Digesting Beckett's Dominance Options

[mr guido]

EV, thanks for posting the pitch data. I too love Enhanced Game Day. It allows for some interesting observations.

Beckett threw a curve in just 11 of the 31 plate appearances... but of those 11 PAs, 7 of them featured multiple curveballs (and 2 of the other 4 ended at-bats)... and in 6 of those 7, curves were thrown back-to-back. This tells me that they were only mixing in curves to hitters who are susceptible to them, and pounding them with it if so. A "I know you can't hit this so good luck" approach rather than using one pitch to set up another.

Figgins & O-Cab combined for 28 pitches, of which 24 were fastballs. Meanwhile Pinkeye, Kotchman, Aybar, Kendrick, and Morales (the kids under 24 years old & the old man w/ the gimpy eye) got 36 pitches, of which only 18 were fastballs. Vlad got all fastballs except for the stretch of 4 consecutive curves.

[86spike]

According to that list, Jason Schmidt did it to Florida in the 2003 ALDS.

Opposing pitcher: Josh Beckett.

derr... reading skills rusty

So here's a question... do you think that dominant performance is the type of thing that can/could/would put the LAA batters into slumps for the next few games? They got nothing to hit all night... and now have to sit another night before getting another crack. I imagine a lot of them would rather get right to it tonight.

[William Robertso...]

To put it in SoSH terms, Josh a'Beckett's performance has taken 6 pages and counting to say enough about it. My first thought last night was that in Sox terms it was Ruthian. I always like that word. But thanks to dog, Eric, and others for digging up some good stats and comparisons.

[MentalDisabldLst]

More on topic - at no point did I feel Beckett was even reaching back for anything extra. Even after each single or the Ellsbury catch in the 9th he looked really comfortable - more a sense of "well, I placed it where I wanted it, sometimes these balls fall in".

It's pretty hard to reach back for something extra when every pitch you're throwing is already 96 or 97. Seriously, look at gameday, the guy threw like 3 fastballs under 95mph the whole game. His pitching mechanics were an absolute clinic of consistency.

edit: and you can read EV's chart on the last page to prove it

This post has been edited by MentalDisabldLst: Oct 4 2007, 03:50 PM

[IronManny]

It's pretty hard to reach back for something extra when every pitch you're throwing is already 96 or 97. Seriously, look at gameday, the guy threw like 3 fastballs under 95mph the whole game. His pitching mechanics were an absolute clinic of consistency.

edit: and you can read EV's chart on the last page to prove it

I'm not sure where that gameday data is coming from exactly - but I'm a little surprised to see those numbers. I was at the game watching the gun readings all night, and according to those Beckett was around 94 with his fastball all night. I definitely didn't see any 97s or 96s. This reminds me of hearing about juiced up guns so Zumaya could break 102 in Comerica Park.

Miles Per Hour aside - Beckett was clearly filthy last night.

[Eric Van]

I'm not sure where that gameday data is coming from exactly - but I'm a little surprised to see those numbers. I was at the game watching the gun readings all night, and according to those Beckett was around 94 with his fastball all night. I definitely didn't see any 97s or 96s. This reminds me of hearing about juiced up guns so Zumaya could break 102 in Comerica Park.

Miles Per Hour aside - Beckett was clearly filthy last night.

Enhanced Game Day velocities are accurate, they're just measured a bit earlier in the path to the plate than the "slow gun" used at the ballpark. (The ball slows down continually as it approaches the plate to wind resistance; at the plate it's 9-10 mph slower.) The network was using a "fast gun" that apparently picks up the ball at the same point as the EGD cameras, because the two were agreeing.

Since the fast guns agree with the EGD readings, I think you'll see that become the standard (which will have the added advantage of giving old-timers more things to bitch about).

[Eric Van]

I agree with Eric. Given that the vertical break and the speed is comparable to that of the 4-seam fastball, I agree that this is a cutter and not a slider. A slider usually is thrown a bit slower and with more vertical drop than a fastball.

Hi, Alan! Are we going to need an Enhanced Game Day hacking thread?

I called the pitch a slider originally because I've heard that he throws one occasionally and didn't want people to say "Beckett doesn't throw a cutter." But it is a cutter, just like Tavarez's and Sabathia's sliders are slurves. It will be interesting to see whether the objective classification of breaking pitches by EGD hackers has any influence on the terminology. But right now it's a great tool for figuring out what a guy's repertoire is (Tavarez, e.g., has two different two-seamers, his bread-and-butter sinker and one which rides in much harder).

BTW, I have a DVD dub of the game where Dice-K threw a seemingly obvious gyroball -- but there was no EGD data for the game. And as I watch games in high def and look at the EGD data, it seems clearer than ever that a gyroball is essentially an intentionally thrown hanging slider that works because it doesn't have the expected break. The usual slider axis of rotation is not that far off from the bullet-spin gyro axis.

[Eric Van]

EV, thanks for posting the pitch data. I too love Enhanced Game Day. It allows for some interesting observations.

Here's a tip -- the first thing you should do if you're looking at pitch location is plot the human-entered data (x, y) against the computer-generated data. They don't always agree. Then (if possible) check the discrepancies against video of the game.

In the Beckett game, there were four pitches that the human got really wrong (11, 14, 19, 108). But the tricky part is pitches 42-46, Cabrera's PA leading off the 4th. The Px and Pz data for pitch 42 agree with x and y but it's not the pitch Beckett threw. The x and y data for pitches 43 to 46 belong to pitches 42 to 45. Very strange.

[alannathan]

Here's a tip -- the first thing you should do if you're looking at pitch location is plot the human-entered data (x, y) against the computer-generated data. They don't always agree. Then (if possible) check the discrepancies against video of the game.

In the Beckett game, there were four pitches that the human got really wrong (11, 14, 19, 108). But the tricky part is pitches 42-46, Cabrera's PA leading off the 4th. The Px and Pz data for pitch 42 agree with x and y but it's not the pitch Beckett threw. The x and y data for pitches 43 to 46 belong to pitches 42 to 45. Very strange.

Hi...Alan here again. Sometimes there are "operator errors" that get the pitches jumbled up, the most famous of which was the Bond's 756th homer pitch, in which a throwback pitch was recorded rather than the actual pitch. To my knowledge, there is no "human-entered (x,y)" data, only the computer-generated data. So, one cannot compare with that.

I am doing my own analysis of Beckett's pitches from last night and will post it here later on tonight.

[OttoC]

Hi...Alan here again. Sometimes there are "operator errors" that get the pitches jumbled up, the most famous of which was the Bond's 756th homer pitch, in which a throwback pitch was recorded rather than the actual pitch. To my knowledge, there is no "human-entered (x,y)" data, only the computer-generated data. So, one cannot compare with that.

The top and bottom of the strike zone is still operator-entered in Enhanced GameDay, although once enough data is built up it will be generated automatically. How well this will work with batters who change their stance depending on pitcher/situation remains to be seen.

Also, EGD has tweaked the break readings about three times by changing the point from which it is first measured. I believe my memory is correct in saying we were told that was produce results that were more in line with what sportscasters would accept. In other words, they were toned down. This means that comparing the pitch breaks for pitchers from game-to-game when comparing eary data to later data. A pitcher showing 80% of the break yesterday that he showed on day 1 is not necessarily pitching any differently.

[missinpedro]

Question how many pitchers are there out there capable of throwing a 2 seamer in the 96-98 mph range. I can only think of two who are close: prime Pedro whose 2-seamer topped out at 94-96, if I recall correctly, and Mariano at his best can dial up his 2 seamer to 95. Besides that I can’t really think of anyone getting their 2 seamer above 93. Looking at EV’s chart and Beckett’s effectiveness with all of his pitches, I would say that the only thing separating Beckett from Pedro in his prime is that Pedro had all the pitches Beckett does only his change-up was 10-12 mph slower and had much more movement. Granted that’s a huge difference and accounts for well over 1 ER per game, but its still pretty damn good to have your 4 seamer, 2 seamer, and cutter be Pedro-esque.

I said it in the game thread and I’ll paraphrase it here: it wasn’t quite Pedro but that was one of the more impressive combinations of pure stuff and cerebral pitching that I’ve ever seen. Not to get ahead of ourselves, but I think we haven’t seen the best of Beckett yet as far as regular season performance is concerned.

[alannathan]

The top and bottom of the strike zone is still operator-entered in Enhanced GameDay, although once enough data is built up it will be generated automatically. How well this will work with batters who change their stance depending on pitcher/situation remains to be seen.

Also, EGD has tweaked the break readings about three times by changing the point from which it is first measured. I believe my memory is correct in saying we were told that was produce results that were more in line with what sportscasters would accept. In other words, they were toned down. This means that comparing the pitch breaks for pitchers from game-to-game when comparing eary data to later data. A pitcher showing 80% of the break yesterday that he showed on day 1 is not necessarily pitching any differently.

What Otto says is exactly right. In my own analyses I have done of the pitch logs, I am able to correct for the different starting point. Actually, as a matter of convention, I compute the break starting 40 ft. from home plate. In the figure is a plot of the vertical vs. the horizontal break, which I calculate myself from the PITCHf/x data logs. These numbers are the computed difference between the actual location of the pitch as it crosses the front plane of home plate and the location expected in the absence of the Magnus force (computed from the 40 ft starting location). The plot is color-coded according to the release speed as follows: black (v>91 mph); red (82<v<91); green (v<82). I associate the green points with the curveball (break is down and to the catcher's right) and the red points with the changeup. You can see the black points can be subdivided into three groups depending on the amount of break in each direction. The four-seam fastball is the cluster around dx=-7.5, dz=10. I haven't yet had a chance to look at this plot in detail. Perhaps one of you out there might be interested in doing that and posting your remarks.

[Eric Van]

What Otto says is exactly right. In my own analyses I have done of the pitch logs, I am able to correct for the different starting point. Actually, as a matter of convention, I compute the break starting 40 ft. from home plate. In the figure is a plot of the vertical vs. the horizontal break, which I calculate myself from the PITCHf/x data logs. These numbers are the computed difference between the actual location of the pitch as it crosses the front plane of home plate and the location expected in the absence of the Magnus force (computed from the 40 ft starting location). The plot is color-coded according to the release speed as follows: black (v>91 mph); red (82<v<91); green (v<82). I associate the green points with the curveball (break is down and to the catcher's right) and the red points with the changeup. You can see the black points can be subdivided into three groups depending on the amount of break in each direction. The four-seam fastball is the cluster around dx=-7.5, dz=10. I haven't yet had a chance to look at this plot in detail. Perhaps one of you out there might be interested in doing that and posting your remarks.[/img]

You've inspired me to go to the work of posting my own chart, which is pretty much the same . . .

[JohntheBaptist]

Question how many pitchers are there out there capable of throwing a 2 seamer in the 96-98 mph range. I can only think of two who are close: prime Pedro whose 2-seamer topped out at 94-96, if I recall correctly, and Mariano at his best can dial up his 2 seamer to 95. Besides that I can’t really think of anyone getting their 2 seamer above 93.

Off the top of my head, I think Bartolo Colon in his prime could get that backup 2-seamer of his into this neighborhood, but I'm not 100% on his ability to sit at ~93 with it.

[Noah]

Carlos Zambrano throws a 2-seam fastball harder than I think I've ever seen from anyone. And when he hits his spot with it inside to lefties, it's ridiculous.

[Paul M]

In terms of sheer velocity, Felix Hernandez can throw a 2-seamer 97-99. And actually maintain it pretty well into the 7th inning. (Kevin Brown in his halcyon days was 95-96 all night and it had great sink).

I think there are actually quite a few (15 or so) that can get 93+ on the 2-seamer, but 96+ is a very short list.