A look at how our starters tire
by dixieflatline on Jun 9, 2008 4:14 PM CDT
71 comments
So for tomorrow's THT article I wrote about how fatigue effects pitcher's fastballs. At the risk of scooping myself I wanted to share how the Crew's starters fastball speed declines as they go through the game. I have left out Sheets as he is an integral part of the THT article so you will just have to wait until tomorrow to see how he fatigues. All bins are in groups of 10 pitches (so the first bin is 1 to 10 pitches the second 11 to 20 etc) and each bin has been divided by the first bin and multiplied by 100 to get a percent. This percent is how much more or less that group is than the baseline. Since most pitchers throw around 90 MPHs a 99% means the pitcher lost nearly 1 MPH compared to the start of the game. Anyway here goes.
Steady Jeff you always know what you are going to get from him. After a few pitches to get warm he stays at a pretty constant rate until about 100 pitches. Notice that he dips in the 30's 60's and 90's and then rebounds in the 40's 70's and maybe 100's. I wonder if those pitches tend to be thrown to the bottom part of the order and he is conserving his energy for later in the game?
Wow Manny is all over the map. It appears like his best fastballs are right in the beginning but it is hard to tell exactly how quickly it is dropping. This irregular pattern is very unusual for pitchers and may indicate that Parra doesn't have a lot of control yet as far as adding and subtracting from his fastball. Whatever the case, by about pitch 90 Parra has lost between 1-2 MPHs off the fastball so having a quick hook with him seems to make sense.
Wow Dave just Wow. Dave's best Fastballs come early and a pretty rapid decline shortly follows. For a right hander who throws a pretty straight fastball in the low 90's this is not a good sign at all. I wonder if he really belongs in the bullpen where he could max effort for a short period of time?
These are just from Villy's starts and I don't have enough data on McClung or YoGa to make a plot for them. Carlos seems to be ok until about 50 pitches and then sees a pretty fast decline. It sure seemed like when he was starting he would get into trouble around the 5th or 6th inning and maybe this loss in speed is why. The closer his fastball is to his change the worse it is for his change so maybe long reliever/guy who can go multiple innings is the best place for him as well.
9 recs |
71 comments
Email
Print
Comments
awesome stuff josh!
hey, don’t the (relatively) large error bars for parra suggest that he DOES have control adding and substracting velocity to his fastball?
and wow is right, staggering decline by bush. I’ve noticed villy’s decline in games, but never noticed bush’s decline. i think this is the future of pitcher management. not using pitch counts anymore as a proxy for fatigue, but rather just measuring fatigue.
Bring Back The Old Logo!
by jacob on Jun 9, 2008 4:33 PM CDT reply actions 0 recs
Hmm you make a good point about Parra.
I’ll have to do some more looking into him. Honestly, no other pitcher that I have looked at has such a strange fatigue curve. I just don’t know quite what to make out of it. I’ll be really curious to see if that continues though the ASB and beyond. I agree pitch counts by themselves are almost certainly too crude a metric. Hopefully a much improved metric involving PITCHf/x values will be created soon.
by dixieflatline on Jun 9, 2008 5:17 PM CDT up reply actions 0 recs
Perhaps different speeds for the two seam and the four seam fastball.
by ol Pete on Jun 9, 2008 5:34 PM CDT up reply actions 0 recs
So totally off topic but kind of cool
If you google PITCHf/x my blog is now the first site that appears followed by one of my hardballtimes articles, Dr. Nathan’s PITCHf/x page, and then sportvision’s page.
by dixieflatline on Jun 9, 2008 5:30 PM CDT reply actions 0 recs
yeah, because you're spamming my site with links
:)
Also, cheese.
by Jeff Sackmann on Jun 9, 2008 6:20 PM CDT up reply actions 0 recs
Links?
There were no links in this article just some pretty pictures. How can anyone get upset at pretty pictures :)
by dixieflatline on Jun 9, 2008 7:35 PM CDT up reply actions 0 recs
*this* time...
also, Google’s algorithm is getting more and more customized. I ran a few queries and you didn’t come up first on any of them. (You were always in the first 5, though.) I’ve read that they track certain aspects of your search history and tweak the results a little based on that. For instance, whenever I search for a university name, it ALWAYS gives me the athletic department no lower than the 2nd or 3rd result.
Also, cheese.
by Jeff Sackmann on Jun 9, 2008 7:45 PM CDT up reply actions 0 recs
Well if Parra is throwing a two seamer I am just not seeing it
The four seamer is clearly the cluster around 92 MPHs with vertical movement of around 10 inches. That is classic four seam goodness there. His change is around 84 MPH with a large spread in vertical movement but if there were some two seamers in there you would expect them to be in the high 80’s with about 5 inches of vertical movement.
by dixieflatline on Jun 9, 2008 5:46 PM CDT reply actions 0 recs
He throws a two seamer and an occasional cutter. He used to throw a splitter as well.
by ol Pete on Jun 9, 2008 10:44 PM CDT up reply actions 0 recs
Yeah I saw the split last year but like you said not this year
If he is indeed throwing a 2 seamer and a cutter they don’t move much differently than his 4 seamer.
by dixieflatline on Jun 10, 2008 8:47 AM CDT up reply actions 0 recs
More fantastic stuff, dixie
Thanks a bunch, man.
He's extremely quick and good.
by battlekow on Jun 9, 2008 6:09 PM CDT reply actions 0 recs
Bush
So why does faster = more runs for Bush. His early innings have been killer for the Crew yet he has more speed. Any idea how this correlate to his control/location over the same time period? Maybe he is overthrowing early for more gas but leaving everything out over the plate. He loses the velocity and gains control?
Great stuff Dixie!
The Wallbangers won because they played the game like kids! Let's do that again!!!
by 80badger on Jun 9, 2008 6:29 PM CDT reply actions 0 recs
That is the biggest problem with the PITCHf/x right now
How to calculate control is just not something that is obvious. You can do things like check how many pitches hit the corner or were thrown in the heart of the plate but neither is really a great way of doing that. It certainly is possible that Bush is indeed overthrowing at the start and that is causing problems until he settles down.
by dixieflatline on Jun 9, 2008 7:29 PM CDT up reply actions 0 recs
Very cool
I actually found it interesting how relatively little fatigue effects the throwing speed of all of them. 1-2 mph is probably significantly more important than it looks on paper, but it’s still less than I would have expected at the 100-pitch mark.
Any plans to work out something similar for pitch control/placement? Not sure how it would be possible to do that, but it would be interesting to see the correlation between loss of control and loss of velocity.
"My thing is that you guys eat meat all day and you can't hit a ball. That's my thing." - P. Fielder
by SRB on Jun 9, 2008 7:04 PM CDT reply actions 0 recs
Yeah I was shocked at the smallness of the differential as well
Tomorrow you can the combination for all pitchers put together and that differential is really tiny. It is definitely counterintuitive but definitely true because there is so much data right now.
by dixieflatline on Jun 9, 2008 7:31 PM CDT up reply actions 0 recs
super cool, but I'm confused about the scale
Those patterns, especiall the cyclical patterns that you see for Soup and Manny somewhat, are really intriguing.
I don’t understand the y-axis though. I first assumed that “fastball speed percent” was as a percentage of the pitcher’s average fastball speed. How come Suppan’s pitches are almost all over 100% and Bush’s are almost all under 100%? If they are as a fraction of the average pitcher’s fastball, than it would look like Suppan is the hardest thrower of the bunch, which can’t be. Any help?
by keephopealive on Jun 10, 2008 7:38 AM CDT reply actions 0 recs
They are all a percent compared to the baseline 10 pitches
This should show warm up time and fatigue quite well and also normalize out if a pitcher doesn’t have his good fastball that game. It all is about how the pitcher’s fastball changes from the initial 10 pitches.
by dixieflatline on Jun 10, 2008 8:49 AM CDT up reply actions 0 recs
yep, sorry
I’m an idiot, since you said it up top. I guess I should learn to read the instructions before jumping down to the cool pictures.
by keephopealive on Jun 10, 2008 10:53 AM CDT up reply actions 0 recs
Could you clarify?
So these percentages only compare pitch speed within a given start, correct? You’re then averaging out all of those comparisons for the 2008 season? Meaning, within a particular start, Jeff Suppan’s fiftieth-pitch fastball, will be about 1% faster as measured against his first-pitch fastball in that same start? So that’s what we should basically expect on Saturday.
What you’re not doing, is taking his fiftieth pitch fastballs lumped together from each of his 13 starts (he’d have had maybe 8 of them), then comparing them to the first-pitch fastballs from all 13 of his starts lumped together (say 12). Is that correct?
Hopefully my questions are clear. I’d be happy to help clarify them if they’re not.
Thank you for the interesting post.
Don't try to do too much with it. Just take the ball the other way.
by shooty babitt on Jun 10, 2008 9:22 AM CDT reply actions 0 recs
I would certainly hope that he'd have a first fastball in every start
Otherwise he had a really bad day.
He's extremely quick and good.
by battlekow on Jun 10, 2008 9:25 AM CDT up reply actions 0 recs
Yeah this is an important point and I haven't found a great way to word this so bear with me
The second bin is pitches 11-20. For each start I take the fastballs in that bin and divide them by the fastballs in the first bin (1-10) so I get a value for the second bin for each start. I then combine all 8 and the results you see are the combined values compared to the first 10 from the same start. The reason I am doing this is because sometimes a pitcher doesn’t have his good fastball and everything is shifted down a bit but by comparing to the fastball he has for a particular game it really shows how fatigue is effecting him. If I did it the other way and average the second bin first and then divided I would have a small bias of how good of a fastball he had on each day.
by dixieflatline on Jun 10, 2008 9:49 AM CDT up reply actions 0 recs
Thank you...
.for the clarification, and because I was struggling to write a long clarifying email myself given battlekow’s comment.
The bias problem is why I asked the question. The way you have chosen to do it controls for things like weather conditions, minor health concerns, etc. or anything else that might shave an mph or two of a fastball within a given day.
I empathize with your struggle to word it, because I was having a hell of a time even wording the questions.
Don't try to do too much with it. Just take the ball the other way.
by shooty babitt on Jun 10, 2008 9:55 AM CDT up reply actions 0 recs
Yeah if I could do a better job at simply and concisely wording the math I'd be a much better writer
I am trying and I think I am getting better but I still have a ways to go.
by dixieflatline on Jun 10, 2008 10:00 AM CDT up reply actions 0 recs
You're definitely getting better, no worries
This is some of the best work out there, dixie.
On a note unrelated to the data itself but to the discussion of language, I find it interesting that “bin” and “bucket” have become so widely used to describe clusters of data points in baseball studies. Just something I’ve noticed.
He's extremely quick and good.
by battlekow on Jun 10, 2008 10:02 AM CDT up reply actions 0 recs
bin
bin is a standard statistical analysis term
bucket? not so sure:P
by PagsBrewCrew on Jun 10, 2008 10:15 AM CDT up reply actions 0 recs
Yeah I don't think I have used bucket before but I have seen it around
Not quite sure what to make of bucket. I tend to like cluster myself. I am glad you think I am getting better battlekow it is something I am trying very hard to improve.
by dixieflatline on Jun 10, 2008 10:56 AM CDT up reply actions 0 recs
I seem to see "bucket" all the time
but actually, I’m thinking of GMAT prep, where people talk about “buckets” of questions, as differentiated by difficulty. Same idea, somehow the awful word caught on.
Another term I’d like to see die is “drill deeper.” We’re looking for insight here, not oil. (Then again, if I were writing a script one day to isolate someone’s true talent level and I struck oil, that’d be awesome.)
Also, cheese.
by Jeff Sackmann on Jun 10, 2008 11:08 AM CDT up reply actions 0 recs
True prospects have oil in their veins.
Obscure baseball records and more at my blog, Recondite Baseball.
by TheJay on Jun 10, 2008 11:19 AM CDT up reply actions 0 recs
so THAT's why they get the big bonuses.
Also, cheese.
by Jeff Sackmann on Jun 10, 2008 11:27 AM CDT up reply actions 0 recs
Ha
Never would’ve seen that reference coming. Speaking of which, I hear they are working on another movie, although it’s suppose to be more along the lines of the earlier episodes as more of a stand alone storyline.
"He's been very, very impressive," Yost said. "I mean really impressive. I mean really, really impressive."
by MadJimiBrewha on Jun 10, 2008 2:57 PM CDT up reply actions 0 recs
The monster of the week episodes are so much better than the mytharc ones
The conspiracy made absolutely no sense by the end of the show. I’m happy to see them going back to flukemen and ice worms.
He's extremely quick and good.
by battlekow on Jun 10, 2008 3:05 PM CDT up reply actions 0 recs
Apparently there is a hole in one of my buckets
by dixieflatline on Jun 10, 2008 11:22 AM CDT up reply actions 0 recs
Well fix it, dear Dixie,
dear Dixie
dear Dixie
well fix it, dear Dixie
dear Dixie
fix it!
"I will agree that the attitude [at BCB] is ridiculous and they have done so much to instigate animosity and then block us from responding. Real mature!"
by roguejim on Jun 10, 2008 2:24 PM CDT up reply actions 0 recs
With what shall he fix it?
Dear Jimmy, dear Jimmy?
With what shall he fix it?
Dear Jimmy,
Fix it!
He's extremely quick and good.
by battlekow on Jun 10, 2008 2:31 PM CDT up reply actions 0 recs
The longest 3:48 of your life
Someone in the comments says the rabbit (the live grey one) looks like it wants to die. I don’t blame it.
"I will agree that the attitude [at BCB] is ridiculous and they have done so much to instigate animosity and then block us from responding. Real mature!"
by roguejim on Jun 10, 2008 2:44 PM CDT up reply actions 0 recs
Great Stuff
Good read. I enjoyed your THT article especially about the movement loss over time on the fastball. Is it possible to plot that as well for our beloved Crew?
by SgtClueLs on Jun 10, 2008 9:40 AM CDT reply actions 0 recs
Josh, I submitted your THT piece to BTF
Hopefully it will go up and generate some good discussion.
He's extremely quick and good.
by battlekow on Jun 10, 2008 10:11 AM CDT reply actions 0 recs
Excellent thanks.
Yeah the more input the better with things like this. That is why I try to fully explain everything such that if someone wanted to they could recreate the results. You get much better user input than the black box techniques that some websites seem to favor.
by dixieflatline on Jun 10, 2008 10:59 AM CDT up reply actions 0 recs
Here you go
He's extremely quick and good.
by battlekow on Jun 10, 2008 2:19 PM CDT up reply actions 0 recs
Ok so here are the movement fatigue curves
Pretty consistent actually and not really following the league trend of a big drop early and then kind of hanging on. I wonder if the large error bars for Parra and Suppan are lack of statistics or due to fatigue messing with their mechanics? They both are very consistent before 90 pitches. Villy takes a huge dive after 100. That probably also makes his fastball and change up move closer together as well. Bush seems rather consistent up to 50 pitches and then drops a bit but stays consistent at his new level until maybe ramping it up right before he leaves.
As for Sheets I have no idea what is going on. Remember it takes him over 30 pitches to hit his peak for his fastball speed so maybe that early dip is his arm moving faster but the spin not catching up? Not sure if that makes sense. In any case once he reaches peak for speed he has solid movement until about 90 pitches when he drops again.
by dixieflatline on Jun 10, 2008 12:39 PM CDT reply actions 2 recs
This might be an interesting way to examine the old saw that sinkerballers get better movement when they're a little fatigued
He's extremely quick and good.
by battlekow on Jun 10, 2008 2:21 PM CDT up reply actions 0 recs
Great Stuff
I’m glad we get an inside scoop on your projects. It is very informative and greatly appreciated. Thanks
"He's been very, very impressive," Yost said. "I mean really impressive. I mean really, really impressive."
by MadJimiBrewha on Jun 10, 2008 1:06 PM CDT reply actions 0 recs
Great Work!
I would just be a little concerned about trying to identify trends from a relatively small number of starts from so early in the season. Also, Sheets was coming back from injury (and had that tricept issue), Parra wasn’t even starting for most of last year and Bush seemed to have some issues with velocity over his first two or three starts. We might see some very different trends from those 3 guys come September.
Is the data available to repeat this effort with the 2007 data?
by rluzinski on Jun 10, 2008 2:49 PM CDT reply actions 0 recs
russ!
welcome. hope you’ll continue to visit, and contribute, i know your insights would be much appreciated here.
I agree with russ, that another look at this data in a couple months could be quite interesting, and a comparison to ‘07 would be interesting as well. Also, I think it’d be pretty interesting to look at two bush buckets, pre- and post-AAA stint and see if his pitches have changed at all as a result of some supposed mechanical tweaks.
Finally, I think Josh does a great job of not going to far with drawing conclusions from the data and focuses much more on simply presenting it.
Bring Back The Old Logo!
by jacob on Jun 10, 2008 3:17 PM CDT up reply actions 0 recs
I miss you Russ
I should visit more often at brewerfan. I really enjoy your blog BTW. There is data from 2007 to repeat this and if I get a chance maybe I will repeat the experiment. The problem with the 2007 data though is it is so hit and miss. You can have a start recorded and then nothing for four starts then two get recorded and then nothing for three starts. You are absolutely right about the small sample here and that is why I try hard to add error bars to things like this. To openly show what the small sample means.
I am also planning on doing work about how pitchers fatigue as the year goes but obviously I need more data for then. Once that is established then I can look at things like reworking pitcher abuse points and how damaging a 120 pitch start really is.
BTW, isn’t that was a brewer fan is all about? Trying to identify trends from a relatively small sample and applying that to how much Ned Yost sucks?
by dixieflatline on Jun 10, 2008 3:47 PM CDT up reply actions 1 recs
error bars
stdev or standard error? or something kooky like confidence interval?:P
by PagsBrewCrew on Jun 10, 2008 5:03 PM CDT up reply actions 0 recs
stdev
by dixieflatline on Jun 10, 2008 5:57 PM CDT up reply actions 0 recs
Glad to be here
You guys do some nice work around here. I’ve accidently stumbled over here enough times to know that I need to check it out more often. I’m still not even sure how things work around here. Anyone can write a “Fan Post”?
You can still get some nice information in the major league forums at bf.net but the arguments over there can get pretty silly, pretty fast. For whatever reason, I still feel compelled to fight the fight, however. I guess I’m a battler.
by rluzinski on Jun 11, 2008 7:48 AM CDT reply actions 0 recs
Yup, anyone can write a FanPost
Basically, it’s a short article. Anyone can also write a FanShot, which is just a quick link to an article, picture, video, etc. That’s what those blue highlighted things on the main page are, FanShots that we’ve “frontpaged”. We usually put one up as a post-game thread if the Brewers won, the Cubs are still on, or there’s any other reason that people are going to hang around and chat after the game.
He's extremely quick and good.
by battlekow on Jun 11, 2008 8:28 AM CDT up reply actions 0 recs
That just means you are gritty Russ
In any case we would love to have you check the stuff out over here.
by dixieflatline on Jun 11, 2008 8:47 AM CDT up reply actions 0 recs
I agree
I think it’s more of a habit to hang in BF.net. The threads aren’t very well researched and half the arguments are “from the gut”. There are some good posters there, but a pretty high quotient of bad ones. I can thank Marty or posting on FNY.com a link here once, and now every day make the Frosty (Or Plastic) Mug one of my morning reads.
by SgtClueLs on Jun 11, 2008 9:46 AM CDT up reply actions 0 recs
Could we print this thread...
on the toilet paper used in manager’s biffy?
by heybatterbatter on Jun 19, 2008 10:04 PM CDT reply actions 0 recs
Er, biffy?
He's extremely quick and good.
by battlekow on Jun 19, 2008 10:09 PM CDT up reply actions 0 recs
Err...
“I scoffed at bidets until I went to Europe. I have wanted to do something about it ever since. I am the opposite of a handyman so I put it off… for 24 years! finally I ordered the Biffy. Installed in 10 minutes with just the two tools you described. Why did I wait? It is great. “Refreshing” is a word you keep using. It sounded silly. Well, it is refreshing, darn it. This product is a 10 out of 10. Thanks.”
It’s great after an interactive gangbang.
"I will agree that the attitude [at BCB] is ridiculous and they have done so much to instigate animosity and then block us from responding. Real mature!"
by roguejim on Jun 20, 2008 10:38 AM CDT up reply actions 0 recs
Nice...REAL nice.
Imagine trying to get your bottom clean with a child’s squirt gun – then imagine using a small showerhead. Now you have some idea of the difference. This feature is patented.
Obscure baseball records and more at my blog, Recondite Baseball.
by TheJay on Jun 20, 2008 11:03 AM CDT up reply actions 0 recs
Umm
Really, very little good can come from a sentence starting with “Imagine trying to clean your bottom.”
"I will agree that the attitude [at BCB] is ridiculous and they have done so much to instigate animosity and then block us from responding. Real mature!"
by roguejim on Jun 20, 2008 11:12 AM CDT up reply actions 0 recs
Great reading
in the can for Ned while he ponders his moves before heading to the field.
Can’t hurt, can it?
by heybatterbatter on Jun 19, 2008 10:06 PM CDT reply actions 0 recs
I'd really like to think that someone over at Miller Park is reading
I don’t think it is Ned but it would be nice if someone in the front office is seeing this stuff and maybe replicating/improving on it. Who knows?
by dixieflatline on Jun 21, 2008 3:53 PM CDT up reply actions 0 recs
Maybe Tom H can point it out to Ned
He's extremely quick and good.
by battlekow on Jun 21, 2008 4:01 PM CDT up reply actions 0 recs
yeah
I hear Tom H likes gangbangs.
Also, cheese.
by Jeff Sackmann on Jun 21, 2008 4:42 PM CDT up reply actions 0 recs
If there is something TH likes more than gangbangs
by dixieflatline on Jun 21, 2008 4:45 PM CDT up reply actions 0 recs
Nah, first he has to say "gangbangs--whatever those are"
Then, 45 minutes later, he posts a blog entry extolling the virtues of Brewers gangbangs.
He's extremely quick and good.
by battlekow on Jun 21, 2008 5:01 PM CDT up reply actions 0 recs
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recent FanPosts
NL Central Standings
| W | L | PCT | GB | STRK | |
|---|---|---|---|---|---|
St. Louis
|
91 | 71 | .561 | 0 | Lost 6 |
|
Chicago
|
83 | 78 | .515 | 7.5 | Lost 1 |
|
Milwaukee
|
80 | 82 | .493 | 11 | Won 3 |
|
Cincinnati
|
78 | 84 | .481 | 13 | Won 2 |
|
Houston
|
74 | 88 | .456 | 17 | Lost 3 |
|
Pittsburgh
|
62 | 99 | .385 | 28.5 | Lost 2 |
(updated 11.9.2009 at 4:19 AM CST)
Certain photos copyright © 2009 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
