Phil Hughes

[drjim]

From what I'm looking at, the 4-seem was his best pitch last year. This year its his worst pitch.

 

I wonder about the accuracy of pitchf/x's fastball classifications, but, it says he's added a 2-seemer this year? The results on that are quite a bit better than his 4-seem.

 

The cutter has straightened out as well.

I'm not 100% how they pull pitch f/x but if it is the same service that mlb.com uses there is reason to be suspect on how they differentiate pitches.

 

The publicly available data might not be the best, but should at least be directional.

[jimmer]

'PITCHf/x is a pitch tracking system, created by Sportvision, and is installed in every MLB stadium since around 2006. This system tracks the velocity, movement, release point, spin, and pitch location for every pitch thrown in baseball, allowing pitches and pitchers to be analyzed and compared at a detailed level. Two mounted cameras in each stadium are used to track each pitch and establish the each aforementioned aspect.'

 

Probably better than us fans or the TV guys trying to guess what was thrown.  Since release point is seen, the grip is too. And it's not like it's new, they've had 10 seasons to tinker with it to make it the best they can. Actually, longer than 10 years since they tinkered with it before being used at the major league level. :-)

[drjim]

What I'm saying is if the output from pitch f/x is the same we get in real time on mlb.com then the pitch identification is very inconsistent. Direction, spin, location are probably acceptable.

 

One of my jobs is to track every pitch and use pitch f/x as a check. Their pitch identification leaves something to be desired.

 

If this is the data that is then ripped for the pitches you cite above, that would explain some of the questions you have.

[nicksaviking]

Don't fly ball guys generally have lower babips?

Yes because HR aren't "in play" which of course is a bit of an issue.

[nicksaviking]

I wonder if his large drop in velocity is due to his choice of fastball or if his choice of fastball is causing the drop in velocity.

 

RHP don't get many Ks with a 90 MPH heater.

[jimmer]

 

One of my jobs is to track every pitch and use pitch f/x as a check. Their pitch identification leaves something to be desired.

 

Assuming, of course, you and/or whatever you are using to double check pitch fx is 100% accurate where you can say, for certain, their pitch identification is very much lacking. Maybe the truth is somewhere in between. Maybe most pitches are obviously one type or another and there's only a minimal amount that are misidentified making the huge difference in pitch selection still valid even if a few pitches are slightly misidentified.

 

In any event, we can only use info available to us to discuss (in this case as well as advanced stats). When looking at the same system's info (their imperfect, but consistently imperfect, year to year data) from 2013, 2014, and 2015, there has been a large change in his pitch selection. Maybe it's not quite as large due to slight imperfections in pitch fx data or maybe it's slightly larger due to slight imperfections in pitch fx data.

 

His regular curve is pretty much gone where it was a big pitch for him in 2013.  His cutter become a very good and often used pitch for him in 2014 and still is in 2015 and his two seamer is a bad pitch he's decided to use even more.  His fastball has also lost velocity and it's costing him

[kab21]

I also don't like this version of Hughes as much going forward.  It is great that he isn't walking anyone but it is tough to be a fly ball pitcher that doesn't strike out many batters.  he is no longer potentially an ace lite but rather a solid 3/4.  Nothing wrong with that but the Twins really needed him to be a #2.

 

 

 

 

[Brock Beauchamp]

Yes because HR aren't "in play" which of course is a bit of an issue.

Also infield pop flys, which drive down BABIP in a good way.

[Secondary User]

 

From what I'm looking at, the 4-seem was his best pitch last year. This year its his worst pitch. 

 

I wonder about the accuracy of pitchf/x's fastball classifications, but, it says he's added a 2-seemer this year? The results on that are quite a bit better than his 4-seem.

 

The cutter has straightened out as well.

 

As someone who works in the sports media industry, I can tell you that it's usually ok, but will have it's fair share of times where it's just garbage

[drjim]

As someone who works in the sports media industry, I can tell you that it's usually ok, but will have it's fair share of times where it's just garbage

Don't let jimmer see this.

[DaveW]

I also don't like this version of Hughes as much going forward. It is great that he isn't walking anyone but it is tough to be a fly ball pitcher that doesn't strike out many batters. he is no longer potentially an ace lite but rather a solid 3/4. Nothing wrong with that but the Twins really needed him to be a #2.

I still think he has the potential to be that #2 type, the results have certainly been there as of late.

[kab21]

 

I still think he has the potential to be that #2 type, the results have certainly been there as of late.

Really?

 

May - K rate BB rate ERA - 10.8% 2.9% 4.64 ERA

June - 13.9% 2.4% 3.76 ERA

July - 10.6% 1.9% 2.49 ERA

 

I think you are dealing with sample size issues (4 games).  If he is K'ing 12-13% of hitters he is most likely a high 3's ERA and possibly even a low 4's ERA guy.  He needs to return to the 18-20% rate (last year 21%) to be a #2.

 

 

[DaveW]

His HR rate was also overdue to come down.

[kab21]

 

His HR rate was also overdue to come down.

Why is this relevant?

[DaveW]

Why is this relevant?

Because that helps the old ERA quite a bit as well.

[kab21]

 

Because that helps the old ERA quite a bit as well.

 

Of course it would but his HR/FB% would have to be abnormally high in the first place for it to help.  It isn't.  He is a FB pitcher that doesn't strike out many batters.  That means that there are a lot of FB's and he is going to give up a lot of HR's.  Basically he is going to be among the league leaders in HR's every season so it really doesn't help the ERA.

[jimmer]

 

I still think he has the potential to be that #2 type, the results have certainly been there as of late.

Hughes has been in the league a long time and in 4 or the last five seasons (this one is included) he's had a FIP of 4.50 or higher.  If he was going to be a #2, he'd be one.  Last season was a huge outlier.  

[jimmer]

 

Don't let jimmer see this.

why not?  He's agreeing with me.  He just said it's usually okay with pitch id (notice the post he responded was about pitch classification).  You said it's good for some things but basically very inconsistent for pitch identification and their pitch identification leaves something to be desired. Your statement doesn't jive with his usually okay assessment of pitch ID.

 

And, also, we still have to wonder what flaws the system he is sourcing has itself, just like we do with yours. Should we believe they systems you too look at are perfect so as to be sure every time those say pitchfx is wrong it's for sure wrong instead of it being your system that's wrong?

 

In any event,  I don't assume pitchfx is perfect.  I never said it was, but for the conversations we were having, it was more than fine to use.

[drjim]

 

why not?  He's agreeing with me.  He just said it's usually okay with pitch id (notice the post he responded was about pitch classification).  You said it's good for some things but basically very inconsistent for pitch identification and their pitch identification leaves something to be desired. Your statement doesn't jive with his usually okay assessment of pitch ID.

 

And, also, we still have to wonder what flaws the system he is sourcing has itself, just like we do with yours. Should we believe they systems you too look at are perfect so as to be sure every time those say pitchfx is wrong it's for sure wrong instead of it being your system that's wrong?

 

In any event,  I don't assume pitchfx is perfect.  I never said it was, but for the conversations we were having, it was more than fine to use.

 

He said exactly what I said. My job doesn't produce data that is used by mlb (it is for other sources for other purposes) so it doesn't even matter what our output is as far as pitch f/x is concerned.

 

What I'm saying is that the exact same pitch can be logged as different types from pitch to pitch and especially from start to start. So if you are looking at different pitch usages from year to year or even start to start, it is from inconsistent pitch logging. I was trying to provide some context for people that are looking at pitch type and potential differences from year to year.

 

They do a very sound job with speed, location and break - because that is the automated aspect of pitch logging. The human part, pitch type logging, is extremely inconsistent. I'm not saying I'm "better" I'm saying their logging often makes no sense, unless you think the same pitch with the same speed and same break should be classified as different pitches.

 

Pitch f/x is fine, but for pitch types you might find more variance than you would think, and it is perfectly possible, if not likely, that Hughes hasn't actually done anything all that different in the comparisons that you are making.

[Secondary User]

Let me give an example.  For most pitchers, it's not a big deal.  When you have a guy throwing a traditional FB/CB/SL/CH set it's usually easy.  The pitches usually have enough speed/break differences that distinguishing them is easy.  But I covered a Marlins game where David Phelps pitched.  He has 4 pitches that regularly register in the high 80s, including a cutter and slider that don't feature a heck of a lot of movement.  Guys like this, it seems like half the time the id'er is just guessing haphazardly. 

 

Most of the time, it's not a big deal, but 'most of the time' is pitchers who have easily distinguishable pitches

[nicksaviking]

 

Let me give an example.  For most pitchers, it's not a big deal.  When you have a guy throwing a traditional FB/CB/SL/CH set it's usually easy.  The pitches usually have enough speed/break differences that distinguishing them is easy.  But I covered a Marlins game where David Phelps pitched.  He has 4 pitches that regularly register in the high 80s, including a cutter and slider that don't feature a heck of a lot of movement.  Guys like this, it seems like half the time the id'er is just guessing haphazardly. 

 

Still, the radar gun isn't guessing.  MLB hitters don't swing and miss at too many 90 MPH heaters from RHPs regardless of the type of fastball it is.