyman — for some reason
I can’t quite pinpoint, I feel strongly about putting “Gil Meche” and
“crapshoot” in the same sentence.

All that said… I’d much rather see him in the rotation than Aaron Sele,
Ryan Franklin or Felix Hernandez.


35 Responses to “Position Roundtables: #4 Starter”

  1. Paul Covert on March 29th, 2005 10:45 pm

    Given the Mariners’ position in 2005– probably a year away from contention, and needing to find out whom they can count on for ’06 when they’ll need a solid rotation– I’m glad it looks like Meche has a rotation spot. If he can establish himself, and if he’s willing to stick around for a reasonable price, then I can see him as a part of the next big-time Seattle team. I like him as a pitcher, and hope he can pull it off. But if not, we need to find that out and move on to somebody else.

    As for the peripheral-stats issue: My handy metric on that is (K-W-5*HR)/BFP, based on a mix of the DIPS and linear-weights theories. (The major league average is about -.06; zero, i.e. breaking even, is above average. See further explanation below.) In Meche’s case, his career by half-season looks like this:

    1999 2H: 47K, 57W, 9HR, 375 BFP (metric: -.15)
    2000 1H: 60K, 40W, 7HR, 363 BFP (-.04)
    2003 1H: 78K, 35W, 17HR, approx. 463 BFP (-.09)
    2003 2H: 52K, 28W, 13HR, app. 322 BFP (-.13)
    2004 1H: 41K, 29W, 6HR, app. 207 BFP (-.09)
    2004 2H: 58K, 18W, 15HR, app. 358 BFP (-.10)

    Given that hits on balls in play are unreliable as a skill indicator, except in very large samples, and given the severity of HR allowed as opposed to other outcomes, it is not clear to me that late-2004 Meche was better than early-2004 Meche. He certainly came back from Tacoma a different pitcher; but perhaps not a better one, because of all the HR he allowed.

    For that reason, one of the most encouraging signs this spring is that Meche hasn’t allowed a home run yet, in 13 2/3 innings (probably about 60 batters faced). At last year’s second-half pace he would have allowed two or three by now. The sample is small, of course, and so I’m still taking this with an Arizona-sized grain of salt. But if Meche can start keeping the ball in the park this year, while keeping the K-W differential under control, then he can become a very good pitcher indeed. If balls keep flying out of the yard on him, though, I’ll be content if the team at year’s end says, “Sorry it didn’t work out; better luck with your next employer.”


    Further explanation of formula used above: Given last year’s ML averages of .266/.333/.428, and thus about .389 TB/PA, the runs created formula gives linear weights of about .333+.389=.722 for a single, .389 for a walk, .333 for an extra base, and -.333*.389=-.130 for an out. Furthermore, the average team scores about .19 runs per out, so that making an out takes that much opportunity off the board. Finally, the average ball in play results in about .3 hits, .075 extra bases ((2B+2*3B)/H), and .7 outs. So I estimate the following average values:


    So a ball in play is almost a neutral result, a strikeout and a walk are roughly equal and opposite, and a HR is about five times as bad as a walk. I fully admit the roughness of the approximations, but for a quick and handy estimate of how well a pitcher’s done in a relatively short sample of performance, K-W-5*HR works pretty well.

    (Also, for the reasons described above, I prefer to think in terms of strikeout-walk differential (K-W) rather than ratio (K/W). I’d rather have a guy who strikes out 10 and walks 5 in a game, than a guy who whiffs 2 and walks 1. But I suppose I’m not likely to change the world on that one….)

  2. Taylor Davis on March 29th, 2005 11:33 pm

    “Balls put into play against him turned into outs at a far greater rate than they did before his promotion”

    Certainly part of becoming a good pitcher is making this happen…

  3. Rob on March 30th, 2005 12:19 am

    Or it is just luck…

  4. Adam on March 30th, 2005 12:42 am

    I still love to watch Meche pitch when he’s on.

    I still have a tape of Meche in Yankee stadium dominating from 2003.

    He had a pin point 2 and 4 seam fastball.

    He had a pin point change up.

    He had a pin point 12 to 6 curveball.

    He had a pin point slurve down in the zone.

    He was unhittable.

    Also, another amazing thing. He appeared to know how to pitch.

    He would start off the game throwing harder stuff, and then move onto his softer stuff almost like a mechanic.

    That lasted the majority of the first half of 2003.

    For whatever reason, they changed his mechanics…at about the same time of the collapse. He stopped swinging his hands over his head.

    It would be foolish to believe that was the biggest factor in his demise…I still don’t think it helped.

    While the slight mechanical change didn’t affect his volicity, everything else went to hell. Was it a coincidence? Was his rhythm thrown off so much it led to his collapse. Did he get overworked? Confidence?

    I don’t think we’ll ever see the 2003 first half Meche. Why? Because he has dwindled his pitches to 4 seam fastballs over the middle, and big curveballs.

    He may be a more consistent pitcher because of this, but his upside is much more limited.

    And this is why he’s such a question mark. Go watch those games from 2003 and you’ll see why he was dominating. It was truly amazing.

    You can get by in the Major Leagues with a 93 mph fastball and a nasty curveball…but not the way he was in the first half of 2003.

    It’s too bad. Maybe I’ll see something like that again with Felix.

    If he can ever get even that changeup back…he’ll be a whole new pitcher.

    Another example.

    Freddy Garcia, in 2001.


    Freddy Garcia, in 2004.

    Totally different pitchers.

    Why did he suddenly get so messed up? I have no idea. The potential in both players is off the charts…but they both could never find it again. So they had to adapt, their new version is “good” but will never be “great” …Hard game this baseball thing is.

  5. Bela Txadux on March 30th, 2005 4:54 am

    Thanks for the insight Adam. I, too, recall just how dominating Meche was, but I didn’t see/hadn’t heard the attendant mechanical change at the time he ‘lost it.’ Now, the word WAS that he was using the two-seamer after he came back up in later ’04, so perhaps he is inching back to a complete pitch mix, although I suspect your ‘two pitch pitcher’ comment is derived from observation. I continue to look for the Meche that was—because he was dominating at times. I’ve never given all that much credenced to his ‘mental mistakes’ which are not supposed to have been corrected by ‘a new approach.’ Mechanics matter, and if he lost his location due to a mechanical change, perhaps we’ll never see the Meche that was again. But unlike many who comment on him here, I believe that he _can_ dominate, and if he does the M’s staff looks a great deal better. Whether he will, who knows?

  6. backwardsk on March 30th, 2005 5:36 am

    ahhh yes…but Dave, Ocean wound up with Tess by the end of the movie. So what does that say about Meche? And what does that say about Andy Garcia?

  7. Dave on March 30th, 2005 5:38 am

    (Also, for the reasons described above, I prefer to think in terms of strikeout-walk differential (K-W) rather than ratio (K/W). I’d rather have a guy who strikes out 10 and walks 5 in a game, than a guy who whiffs 2 and walks 1. But I suppose I’m not likely to change the world on that one….)

    As always, Paul, great stuff. That’s why you’re Super Reader.

    But isn’t your preference stated above very context dependant? I’d argue that, given equal ratios, the preference for a high walk and strikeout pitcher or low walk and strikeout pitcher would depend on the surrounding effects, mainly his teammates and his ballpark.

    For instance, the 2003 Mariners turned 73 percent of all balls in play into outs and were absolute death to extra base hits. The tandem of the incredible outfield defense and Safeco Field’s extra base hit supression turned the Mariners into the ideal situation for a contact command pitcher. Ryan Franklin, obviously, is the case study, as he enjoyed tremendous success almost exclusively due to the performance of his teammates and ballpark. Having a high walk/high strikeout pitcher on a team that takes away outs on balls in play by such a large margin over the average serves to negate one of the main strengths of the team.

    I’d argue that in low offensive environenments such as the ones created by the 2003 Mariners, command and contact pitchers would be the way to go. The flip side would also be true; the Yankees, for instance, should be acquiring as many strikeout artists as possible, trying to limit the balls Bernie Williams and the fellow brutal Yankee defenders have to run down.

  8. ajp on March 30th, 2005 8:24 am

    I would also point out that I would much prefer a pitcher with a 6:1 K:BB ratio over one with a 10:5 ratio.

  9. Russ on March 30th, 2005 8:37 am

    Good write-up on Meche. He is hard to watch as when he is on, it looks so effortless and when he is off, nothing goes his way. I think he is another head case. I do think this is his make or break season. Either he settles in and becomes consistent (good, not great) or he never does get his poop together.

    From the “what its worth” file…

    Just a few moments ago, Colin Cowheard said he felt the Mariners are better than the Vegas line at 81 wins. He said that while we lacked any aces, we had a lot of good #2 pitchers. With what he felt are good(not great) pitchers along with the addition of Betre/Sexon to the existing Ichiro/Boone line-up, he thought we are very underated. His words, not mine.

    It is interesting that some national guys are taking a flyer on the Mariners. Obviously one takes all this with a grain of salt, especially when we remeber yesterdays conversation regarding the amount of time the national guys have to spend on each team.

  10. Evan on March 30th, 2005 8:44 am

    Paul’s preference for a counting stat rather than a rate stat means that we need more information to go with it. The K-W differential doesn’t tell us anything without also knowing innings pitched. The K/W ratio does tell us something if that’s all we have. A guy with a K/W of 10 is muh more likely to get a K than a walk. A guy with a K-W of 10 could be David Wells after one start, or it could be Aaron Sele at the end of the season.

  11. Paul Covert on March 30th, 2005 9:20 am

    Re. #10: Yes, Evan, that’s quite correct; I was aware of the simplification when I wrote it, but it was late and I’d rambled on pretty long already. (K-W)/BFP would be my preference, or (K-W)/IP if you don’t have the BFP data.

    Re. #8: Yes, AJP, that’s correct also.

    Re. #7: On the context-dependence Dave suggests– well, that’ll of course be at least somewhat true, but I haven’t worked out whether it would spoil the approximation. Let me do so now:

    2003 M’s (.727 DE): BIP = .273*(.722+.25*.333)-.727*(.32) = -.013. For an average team with about a .700 DE, it’s about +.017 (the .04 above involved a math error; sorry about that). So the difference between a very good defensive team and an ordinary one makes only a small difference here; it would take about 10 BIP for the difference in defense to make up for one extra strikeout by the pitcher.

    While I’m at it, let me also apply this to create an estimator of Defense-Independent ERA:

    If the pitcher allows every ball to be put in play, with an average defense, he’ll see a .300 OBP against and .375 TB/BFP allowed, so that the runs created formula gives .1125 R/BFP. Ignoring the relatively small contribution of baserunning, we also get (1-.7)/.3 = .233 IP/BFP, and thus a 9*.1125/.233 = 4.34 RA (per 9 IP). Since the RC formula assumes an average error rate, we can expect a DI-ERA of about .92*4.34 = 3.99 (round this off to 4.00 if you like).

    (For the 2001 M’s, this comes out as 3.18. But even a high-strikeout, high-walk pitcher– e.g. 10K, 5W, 30BFP– will still see a ball in play about half of the time, and a real-world high-contact pitcher about 80%– even Dan Kolb was only 83% last year– so that the difference between high-contact and low-contact will be about 0.25 or 0.3 ERA points additional advantage to the high-contact guy with a great defensive team.)

    Now, an extra K-W is worth about .35 runs (if you want to be more precise, use 0.9*K-1.1*W; that’s a little bit fairer to the contact pitchers, but in small samples the difference is marginal), and an extra HR allowed costs about 5 times that much; so for an average defensive team, we can estimate DI-ERA as:

    4.00 + 0.35 * (K-W-5*HR)/BFP.

    Hope that all makes sense….

  12. joebob on March 30th, 2005 9:46 am

    Re #2: actually pitchers have very little control over what happens to a ball once it is actually put into play, that’s the whole point of defense independent pitching statistics. There are some pitchers who are exceptions to this rule, generally ones that have a very extreme gb:fb ratio, but gil meche is not one of these pitchers.

  13. Paul Covert on March 30th, 2005 10:24 am

    Arrgh– of course I meant

    4.00 0.35 * (K-W-5*HR)/BFP. (Minus, not plus.)

    Also, another way to think of the difference between an average and an excellent defensive team is to note that, for an average team, it takes about (.39-.017)/(.32+.017) = 1.1 strikeouts to make up for a walk, while with an excellent defensive team it’s more like (.39+.013)/(.32-.013) = 1.3. For a poor defensive team it could be as low as 0.9.

  14. Joel on March 30th, 2005 10:27 am

    I blame Ben Davis for Meche and Pinerio’s bad starts

    When Meche came back from the minors Ben Davis was gone and he pitched better. We saw the same thing with Garcia where Ben Davis has this nasty habit of calling fastballs when the hitter is expecting fastballs.

    It’s just a thought but can anyone look into indivdual stats with a certain catcher and compare the stats

  15. JMHawkins on March 30th, 2005 10:33 am

    Are you really willing to give a pitcher credit for “missing bats” but not for having more BIP turn into outs?

    Or do you think there might be a connection between “almost” missing the bat and easy grounders and routine popups? Those line drives that happen when the pitcher doesn’t miss any of the bat sure seem hard to turn into outs.

  16. Dave on March 30th, 2005 10:42 am

    JM, this has been studied far more extensively than you can imagine. You can do a google search for “DIPS”, “Voros McCracken”, “Tom Tippett”, or “Batting Average Balls in Play” and you’ll be amazed at the amount of research and evidence that is available on the subject. The basic conclusion (and there are some minor exceptions and variations) is that a pitcher has little control over whether a batted ball is turned into an out or not. This is different than saying they have no control over a batted ball, because some pitchers certainly can induce ground balls or fly balls at significantly greater than average rates. The key, though, is that they almost certainly don’t possess an ability to turn those ground balls or fly balls into outs. That’s up to the defenders.

    And Joel, what your suggesting is widely available as catcher ERA. It is a flawed stat for various reasons, but it basically throws your theory out the window. Davis wasn’t the problem. For most of his tenure, the staff’s ERA was lower when he was catching than when Wilson was on the mound.

  17. Ralph Malph on March 30th, 2005 10:44 am

    #15: yes, you can give a pitcher credit for missing bats but not for getting more outs on BIP.

    The fact is that, except for a very few pitchers, there is no correlation from year to year in BABIP (batting average on balls in play). There is a correlation from year to year in K rate.

    It may not seem like common sense but it is a provable fact.

  18. Dave Paisley on March 30th, 2005 11:14 am

    Following up Joel #14 and Dave #16, Meche pitched only three games caught by Davis – April 10, 21 and May 2. He gave up only 2, 3 and 2 runs in those three starts. His worst early start was April 16, caught by Wilson, and his later meltdowns (particularly May 27) were Wilson games.

    The evidence points clearly away from Davis being the problem (and by that I’m not saying Wilson was the problem either.)

  19. Jesse on March 30th, 2005 11:40 am

    Dave, about the balls put in play, you mention there’s some exceptions, but are there really not that many pitchers who are good at, say, inducing pop ups? What about pitchers who break bats a lot like Rivera? Is that just observation not supported by statistics? It seems to me that if a pitcher has a lot of late movement on their pitches, that would mean that batters don’t hit the ball with the good part of the bat, that’s contact that is easier for the defense to turn into an out I’ll read up on DIPS directly, but I have a lot of trouble believeing that pitchers can’t affect the kind of contact batters get off of them at all.

  20. Kevin Wright on March 30th, 2005 11:54 am

    Personally what I want to see for next years starting rotation, Joel, Gil, Bobby, Felix and Rafael. Dominating starters start to finish. It would have the potential to be the best rotation in baseball!

  21. joebob on March 30th, 2005 11:57 am

    Off topic, but I have to say that I am starting to get ramped up for the season. Optimism chemicals are starting to flood my brain and I can’t help but feel excited about the Mariner’s chances this season.

    GO M’s!!

  22. Dave on March 30th, 2005 11:58 am


    We only have publically available data like infield fly % for 2004 (available at the the Hardball Times), so its hard to do a study on year to year correlation between types of balls allowed. For instance, last year, Mariano River certainly was an outlier. His line drive percentage of 10.8% was significantly lower than the league average of 17.7% and his infield fly percent of 30% was twice as high as the league average of 15.5%. Despite those huge differences, though, it only amounted to a three percent above average mark on balls in play overall; he had 72 percent turned into outs versus the league average of 69 percent.

    We don’t have enough evidence to say that Rivera does or does not possess the ability to deflate line drives and inflate fly balls, but we can look at his prior year BABIP’s to get a guess. His BABIP in 2004 was .282, but .299 in 2003, .250 in 2002, .275 in 2001, .249 in 2000, .215 in 1999, and .245 in 1998.

    Those numbers are pretty stiking, honestly. He’s consistently held batters way under the .300 mark, so it looks like Rivera’s cutter does give him an advantage on BABIP, perhaps in the same way a knuckleballer consistently beats expected average on balls in play. However, even though he’s consistently above average, he’s also all across the board as to what level. It certainly appears that, even with the help that his cutter may give, there’s still a lot of random noise in there.

    So yes, Rivera may be an extreme example that violates the general concept. But for a huge majority of pitchers, the rule holds true.

  23. Colm on March 30th, 2005 11:58 am

    Getting slightly sidetracked on the DIPS/BABIP questions here:

    If memory serves Jamie Moyer was a notable exception to the general rule that pitchers have little influence on BABIP. Possibly the most significant exception in the last ten years.

    What strikes me about this (other than pining for the Moyer we had 1996-2003) is that Moyer about the only pitcher to whom this applies who is neither a knuckleballer nor an extreme goundballer. Any stat-freaks got a theory on this?

  24. Colm on March 30th, 2005 12:00 pm

    Okay, viz Dave’s comment, I guess Rivera is another exception. Maybe he’s not so exceptional after all.

  25. Dave on March 30th, 2005 12:02 pm


    Actually, its extreme flyballers who often outperform their expected BABIP. Fly balls are turned into outs at a 73 percent clip versus just 67 percent for groundballs, so a guy like Ryan Franklin should theoretically post a lower BABIP than a guy like Derek Lowe. Things even out, though, because the fly balls that do go for hits are a lot more likely to be extra base knocks, so even though the BABIP might be different, the overall effect on a pitcher’s usefulness isn’t huge.

    I’m considering doing a study on SLG on BIP this summer. It could yield some interesting results.

  26. Paul Covert on March 30th, 2005 12:34 pm

    Note: written re. Jesse’s #19. First attempt to post didn’t take. Sorry if I’m getting too far behind the discussion here.

    Mariano Rivera’s career has included 540 non-HR hits on approximately 1997 balls in play, for an average of .270, compared to a typical .290 to .300. Jamie Moyer since 1996 has allowed a BABIP of .273, Tim Wakefield .277 for his career. (The greatest pitchers in recent memory tend to be better than average here, but not drastically so: Pedro and Maddux .279, Clemens and Mussina .286, Johnson .291. It’s their K, W, and HR rates that do more to set them apart.)

    So– continuing the thought from Dave’s #16, if I may– it’s not that pitchers have no differences in their skills that affect the results on balls in play. But the differences are on the order of 1 hit per game, even comparing the best to the worst. If a pitcher has a run with either an awful lot of hits falling in, or very few, it’s usually a short-term thing that will even out in time (mostly randomness, perhaps exacerbated by defense, as with the early-2004 M’s).

    Part of the difficulty is our natural tendency to judge by result. Yes, I know some of you are thinking: “But of course we judge by results– isn’t that the whole point? If a guy throws a pitch that the batter grounds to the shortstop, doesn’t that usually mean it was a good pitch?” But hear me out.

    Anyone who’s played a board game like Risk is probably familiar with the phenomenon that sometimes a player seems to go on a great run with the dice where everything’s coming up boxcars. (Presumably the same is true with Vegas-style gambling, but that’s a world with which I am rather less familiar.) Does this mean that you’ve suddenly developed a wonderful skill at throwing dice (or pulling slot maching handles or whatever)? No– it just means that you’ve gotten some breaks.

    Next, suppose that you’re on a pitcher’s mound– but that instead of actually pitching, you’re feeding balls into a high-tech pitching machine like “Abner,” which actually replicates major-league pitches. And suppose that the machine throws the same pitch to the same hitter twenty times (mixed in with other pitches, so the hitter doesn’t know what’s coming). Will the hitter get the same result all twenty swings? Of course not; he’ll probably connect solidly a few times, pop a few up, foul a few off, and miss a few entirely. But that needn’t mean that the pitcher was doing anything differently.

    But when a real-life pitcher does the same thing, then an illusion of control suddenly kicks into our brains. We start thinking that any pitch that allows a base hit– or, at least, any pitch that was struck solidly– was therefore a bad pitch, and any pitch that gets popped up was a good one. On rare occasions we make exceptions: “he got away with a mistake pitch there,” or perhaps “Vlad really dug that one out of the dirt.” But we fall victim to the illusion far, far more often than not.

    So then… can we judge pitchers by results? Yes– in the long term. A pitcher who consistently gets guys out is therefore a good pitcher. But if a pitcher gets guys out for a few innings, that could mean he’s a good pitcher; or it could mean he’s Rusty Meacham.

    Only time will tell.

  27. stiletto on March 30th, 2005 12:37 pm

    nice piscopo reference there, jeff.

  28. Jesse on March 30th, 2005 12:42 pm


    Actually McCracken addresses Moyer in his article. I should know, I just read it for the first time. He writes:

    “People have a hard time diagnosing who the pitchers are that are very good at preventing hits on balls in play. You’ll often hear people use names like Randy Johnson, Jamie Moyer and Andy Pettitte in protest of the concept, but by any definition you want to use, these guys are not particularly good in the stat.”

    So, apparently, Moyer, not so much.

    Anyway, I hope the late movement to the hands is an outlier because the announcers wouldn’t stop blabbing about how great Madritsch was at that when the game was on ESPN2 a couple of weeks ago, either his fastball or his change, I can’t remember. Visually, I know watching that happen again and again for Rivera has been pretty amazing.

    Are there free places to look at all these DIPS stats for last year or am I just going to have to buy one of these BP volumes I keep hearing about? I still feel a little skeptical, but hearing that the same pitchers bounce around from year to year a lot is pretty compelling. I’m definitely interested to see which pitchers consistently exceed such expectations.

  29. Dave on March 30th, 2005 12:54 pm

    DIPS is freely available in a lot of places. The Hardball Times stats page is my favorite (though they call it FIP – fielding independant pitching) because their page also contains a ton of other data you can’t get anywhere else. ESPN actually tracks DIPS now, too, though you have to hunt for it. There’s also a really nifty excel worksheet to caluclate DIPS on your own, if you’d like, that can be found here.

    If you want to get REALLY geeky, try Tango’s baseball page. There’s some good stuff there if you can get through the lack of effort at packaging it for normal human beings.

  30. Jesse on March 30th, 2005 12:56 pm

    Dave, thanks.

    Colm, I’m reading the Tippet article and it appears you are right and McCracken was wrong. Apologies.

  31. Jesse on March 30th, 2005 1:32 pm

    OK, I just finished the Tippett article, and not to beat a dead horse, but I think it comes to a very different conclusion than batters have very little control over what happens to balls put in play. Has the thinking progressed a lot since he wrote this one?


    Because I think he makes a very convincing (and pretty exhaustive) case that some pitchers are very good at controlling that. So, to bring this back to Meche, to be sure, there’s a decent chance the improvement was just noise, because a lot of what happens to balls after they are hit is some combination of chance and defense. But Moyer turned a huge corner in 1996 and never looked back (until ’04 maybe, but still, 8 years). They’re obviously not similar pitchers, but it makes it hard for me to think about the probability of Meche being able to sustain success.

    Any idea what kind of secondary indicators we might look for to see if he’s actually gotten better at this? I guess I can agree it seems unlikely, but I feel like it’s a lot muddier than this discussion indicates. Either way, the way Tippett frames it definitely makes me think that better outcomes on Meche’s BIP were probably because he was pitching better. Am I missing something? Is there newer research, or just newer analysis of Tippett’s research, we should all be up on?

  32. Dave on March 30th, 2005 1:36 pm

    Tippett’s article is a lot more even handed than Voros’ over-the-top claims. I try to lean more towards Tippett’s findings, that most pitchers have little control over balls in play being converted into outs, rather than Voros’ claims that nobody could do anything about batted balls that didn’t leave the yard.

    Guys like Moyer (and probably Rivera) are certainly reminders that this doesn’t apply to everyone, so we can’t say that Meche certainly won’t go on a 10 year run with BABIP’s of .270 or something. But we can say is that its pretty unlikely, and Meche doesn’t fit any of the normal qualifications for a DIPS beater. He’s not left-handed, doesn’t throw a knuckle ball, and doesn’t have a long history of beating expected average on balls in play, so basically, if it happens, it’s going to be out of the blue.

  33. Evan on March 30th, 2005 2:06 pm

    So, to reword that for the DIPS-rookie, while some pitchers do show an unusual ability to significantly prevent hits on BIP, the vast majority of pitchers show very little such ability. Therefore, when projecting the performance of younger pitchers, assuming average future DIPS performance is far more reasonable than assuming an (extremely unlikely) atypical future DIPS performance.

    Some pitchers can exert significant control over BABIP. But it’s such an unusual skillset (outside knuckleballers, who all seem to do it) that without a strong multi-year trend, there’s no reason to believe that any given pitcher can do so.

  34. Brian Rust on March 30th, 2005 2:08 pm

    Isn’t BABIP really more of a PIDS than a DIPS?

  35. Ralph Malph on March 30th, 2005 2:37 pm