45 replies to this topic

[Sprowl]

Josh Beckett is a different pitcher in 2011. His velocity is down from 2007 (92.7 vs. 94.6 mph), but the loss of velocity isn't costing him runs. Somehow he seems to have made a transition from an 11 o'clock power pitcher with a knockout curve to 10 o'clock veteran pitcher with 5 good choices and a keener grasp of how batters can be set up and put away. He uses the fastball to get strikes early and the offspeed pitches to get the out: cutter, curve and changeup have all been outrageously successful so far in 2011.

The cutter has gone from 2% of pitches in 2008 to 16% in 2011. It has also slowed down and slid more, with horizontal break in on the hands of a left-handed batter. Buchholz's slider has become a cutter; Beckett's cutter has become more of a slider.

Beckett almost lost the curve in 2010, probably as a consequence of his gradually descending arm angle. More slurvy action (ie, horizontal motion at the expense of vertical drop) made the pitch harder to command. As a consequence, its usage has dropped from ~25% in 2007-9 to 18% in 2010-11. Now it is used more for specialty purposes, especially thrown inside to move a RHB off the plate. When Beckett has the full movement of the pitch under command, it buckles the batter's knees but breaks back over the inside of the plate for a called strike.

Beckett's approach against LHB showed up well against the Indians: 4-seamers up and away, cutters down and in, and changeups out of the zone low or away:



A cautionary note: Beckett's flyball ratio is way up (GB/FB = 1), and his HR/FB ratio (4.9) and BABIP (.224) are unsustainably low. He's keeping the hitters off balance, so part of that good luck has been earned, but how long can he keep that up? I'm thinking that eventually he'll have to unearn some bad luck instead.

[Van Everyman]

He threw an absolutely nasty two-seamer tonight (to Asdrubel?) for a K tonight. It seems to me that power pitchers who learn to throw that pitch have an easier time pitching to contact w less pure stuff. Clay has been doing the same -- which may explain his lower K rate but would also explain his better overall results.

As for Beckett's GB/FB ratio, I think you're on to something. It's not like many of the flyballs I'm seeing are going very far. Between the buckling at the two-seamer/cutter, and the far better action on his curve over 2010, nobody's hitting much of anything w authority against him.

One question: is Beckett's reduced velocity actually the product of throwing slower or just less four-seamers? Bc it seems to me be can still get it up there when he wants.

[Jed Zeppelin]

Sprowl, the thread title and your post reminded me of what G38 said about Beckett before last season and his contract extension:

He’s getting to a stage in his career where the mental work, the prep work, is going to meet and surpass the physical effort. Which in his case is saying something because he’s a kid who does work his ass off and does want to be the best. When he does combine max effort in the mental preparation arena (which is not to say he hasn’t before, but as you age you learn different things about yourself and the game) with his physical preparation, things could get interesting.

Gives a different angle than the charts and the stats but it sure seems like this is the kind of thing we're seeing from Beckett this year.

[phragle]

Josh Beckett is a different pitcher in 2011. His velocity is down from 2007 (92.7 vs. 94.6 mph), but the loss of velocity isn't costing him runs. Somehow he seems to have made a transition from an 11 o'clock power pitcher with a knockout curve to 10 o'clock veteran pitcher with 5 good choices and a keener grasp of how batters can be set up and put away. He uses the fastball to get strikes early and the offspeed pitches to get the out: cutter, curve and changeup have all been outrageously successful so far in 2011.

The cutter has gone from 2% of pitches in 2008 to 16% in 2011. It has also slowed down and slid more, with horizontal break in on the hands of a left-handed batter. Buchholz's slider has become a cutter; Beckett's cutter has become more of a slider.

It's a good cutter. I definitely support its increased usage. It hasn't turned into a slider though. It is still about 3 mph off his fastball like it has always been, instead of the 10 or so mph like a traditional slider. So I think the fact that it is slower is just due to slower overall arm speed on all his pitches. He has got a little more horizontal movement on it though, probably due to the increased control and consistency of the pitch.

(To non Pitch FX experts like Sprowl)
Here is what a cutter looks like on Pitch FX. The cutter is the black squares like the key says on the side. This is from Beckett's start earlier tonight.

Fastball coming in at about 93, cutter at 90.


Also it looks like the Fenway cameras are still making the charts read about 4-5 inches low, like Sprowl, PedroKstheBambino, and I talked about in the gamethread the other day.

A cautionary note: Beckett's flyball ratio is way up (GB/FB = 1), and his HR/FB ratio (4.9) and BABIP (.224) are unsustainably low. He's keeping the hitters off balance, so part of that good luck has been earned, but how long can he keep that up? I'm thinking that eventually he'll have to unearn some bad luck instead.

Yeah he is due for some regression. I'd like to see him cut down on the walks and get more ground balls, but I guess you can say that about everyone.

[phragle]

One question: is Beckett's reduced velocity actually the product of throwing slower or just less four-seamers? Bc it seems to me be can still get it up there when he wants.

No he is getting older and throwing slower. It is not the end of the world though, he has a history of still pitching well without his best stuff. He is just going to have to rely less on his heater, and that is what he is doing, and did in 2010. As the mph decreases so did the usage.


Year, fastball velocity, percentage fastballs thrown.
2009, 94.3 mph, 60.2%
2010, 93.5 mph, 55.2%
2011, 92.7 mph, 52.8%

[Al Zarilla]

Every time I see Beckett pitch this year I think of the old pitcher's expression rock and fire. I don't know if he's more deliberate with the windup or if he pauses on his back foot a bit longer, but he looks a tad different. Works for me.

[twothousandone]

He uses the fastball to get strikes early and the offspeed pitches to get the out

The cutter has gone from 2% of pitches in 2008 to 16% in 2011.

Now [the curve] is used more for specialty purposes, especially thrown inside to move a RHB off the plate.

4-seamers up and away, cutters down and in, and changeups out of the zone low or away:

Since the catching thread has morphed, I'll ask here how this change in Beckett came about -- did Young push it from when he first joined, does Varitek get credit for calling it this way, would Saltalamacchia not be able to call it the same way?

And, given the occassional effectiveness of advanced scouting, would a switch next time out -- starting hitters off with curves in the first inning -- enabled him to keep the advantage of keeping hitters off-balance? Shouldn't somebody with the Tigers have figured out what Sprowl and phragle have? Shouldn't Cabrera, in particular, go up there ready to swing at a first pitch fastball?

[HriniakPosterChild]

And, given the occassional effectiveness of advanced scouting, would a switch next time out -- starting hitters off with curves in the first inning -- enabled him to keep the advantage of keeping hitters off-balance?

I've never thrown pitch in any league, but I've read that the curveball is a pitch you might not have working on a given night.


From Tek after the win in Cleveland:

"Josh should feel good about himself. We went to a lot more changeups and cutters because he didn't have his curve. He threw some real good sinkers and he had quality location of his fastball."

I've also read that just because the curve isn't working in warmups or in the first inning doesn't mean it won't be there that game. Figuring out when to take the chance the pitcher can get it working is where Tek makes his money that days. Last night was a night it never clicked for Beckett.

[PedroKsBambino]

Really interesting post/topic. Couple notes on it...

It's a good cutter. I definitely support its increased usage. It hasn't turned into a slider though. It is still about 3 mph off his fastball like it has always been, instead of the 10 or so mph like a traditional slider.

Agreed; one has to consider the fastball velo change when looking at the other pitches...a reminder that interpreting pitch FX data is tricky, great a resource as it truly is.

He's keeping the hitters off balance, so part of that good luck has been earned, but how long can he keep that up?

I'm unaware of any research that suggests that 'keeping hitters off balance' is either defineable or supports a low BABIP. Is this just an assumption (not an unreasonable one, but not clearly with any real meaning either) or is there something I've missed on this? I'm wayyyyy less current than I used to be on BABIP research, so could have missed it.

Edited by PedroKsBambino, 25 May 2011 - 09:23 PM.

[ReggieSmith]

The thing that always impressed me about Beckett was he had "the edge". Not sure if I can quantify that, but Muhammad Ali had "the edge" that willingness and arrogance to stand out and be boisterous. He wants to win, never mentions a physical injury, and you can see him vocal on the sidelines (if Pedroia were a pitcher, he would be Beckett). After last year, I didn't know if I would see "the old Beckett", but he's here and he's performing. I'll never forget him beating the MFY's on their own turf in 2003 on short rest. I'll never question his heart! He also protects his team and isn't afraid to hit someone on the other team when it is the correct thing to do. Having a healthy Beckett is something I didn't expect this year, but it's only icing on the cake.

Edited by ReggieSmith, 25 May 2011 - 10:22 PM.

[Sprowl]

I'm unaware of any research that suggests that 'keeping hitters off balance' is either defineable or supports a low BABIP. Is this just an assumption (not an unreasonable one, but not clearly with any real meaning either) or is there something I've missed on this? I'm wayyyyy less current than I used to be on BABIP research, so could have missed it.

A hitter off balance can be observed when a hitter lunges for a ball or gets jammed, manifesting in weak contact off the end of the bat or, if the ball is far enough in on the hands, a broken bat. I don't know that it has been measured, although I think it probably can be: HitFX velocity data coupled with observation might tell a clearer story than has been told so far.

Beckett is throwing a slower fastball, but appears to be locating it better. In 2010, way too many of his fastballs ended up as meatballs, but this year it doesn't seem to be happening as much. Is that due to luck? Better command of the fastball? More effective action on the cutter and changeup that makes a hitter less likely to sit on the fastball? Some combination of the three, I'm guessing, and the last two might stick around even when the luck evaporates.

I've never seen a convincing quantification of fastball command, and that still seems to me to be the missing link in what makes the difference between Beckett 2011 and Beckett 2010, and between Buchholz 2010 and Buchholz 2008. For fastball command, we're still dependent on anecdote, observation and a few strike zone charts.

[ljs252]

Very interesting discussion and analysis. The thing that jumps out to me is that Beckett really began this reinvention last year. And we all know how bad the results were. Last year, according to Fangraphs, he threw the 4 seam/2 seam about 55% of the time, cutter 15%, curve 18%, and change 11%. Thus far this year the pitch breakdowns are similiar: Fastballs about 53%, cutter 16%, curve 18%, and change about 13%. Additionally, his 2010 and 2011 K and walk rates are similiar with one another so far : 2010 K/9: 8.18 2011 K/9: 8.16, 2010 BB/9: 3.17, 2011 BB/9: 2.81. And, as has already been noted, his fastball velocity is down from last year and he is giving up more fly balls than last year. So, if his pitch selection between last year and this year is similiar and his fastball velocity is down even more, why the greatly increased success?

I tend to agree with Sprowl and others in that I think he is spotting the fastball much better than last year, is healthy, and he is throwing a better changeup and cutter than he ever has. His command of his change is better than I can ever recall in years past and he appears, at least to my eyes, to be getting more movement on it as it really dives down and away from lefties. It has become a great putaway pitch against lefties for him. And he uses his cutter in several different ways with equal effectiveness to both righties and lefties. Beckett has actually had a few starts this year where the curve hasn't been of much use to him and he has gotten by with 4 seam/2 seam/cutter/change. The beckett of old usually got pounded on nights he didn't have command of his curveball.

Its great to finally see him pitching instead of throwing. Beckett's stuff pre-2009 was so good that it was both an asset and liability: When he had the command of the old hammer curve to go along with the 94-97 mph heat he could dominate on any night he pitched. But if he didn't have the curve working the emperor had no clothes and he was reduced to trying to throw fastballs by people all night. And then we all would complain about him being stubborn and a dumb hick from Texas. I think getting his ass handed to him a little bit caused him to look in the mirror and realize he needed to change the way he pitched as he got older. One thing to keep in mind is that it will be interesting to see how the league, especially the al east, adjusts to the "new beckett" as the season progresses. And yeah he is def. due for some regression as his hr, babip, and strand rates are unsustainable.

[Eric Van]

The cutter has gone from 2% of pitches in 2008 to 16% in 2011. It has also slowed down and slid more, with horizontal break in on the hands of a left-handed batter. Buchholz's slider has become a cutter; Beckett's cutter has become more of a slider.

Beckett's "cutter" in fact has the single most telling sign of a slider: effective rotation of less than 1000 RPM (often well under), which means it has the bullet-like spin characteristic of a slider.

It also moves more like a slider than a cutter. This is average H, V (with Lester reversed horizontally) -- remember that for sliders and cutters, a lower V means more downward break:

1.5, 5.2: Lester's super cutter Wednesday
-0.1, 7.5: Buchholz's more ordinary cutter Tuesday
4.6, -0.3: Greinke's super slider in Cle 9/11/09
2.6, 1.6: Wheeler's average slider yesterday
2.0, 1.8: Beckett's "cutter" yesterday

It has the vertical break of a slider, a horizontal break intermediate between the two but quite a bit closer to slider (look at Buchholz and Wheeler for more ordinary amounts of break).

It has the rotation and vertical break of a slider, a spin axis and horizontal break just a bit in the direction of cutter from slider. It's really a tremendously hard slider, with a little less break which is a small price to pay for the extra 3-5 mph of velocity relative to his FB.

[phragle]

Beckett's "cutter" in fact has the single most telling sign of a slider: effective rotation of less than 1000 RPM (often well under), which means it has the bullet-like spin characteristic of a slider.

It also moves more like a slider than a cutter. This is average H, V (with Lester reversed horizontally) -- remember that for sliders and cutters, a lower V means more downward break:

1.5, 5.2: Lester's super cutter Wednesday
-0.1, 7.5: Buchholz's more ordinary cutter Tuesday
4.6, -0.3: Greinke's super slider in Cle 9/11/09
2.6, 1.6: Wheeler's average slider yesterday
2.0, 1.8: Beckett's "cutter" yesterday

It has the vertical break of a slider, a horizontal break intermediate between the two but quite a bit closer to slider (look at Buchholz and Wheeler for more ordinary amounts of break).

It has the rotation and vertical break of a slider, a spin axis and horizontal break just a bit in the direction of cutter from slider. It's really a tremendously hard slider, with a little less break which is a small price to pay for the extra 3-5 mph of velocity relative to his FB.

No no no. That would all be true if Cleveland's PFX cameras were calibrated properly, but they are not, so the data is skewed. Their cameras are even worse than Fenway's. All the pitches have 4-5 inches less (as in negative vertical movement) vertical movement than they should have if the cameras were calibrated properly. A dead give away is that the curveball has much more overall movement then the fastball.

-5.67 / 5.29 Fastball
7.88 / -10.16 Curveball

For that to be possible the curveball would have to have a much higher rotational velocity (RPM) than the fastball, and that is not possible. It doesn't work that way.

Another give away is Beckett's straight change up has negative vertical movement (-0.53), it doesn't work that way either.

So Cleveland's charts read 4-5 inches low, Boston's read 3-4 inches low, and New York's exaggerate horizontal movement. The only other places he has pitched are LAAA and BAL. Fortunately they both look calibrated and accurate.

In LA his cutter was (2.24 / 5.03)
In BAL his cutter was (2.01 / 6.45)

That is a cutter, a pure cutter, not a slider at all.

On top of all that, you know PFX cameras don't and can't record actual spin RPM, right? The Spin Chart is just an estimate of rotational velocity (RPM) calculated directly from the Horizontal Movement and Vertical Movement Chart.

Edit: Here are the LAAA and BAL charts for convenience

http://www.brooksbas...2F&prevDate=421

http://www.brooksbas...2F&prevDate=427

Edited by phragle, 26 May 2011 - 07:05 AM.

[PedroKsBambino]

A hitter off balance can be observed when a hitter lunges for a ball or gets jammed, manifesting in weak contact off the end of the bat or, if the ball is far enough in on the hands, a broken bat. I don't know that it has been measured, although I think it probably can be: HitFX velocity data coupled with observation might tell a clearer story than has been told so far.

Beckett is throwing a slower fastball, but appears to be locating it better. In 2010, way too many of his fastballs ended up as meatballs, but this year it doesn't seem to be happening as much. Is that due to luck? Better command of the fastball? More effective action on the cutter and changeup that makes a hitter less likely to sit on the fastball? Some combination of the three, I'm guessing, and the last two might stick around even when the luck evaporates.

So, to be clear, what I wondered about is why you said part of the BABIP drop was due to variety of pitches instead of just luck. Were it the case that lower BABIP correlated to a pitcher's variety of pitches/speeds, that seems like something that would have popped in the prior BABIP studies, don't you think? Seems shaky to argue that Beckett's current mix of pitches and speeds is the most diverse (or anywhere near the most diverse) of any pitcher who has been included in the numerous studies, right? And if it's not, then why would (in his case this year) the BABIP be skill rather than luck?

Seems like you're really just tossing out a lot of research on BABIP without any basis that I can see. Pitch FX is wonderful, but I continue to be of the view it does not explain all problems all the time and we get in trouble when we try to make it do so.

[effectivelywild]

Really interesting post/topic. Couple notes on it...



Agreed; one has to consider the fastball velo change when looking at the other pitches...a reminder that interpreting pitch FX data is tricky, great a resource as it truly is.



I'm unaware of any research that suggests that 'keeping hitters off balance' is either defineable or supports a low BABIP. Is this just an assumption (not an unreasonable one, but not clearly with any real meaning either) or is there something I've missed on this? I'm wayyyyy less current than I used to be on BABIP research, so could have missed it.

Well, I don't know if "keeping hitters off-balance" will result in a noticeably different BABIP (after all, for all of the times that they aren't expecting your next pitch, there will be times when they guess right as to what's coming and will pound it) but I imagine it would also show up in K and HR rates. After all, if you're sitting dead red and you get a change up or a slider, you're more likely to miss the pitch---same thing if you're thinking "ok, this is definitely going to be a change up" and you get a cutter---your timing is going to be off and you're more likely to miss the pitch. Similarly, I imagine "keeping hitters off-balance" would lower your HR rate, as its hard to hit the ball hard if you have no idea what pitch is coming next.

[Dick Drago]

I hadn't seen him pitch very much this year, and I was amazed at how good his change-up was the other night. It looked almost like a splitter the way it would suddenly dive. The Cleveland broadcasters referred to it as a splitter on several occasions.

It is really fun to watch his transformation; last year, not so much. One of the things that makes baseball great is observing the career trajectories of various players. Some become stars then slowly fade away, others reinvent themselves and continue to be effective. Now we just need John Lackey to do the same!

[KiltedFool]

A hitter off balance can be observed when a hitter lunges for a ball or gets jammed, manifesting in weak contact off the end of the bat or, if the ball is far enough in on the hands, a broken bat. I don't know that it has been measured, although I think it probably can be: HitFX velocity data coupled with observation might tell a clearer story than has been told so far.

Beckett is throwing a slower fastball, but appears to be locating it better. In 2010, way too many of his fastballs ended up as meatballs, but this year it doesn't seem to be happening as much. Is that due to luck? Better command of the fastball? More effective action on the cutter and changeup that makes a hitter less likely to sit on the fastball? Some combination of the three, I'm guessing, and the last two might stick around even when the luck evaporates.

I've never seen a convincing quantification of fastball command, and that still seems to me to be the missing link in what makes the difference between Beckett 2011 and Beckett 2010, and between Buchholz 2010 and Buchholz 2008. For fastball command, we're still dependent on anecdote, observation and a few strike zone charts.

Another aspect of hitter off-balance would be swinging earlier or later than would be typical for that batter. Out ahead of the change-up or a late defensive swing at a heater when they were obviously sitting offspeed. But then you'd have to quantify what was a typical time to start the swing for that batter and likely further define it for different pitcher types. Catching the data point where the pitcher has been living away and the batter is defending away so the pitcher goes inside, no idea how you'd catch that.

I can't see a case where a convincing quantification of fastball command could be accomplished, the closest you could get IMO would be whether the catcher's mitt moved. If the catcher doesn't move the glove, then obviously the pitcher commanded that pitch. Other than that you're stuck trying to read the pitcher's mind of where he wanted to throw the ball.

If you could pick up the initial and final locations of the mitt, I guess you could also do precision and accuracy studies and base your analysis of command on that. And you'd be assuming a level of mind meld between pitcher and catcher as well.

[Sprowl]

So Cleveland's charts read 4-5 inches low, Boston's read 3-4 inches low, and New York's exaggerate horizontal movement. The only other places he has pitched are LAAA and BAL. Fortunately they both look calibrated and accurate.

In LA his cutter was (2.24 / 5.03)
In BAL his cutter was (2.01 / 6.45)

That is a cutter, a pure cutter, not a slider at all.

I agree that the vertical movement error (which seemed to show up in Fenway's readings only during this past homestand, and not all season long) makes it impossible to derive accurate spin RPM data for Beckett's cutter during the last two games. When I wrote that Beckett's cutter is becoming more of a slider, I was referring to increased horizontal movement on the pitch in 2011 compared to 2010 and 2009 (prior to 2009 he threw so few of them that it is harder to generalize). Those quantitative changes aren't enormous, but to my eyes they confirm observation that Beckett's cutter looks different and seems to move more.

When the cutter's horizontal movement data are broken down by batter handedness, the picture becomes clearer: Beckett's cutter is definitely a cutter when thrown to lefties (1.7" movement, N = 85)), but the same pitch registers 2.2" (N = 90) when thrown to righties. It confirms observation (again, this is my impression, not a claim of proof) that Beckett tries to get more movement on the cutter when throwing to the same-handed batter: it moves more because Beckett throws it a little bit differently, with more torque from the wrist and slightly lower velocity (~.3 mph lower, to be precise). Hence, more of a slider than before, even if it is still accurate to classify the pitch as a cutter. By way of comparison, most major-league cutters have little if any horizontal movement, but Beckett's cutter is starting to behave a little more like Mariano Rivera's cutter. I wouldn't claim that it gets the same results, but it has been an increasingly effective weapon the last two years.

This assumes that the pitch classification is correct. I think that's fair, since pitchfx has learned to read Beckett quite well, as phragle's charts show. It also assumes that the horizontal movement data are reliable, but does not depend on vertical movement readings. The change is marginal, and if it generated the hypothesis rather than confirming observation, it might be ignored, but I think we're seeing a change that is both real and intentional.

So, to be clear, what I wondered about is why you said part of the BABIP drop was due to variety of pitches instead of just luck. Were it the case that lower BABIP correlated to a pitcher's variety of pitches/speeds, that seems like something that would have popped in the prior BABIP studies, don't you think? Seems shaky to argue that Beckett's current mix of pitches and speeds is the most diverse (or anywhere near the most diverse) of any pitcher who has been included in the numerous studies, right? And if it's not, then why would (in his case this year) the BABIP be skill rather than luck?

Seems like you're really just tossing out a lot of research on BABIP without any basis that I can see. Pitch FX is wonderful, but I continue to be of the view it does not explain all problems all the time and we get in trouble when we try to make it do so.

I am not tossing out any research that I can think of, nor am I using pitchfx to explain all problems. Instead, I am using pitchfx data, with appropriate reservations as stated, where they can help confirm or reject hypotheses arrived at by watching Beckett pitch. If hitfx data were publicly available, I think it would be possible to put BABIP studies on a firmer footing by quantifying the quality of contact -- speed of groundballs, velocity and trajectory of flyballs and line drives. I don't have those data and am not attempting that exercise in this thread.

Luck is not only or always about BABIP -- it is as likely to be about power suppression, as effectivelywild argues below.

Well, I don't know if "keeping hitters off-balance" will result in a noticeably different BABIP (after all, for all of the times that they aren't expecting your next pitch, there will be times when they guess right as to what's coming and will pound it) but I imagine it would also show up in K and HR rates. After all, if you're sitting dead red and you get a change up or a slider, you're more likely to miss the pitch---same thing if you're thinking "ok, this is definitely going to be a change up" and you get a cutter---your timing is going to be off and you're more likely to miss the pitch. Similarly, I imagine "keeping hitters off-balance" would lower your HR rate, as its hard to hit the ball hard if you have no idea what pitch is coming next.

Beckett's K rate is the same in 2011 as in 2010 (~8.2), but the HR/FB rate has gone from 14.2% to 4.5%. The 86% LOB% is fortunate as well. xFIP 3.30 vs. ERA 1.69 suggests that Beckett has been very good, but his results have been better than that. There's still no way to explain why he grooved so many fastballs in 2010, and hasn't done the same this year. We could examine some game-to-game strike zone charts, but that's likely to give us more anecdotes than explanations. Fastball command is still a black box.

[PedroKsBambino]

I am not tossing out any research that I can think of, nor am I using pitchfx to explain all problems. Instead, I am using pitchfx data, with appropriate reservations as stated, where they can help confirm or reject hypotheses arrived at by watching Beckett pitch. If hitfx data were publicly available, I think it would be possible to put BABIP studies on a firmer footing by quantifying the quality of contact -- speed of groundballs, velocity and trajectory of flyballs and line drives. I don't have those data and am not attempting that exercise in this thread.

Luck is not only or always about BABIP -- it is as likely to be about power suppression, as effectively wild argues below.

As above, a set of statements about 'luck' that are not consistent with any research I've seen on the subject, and this is a topic that has been pretty widely researched.

If I'm missing some research, happy to learn about it. I am sure there's a bunch of things written in this area (especially last two years) that I've missed; that acknowledged, the idea that diversity of pitches drives luck (as manifested through BABIP, especially) would be new so far as I know and thus, I think we need to be cautious. That does not mean its wrong, it only means that considering how the hypothesis fits with existing research is something to consider.

[Eric Van]

No no no. That would all be true if Cleveland's PFX cameras were calibrated properly, but they are not, so the data is skewed. Their cameras are even worse than Fenway's. All the pitches have 4-5 inches less (as in negative vertical movement) vertical movement than they should have if the cameras were calibrated properly. A dead give away is that the curveball has much more overall movement then the fastball.

-5.67 / 5.29 Fastball
7.88 / -10.16 Curveball

For that to be possible the curveball would have to have a much higher rotational velocity (RPM) than the fastball, and that is not possible. It doesn't work that way.

Another give away is Beckett's straight change up has negative vertical movement (-0.53), it doesn't work that way either.

So Cleveland's charts read 4-5 inches low, Boston's read 3-4 inches low, and New York's exaggerate horizontal movement. The only other places he has pitched are LAAA and BAL. Fortunately they both look calibrated and accurate.

In LA his cutter was (2.24 / 5.03)
In BAL his cutter was (2.01 / 6.45)

That is a cutter, a pure cutter, not a slider at all.

On top of all that, you know PFX cameras don't and can't record actual spin RPM, right? The Spin Chart is just an estimate of rotational velocity (RPM) calculated directly from the Horizontal Movement and Vertical Movement Chart.

Edit: Here are the LAAA and BAL charts for convenience

http://www.brooksbas...2F&prevDate=421

http://www.brooksbas...2F&prevDate=427

I was aware of the need to use data from one park only, which is what I did. (And to be precise, both the movement and the spin parameters are calculated from the same set of velocity and acceleration parameters, which is the first thing they derive from the camera data.)

A look at the data from his Yankee Stadium start shows a pitch that is pretty much smack-dab right in the middle between a cutter and a slider in terms of spin. The average RPM is about 1100, the spin axis averages 150 or so. Those are both somewhat lower than a cutter. Lester the next day is up around 1400 and 180, which is much more typical.

Really, the actual variety of pitches exceeds our vocabulary. Mo's "cutter" the same night as Lester is unlike either Beckett's or Lester's. There is almost no overlap between the spin parameter clusters of the three pitches. And Rivera alone shows a relationship between RPM and spin angle (diagonally-running cluster), which you get from a specific release technique.

(Those checking this out should remember that Lester's graphs need to be swapped L/R.)

Beckett, Lester, Mo.

[phragle]

I agree that the vertical movement error (which seemed to show up in Fenway's readings only during this past homestand, and not all season long) makes it impossible to derive accurate spin RPM data for Beckett's cutter during the last two games. When I wrote that Beckett's cutter is becoming more of a slider, I was referring to increased horizontal movement on the pitch in 2011 compared to 2010 and 2009 (prior to 2009 he threw so few of them that it is harder to generalize). Those quantitative changes aren't enormous, but to my eyes they confirm observation that Beckett's cutter looks different and seems to move more.

When the cutter's horizontal movement data are broken down by batter handedness, the picture becomes clearer: Beckett's cutter is definitely a cutter when thrown to lefties (1.7" movement, N = 85)), but the same pitch registers 2.2" (N = 90) when thrown to righties. It confirms observation (again, this is my impression, not a claim of proof) that Beckett tries to get more movement on the cutter when throwing to the same-handed batter: it moves more because Beckett throws it a little bit differently, with more torque from the wrist and slightly lower velocity (~.3 mph lower, to be precise). Hence, more of a slider than before, even if it is still accurate to classify the pitch as a cutter. By way of comparison, most major-league cutters have little if any horizontal movement, but Beckett's cutter is starting to behave a little more like Mariano Rivera's cutter. I wouldn't claim that it gets the same results, but it has been an increasingly effective weapon the last two years.

This assumes that the pitch classification is correct. I think that's fair, since pitchfx has learned to read Beckett quite well, as phragle's charts show. It also assumes that the horizontal movement data are reliable, but does not depend on vertical movement readings. The change is marginal, and if it generated the hypothesis rather than confirming observation, it might be ignored, but I think we're seeing a change that is both real and intentional.

Look at my first post, I already agreed that the cutter has more horizontal movement than years past. Fangraphs numbers are just not good or accurate enough for me. The margin of error of the pitch classification algorithm is too large, or at least used to be. The algorithm has improved this year, but its not perfect by any means. Then add the inconsistency from park to park and the numbers become more than useless.


Without even looking at Fangraphs numbers I bet it says Beckett's curve is huge this year. Which is completely untrue. It just has a better more controllable angle. Ok let's see if I'm right. Yup, the numbers say it has about 2 inches more movement both ways. That is total crap, misleading, and plain wrong.

Anyway, Beckett threw a lot of cutters in 2009 and 2010 with 2 or more inches of horizontal movement, it just wasn't consistent movement. Sometimes it would just look like a bad fastball and not cut much at all, but sometimes it would cut 4 inches or more. Now, in 2011, the cutter has had consistent movement, not as many bad cutters and more good cutters. That is my theory on the increased movement.

[Eric Van]

Re BABIP:

If you think of a specific game where a pitcher had terrific BABIP, a game you watched, four times out of five it was clearly skill and not luck -- a game where the pitcher had guys swinging early in the count and hitting one routine grounder or one easy fly ball (less often, a mixture of both) after another.

Now, if the pitchers that we've seen do this could do it consistently, and the were more such pitchers, we would observe a greater variance in BABIP than we do, and more year-to-year correlation as a result.

So the conclusion is that getting that kind of easy contact requires finer command than most pitchers can sustain over a season, and / or requires an approach that hitters will eventually adjust to.

Beckett's BABIP will normalize, but not because he's been lucky. It will normalize because he's been friggin' great and no one can be expected to sustain that forever (look at the cluster of three mediocre starts that Lester just got past).

Now, Clay Buchholz has been shown to be able to sustain a BABIP skill for over a year now, and at some point hitters may well adjust.

Finally, there are the rare guys whose mechanics are inherently deceptive, resulting in a sustainable BABIP skill. Jered Weaver seems to throw three-quarters but has the fastball spin of a guy who nearly comes over the top (his forearm angle is unusual relative to his upper arm). As a result, guys who have not faced him a lot tend to expect more run and less rise than he actually has, which results in a lot of weak fly balls. (Guys like this will have larger-than-usual splits by time around the batting order.)

Edited by Eric Van, 26 May 2011 - 11:53 AM.

[Sprowl]

As above, a set of statements about 'luck' that are not consistent with any research I've seen on the subject, and this is a topic that has been pretty widely researched.

It is not clear to me what statements about luck you are referring to, nor what claims you infer that I have made, since I do not recognize my ideas in your paraphrase. There are three indicators commonly thought to be outside of the pitcher's long-term control, and expected to revert toward league-wide norms: BABIP, HR/FB and LOB%. Beckett has been lucky on all three so far. But it is also the case that BABIP and HR/FB are influenced by the quality of contact that in turn is influenced by aspects of skill like variety, command and movement. Inducing weak contact could come from putting the hitter off balance by changes in the sequence pitch type and location, keeping the ball out of the middle of the strike zone, and/or getting so much movement on the pitch that the batter doesn't hit the ball on the sweet spot. When the data becomes more widely available to the public, we may be able to put numbers on how weak that contact is, and for how long pitchers are able to maintain.

Buchholz's reversion to high HR/FB rates early in the season wasn't "pure" luck, although it was widely expected. He was also throwing slower, straighter fastballs in the middle of the zone. Now that he has regained some of the movement he had in 2010, he's getting better results. Even when a pitcher is pitching badly, the results can sometimes be even worse. I think EV did a breakdown of Lester in April 2010 that argued that a) Lester was throwing more meatballs, but b) batters were doing more with those meatballs than they ordinarily would. The same description applies to Buchholz in April 2011, when every mistake got hit hard. In the ordinary run of events, batters will miss a fair number of mistake pitches.

I was aware of the need to use data from one park only, which is what I did. (And to be precise, both the movement and the spin parameters are calculated from the same set of velocity and acceleration parameters, which is the first thing they derive from the camera data.)

A look at the data from his Yankee Stadium start shows a pitch that is pretty much smack-dab right in the middle between a cutter and a slider in terms of spin. The average RPM is about 1100, the spin axis averages 150 or so. Those are both somewhat lower than a cutter. Lester the next day is up around 1400 and 180, which is much more typical.

Yankee Stadium, both old and new, has had ongoing problems with measurements of horizontal movement, so we may need to look at data from parks that are not CLE, NYY or BOS (for the last two weeks; otherwise Fenway's movement data have been pretty consistent). IIRC, Jnai and Stupendous Man worked out a table that ranks different ballparks in terms of whether they exaggerate or diminish velocity; they may have done the same for the movement parameters, which would help in coming up with a set of parks that would provide the most reliable baselines for comparison.

In the 5/15 game in NY, all horizontal data seem to be shifted ~2" to the right, adding more movement onto Beckett's and Rivera's cutters than they usually get, and removing apparent movement from Lester's cutter, which shows up as having negative cut (ie, close to 4-seam fastball action). The 5/15 horizontal movement appears to be at the extreme of Rivera's recorded range -- when Rivera pitches on the road, his cutter usually shows less horizontal cutting movement. Even so, if we chose data from another road park to examine, I would guess that it would also show that Beckett has indeed gradually gotten more horizontal movement on the cutter over the years, because our eyes told us that first.

Look at my first post, I already agreed that the cutter has more horizontal movement than years past. Fangraphs numbers are just not good or accurate enough for me. The margin of error of the pitch classification algorithm is too large, or at least used to be. The algorithm has improved this year, but its not perfect by any means. Then add the inconsistency from park to park and the numbers become more than useless.

Fangraphs/BIS pitch classifications are bad for some pitchers (ie, Lackey, who has a similarly unorthodox delivery as EV describes for Weaver), but for others, they are quite reliable, and more so in recent years. For Beckett, both BIS and PitchFX do a very good job of recognizing his stuff, in spite of some problems distinguishing 4- from 2-seamers. Not perfect doesn't mean useless -- it means that interpretation is required and some data need to be set aside. Season-long averages, however, tend to smooth over park-to-park variation, unless there is some persistent problem with the home park (eg, Fenway's velocity readings, which tend to make every pitcher look a little bit faster). If season-long averages confirm other notions based on observation or anecdote, dismissing the numbers entirely strikes me as throwing the baby out with the bathwater.

[PedroKsBambino]

It is not clear to me what statements about luck you are referring to, nor what claims you infer that I have made, since I do not recognize my ideas in your paraphrase. There are three indicators commonly thought to be outside of the pitcher's long-term control, and expected to revert toward league-wide norms: BABIP, HR/FB and LOB%. Beckett has been lucky on all three so far. But it is also the case that BABIP and HR/FB are influenced by the quality of contact that in turn is influenced by aspects of skill like variety, command and movement. Inducing weak contact could come from putting the hitter off balance by changes in the sequence pitch type and location, keeping the ball out of the middle of the strike zone, and/or getting so much movement on the pitch that the batter doesn't hit the ball on the sweet spot. When the data becomes more widely available to the public, we may be able to put numbers on how weak that contact is, and for how long pitchers are able to maintain.

Puzzling, but let me try it this way.

This:

He's keeping the hitters off balance, so part of that good luck has been earned, but how long can he keep that up? I'm thinking that eventually he'll have to unearn some bad luck instead.

makes intuitive sense. However, we also know that for virtually all pitchers, the stats being referred to in the quoted text (e.g. BABIP, etc) move towards a standard point.

The question has been, why would we assume in Beckett's case that it would move differently, e.g. part is earned and will be kept?

Theory-wise, EV states why that could be true (e.g. maintaining unusually fine control). However, it isn't at all clear to me that we can actually correlate the pitch FX data to BABIP well enough to make this connection prospectively. And if not, then why wouldn't we assume the BABIP 'rules' of moving to standard points are the most likely to govern?

To put it a different way, why would we be confident (in a data and projection sense, not a theory one) that we can identify luck vs skill here based on what we have?

[Sprowl]

Puzzling, but let me try it this way.

This:

makes intuitive sense. However, we also know that for virtually all pitchers, the stats being referred to in the quoted text (e.g. BABIP, etc) move towards a standard point.

The question has been, why would we assume in Beckett's case that it would move differently, e.g. part is earned and will be kept?

Theory-wise, EV states why that could be true (e.g. maintaining unusually fine control). However, it isn't at all clear to me that we can actually correlate the pitch FX data to BABIP well enough to make this connection prospectively. And if not, then why wouldn't we assume the BABIP 'rules' of moving to standard points are the most likely to govern?

To put it a different way, why would we be confident (in a data and projection sense, not a theory one) that we can identify luck vs skill here based on what we have?

I see better what you're aiming at now. I don't know that pitchfx data have been used to explain BABIP variations, except through a few illustrations of good and bad command and a rather coarse set of results arising from separating GB, FB and LD. I tried at one point to come up with a way of combining pitch count, pitch location and pitch type to illustrate what sequences were effective in generating desirable results like a swinging strike or a groundball, but there are too many variables, too many of which can only be approximated, to explain the art of pitching by means of a correlation of results with pitchfx variables.

Part of the cult of BABIP arose because the data are available and easily computed. I think that BABIP reversion should be regarded as a heuristic rather than an iron law, and as a pattern that appears only at a given scope of analysis. That scope is produced by lumping all balls in play into one category. Hitfx data for trajectory, velocity and distance, however, could open up the black box of BABIP by enabling a separation of a lumpy category like Balls In Play into subsets (eg, lazy fly balls should produce a lower BA than sharply hit one-hop grounders). If quality of contact is under partial control of some pitchers, then so is BABIP.

Clearly that partial control is very small because each at-bat outcome is the product of multiple actors, including (at least) the pitcher and the batter, and sometimes the fielders, umpires and managers (via positioning or pitch-calling from the bench). Sometimes the batter will miss a meatball, and sometimes the batter will get a hit on a ball well outside the strike zone, just as the fielders will make excellent plays or fail to handle routine opportunities.

Nevertheless, when pitchers are pitching well, I hypothesize that some of them are earning a proportion of their luck. Not all: some pitchers are "just lucky," and should revert entirely toward the major-league mean. Others are lucky-good because they are pitching well for a period of time at a level that they cannot maintain (EV's example): they're good, but can't keep up that performance, and they too will eventually revert toward the mean. If a pitcher can persistently obtain lower-than-average BABIP over multiple seasons (see the Aceves thread for an example of a persistently low BABIP), then we have several options: 1, call the pitcher unnaturally lucky, at the far end of the bell curve of career luck (after all, every bell curve has some tail or it wouldn't be a curve); 2, call the pitcher really good, because results are what count; or 3, call the metric too coarse to yield insight, and seek a better metric.

The component that is skill does not show up in PitchFX data, afaik, but if it exists, it should show up better as quality of contact in HitFX. For instance, Buchholz in 2010 seemed to get a lot of feeble groundballs on the changeup. It's a bit counter-intuitive, because changeups more typically generate swings-and-misses (if the hitter is off balance), pulled foul balls (if the hitter gets the timing wrong), or solid contact (if the hitter is looking for the change, or recognizes the pitch and makes a good adjustment). HitFX data should make it possible to see whether the observation holds true (average groundball velocity off a Buchholz changeup, compared to other changeups), and if so, what pitch sequences or locations in the strike zone were most likely to produce weak groundballs.

[donutogre]

A little bump for Josh F. Beckett, who has pitched 21 innings against his former tormenters this season, giving up a grand total of 2 ER, 10 hits, 5 walks.

He even heard the people who have been bitching about him supposedly not pitching enough innings lately and went 7 full.

In 7 starts since leaving one early because of rain, Josh has not failed to finish the 6th, and hasn't given up more than 3 ER. He's put up a sterling 1.03 WHIP, good for 9th in the AL. This turnaround has just been a joy to watch, and with 2/5 of the rotation in shambles and starting the season as a question mark, JFB has come up huge.

[Lose Remerswaal]

And he hit 3 Yankees, including one AFTER a warning was given to both benches, and didn't get ejected!

[donutogre]

And he hit 3 Yankees, including one AFTER a warning was given to both benches, and didn't get ejected!

Ha. Dude is walking on water right now

[DanoooME]

He even heard the people who have been bitching about him supposedly not pitching enough innings lately and went 7 full.

And it sure didn't look like he'd make it that far with the Dice-K-esque 1st inning he had last night. He gutted it out and got out of a lot of jams. His breaking pitches don't seem to be the problem; it's his fastball command that's really spotty right now. Could that be a mechanical issue?

[Snodgrass'Muff]

The way he settled down and followed that rough first inning with 6 really solid innings last night has me feeling really good about him. He's been a treat to watch this season, and while last year's "Oh no!" reaction to him losing control early in the game was easy to fall into, the confidence he's inspired for most of 2011 had snuffed it out again by the third inning when it looked obvious that he wasn't in the midst of a full meltdown.

He's not Josh Beckett circa 2007, he's Josh Becket 2011. And both are truly outstanding pitchers.

[Sprowl]

Beckett had no-hitter stuff against the Rays, although it will go down in the records as a one-hitter. He didn't do it with velocity, although 92-94 on the fastball is good, and sustaining it for 9 innings is better. He didn't do it with command, although he walked nobody -- he threw lots of pitches over the middle of the plate. The Rays hit them, but not that hard, and the Sox fielders caught them.



I think he did it with wicked movement on his fastball, cutter and changeup. The fastball looked to have more pop, vertical jump, or late movement. I'm not entirely confident in Tampa's horizontal movement charts yet, so take the movement chart with a grain of salt for now. The vertical lift on that fastball really showed in the way Beckett got swinging strikes on the high fastball, forcing hitters to chase it at the letters. There is also excellent differentiation between the fastball and the two primary alternatives, the cutter and changeup (the curve and the sinker were both scarce in this game). Beckett's changeup is sinking more, and his cutter is cutting more. While the changeup remains primarily a tool for dealing with lefties, the cutter is now being thrown as or more often to righties, both to brush them back inside and to tail away outside.

[Eric Van]

There is also excellent differentiation between the fastball and the two primary alternatives, the cutter and changeup (the curve and the sinker were both scarce in this game).

I actually see 51 four-seamers, 16 cutters, 12 changes, 10 two-seamers, and 8 curves, so it's more like a nice continuum among all the alternatives, and I think that mixing it up that well was a big part of the success. The differentiation between the two- and four-seamer was excellent, as well.

It just bugs me that even with a guy whose pitches cluster as obviously as Beckett's*, and after all the accumulated data, that the pitch/fx algorithm still misclassifies his pitches (last night, 5 two-seamers as four-seamers and 13 four-seamers as cutters, which is probably the result of all the axis measurements being too low. The algorithm doesn't seem to consider park-to-park variations).

*In general, the spin versus speed chart gives the clearest clustering, and is often all you need. RPM can be helpful as a tiebreaker.

[JMDurron]

It's obscene that the lone "In Play, No Out" pitch is located where it is. THAT is what prevented a no-hitter.

I missed the first 3 innings, but watching the last 6, I can't remember the last time Beckett looked so effective while throwing so few curveballs. It seems that Beckett can be dominant with any two of his fastballs (4-seam, 2-seam/sinker, cutter) going well along with one of his other pitches (curve or change). I also can't personally recall seeing Beckett get more out of his changeup in a single start, while using it so often.

[Bucknahs Bum Ankle]

It's obscene that the lone "In Play, No Out" pitch is located where it is. THAT is what prevented a no-hitter.

Ridiculous. At least it happened it the third inning and we didn't have to endure a gut-wrenching broken-up perfect game in the the ninth.

Also, we give the umps so much crap when they screw up, Rob Drake deserves major props for the strike zone called last night. I see one low strike that was called a ball and one obviously inside ball that was called a strike. Otherwise everything around the margins was called just about perfectly. Impressive.

Edited by Bucknahs Bum Ankle, 16 June 2011 - 08:19 AM.

[phrenile]

Hurler again. Tito confirms Beckett's missing tomorrow's start with "we'll go with intestinal turmoil."

[Al Zarilla]

Hurler again. Tito confirms Beckett's missing tomorrow's start with "we'll go with intestinal turmoil."

Getting the trots makes him go back to being a hurler?

[joe dokes]

Getting the trots makes him go back to being a hurler?

My guess is that Beckett is both hurling and trotting.

[Al Zarilla]

My guess is that Beckett is both hurling and trotting.

That's funny, but not to Josh.

[joe dokes]

That's funny, but not to Josh.

Yeah, but I can picture Pedroia standing outside the stall laughing at him anyway.

[Sprowl]

Beckett's back, and he's back bigger than Texas. Bigger than Houston anyway.

Against the Astros, Beckett was pitching like Beckett 2007 -- not entirely, since his velocity peaked at 93, not 97. The throwback element came in the revival of his ability to get batters to miss both his fastball and his curve. He wasn't dependent on the newfangled cutter or midfangled changeup that he uses to persuade AL East hitters that they haven't seen his stuff a thousand times already. He threw the cutter and the changeup, but what really worked him today were his old standbys, especially the curve, which the Astros could not hit to save their lives: 5 swinging strikes out of 18 curves is like the Beckett of old.

The 4-seam fastball had decent velocity, but lots of jump / pop / late movement. As Beckett threw against the Rays, so he threw against the Astros: fastballs up at the letters, with most of the movement in the vertical axis, for 11 strikeouts and zero walks. There is no such thing as late movement, except in batters' heads. When it's in their heads, it gets the job done.

[Flynn4ever]

Beckett's back, and he's back bigger than Texas. Bigger than Houston anyway.

Against the Astros, Beckett was pitching like Beckett 2007 -- not entirely, since his velocity peaked at 93, not 97. The throwback element came in the revival of his ability to get batters to miss both his fastball and his curve. He wasn't dependent on the newfangled cutter or midfangled changeup that he uses to persuade AL East hitters that they haven't seen his stuff a thousand times already. He threw the cutter and the changeup, but what really worked him today were his old standbys, especially the curve, which the Astros could not hit to save their lives: 5 swinging strikes out of 18 curves is like the Beckett of old.

The 4-seam fastball had decent velocity, but lots of jump / pop / late movement. As Beckett threw against the Rays, so he threw against the Astros: fastballs up at the letters, with most of the movement in the vertical axis, for 11 strikeouts and zero walks. There is no such thing as late movement, except in batters' heads. When it's in their heads, it gets the job done.

Is it worrisome that he seems down a bit on velocity? I realize he is not 26 anymore (God knows, neither am I) but I wonder, Sprowl, do you think he can keep getting away with wile? I know pitchers have made careers of it, but I never guessed that Beckett had the mentality for that. Guesses anyone?

[mfried]

Is it worrisome that he seems down a bit on velocity? I realize he is not 26 anymore (God knows, neither am I) but I wonder, Sprowl, do you think he can keep getting away with wile? I know pitchers have made careers of it, but I never guessed that Beckett had the mentality for that. Guesses anyone?

I have the distinct impression that JB has consciously sacrificed 2-3 mph on the 4-seamer in order to gain command within the strike zone. He's a completely different (and superior) pitcher to the "power model" of the past.

[phragle]

I have the distinct impression that JB has consciously sacrificed 2-3 mph on the 4-seamer in order to gain command within the strike zone. He's a completely different (and superior) pitcher to the "power model" of the past.

He has gotten better results, but he hasn't been a better pitcher than in 2007, 2008, or 2009. By ERA or FIP he has, but he is due for some serious regression. His hits/9, HR/9, BABIP, and LOB% are all unsustainable. He is also walking more batters, and striking out less, while getting less ground balls and more fly balls. His 3.65 xFIP is closest to what we should expect the rest of the way.

[JMDurron]

I have the distinct impression that JB has consciously sacrificed 2-3 mph on the 4-seamer in order to gain command within the strike zone. He's a completely different (and superior) pitcher to the "power model" of the past.

I disagree. I think he has added pitches to adjust to gradually degraded "stuff", specifically in terms of velocity. He has increasingly relied on his cutter, 2-seamer, and changeup along with his classic 4-seamer and curveball precisely because he can't hit 97 on his 4-seam fastball with any regularity anymore. He made this adjustment in the 2010 campaign, but hadn't developed enough of a feel for the pitches in his newer bag of tricks to get away with it. Now he has, and he might be capable of sustaining weaker contact due to having more ways to try to keep opposing hitters off balance.

He can get away with just a 1-2 punch against the Astros, but when it's AL lineups, 2011 Beckett needs at least 3/5 of his pitches to be working to be effective. Thankfully, that appears to happen more often than not these days. 2007 Beckett is dead, 2011 Beckett can get equally good results using rather different methods.

[SMU_Sox]

Just a minor point JMD but didn't we determine a lot of Beckett's problems last year originated after 75 or so pitches? I think that either means he was dealing with fatigue/out-of-shape or his stuff wasn't good enough to fool hitters the third time around. It could also be both factors in conjunction. Lackey otoh has no stamina issues but might not be able to command all his stuff due to injury.
Edit: perhaps Lackey has a stamina issue - in some of his start his velocity plummets as the game goes on.

Edited by johnlimberakis, 04 July 2011 - 06:14 PM.