Jump to content

Providing independent coverage of the Minnesota Twins.
Subscribe to Twins Daily Email

The Forums

Article: TB 8, MIN 6: Twins Swept on Gomez Walk-off Homer

Minnesota Twins Talk Today, 06:05 AM
The Twins were haunted by a couple of their old friends this weekend. Denard Span wore out the Twins all weekend, and Carlos Gomez finish...

Article: Game Thread: Twins@Yankees 4/23 6:10PM

Minnesota Twins Talk Today, 06:04 AM
It’s big four game series that kicks off tonight in the Bronx as the Twins and Yankees go to battle in a series that can really help the...

Article: Twins Extend Phil Hughes

Minnesota Twins Talk Today, 04:22 AM
Mark Feinsand from the New York Daily News is reporting this morning that the Twins and Phil Hughes have agreed to a three-year contract...

Article: Twins Minor League Report (4/22): Romero Strikes...

Twins Minor League Talk Yesterday, 11:16 PM
Weather has gotten in the way of at least one minor league game every Sunday this season. That trend continued this week with the Lookout...

Article: Puerto Rican Dream

Minnesota Twins Talk Yesterday, 11:00 PM
They say you should write your dreams down right when you wake up in order to remember them clearly. After returning from a dreamlike day...

Photo

Understanding the "Quality of Pitch" (QOP/QOPA/QOPV) Statistics

Posted by Matthew Lenz , 08 January 2018 · 671 views

quality of pitch (qop)
Understanding the "Quality of Pitch" (QOP/QOPA/QOPV) Statistics I've spent a lot of time over the last few days reading about a relatively new statistic called "quality of pitch" (QOP), which assigns a numerical value to each pitch a pitcher throws. The values can then averaged together to come up with a pitchers average quality of pitch (QOPA) or you can look at a quality of pitch set of values (QOPV) as another tool to measure the performance of a pitcher. The purpose of this post is to provide a simple overview of this data as it may be referenced in future articles.

Background
QOP was first publicly introduced in March 2015 by Jason Wilson and Wayne Greiner. Since then it has been written in various publications such as "Baseball America", the "Fangraphs", and by Yahoo Sports! columnist Jeff Passan among others. Meanwhile, Wilson and Greiner have presented their findings at the 2015 SABR Analytics Conference. In short, this statistic was introduced and quickly regarded as a good tool to measures a pitchers performance in a way the baseball community has not previously done before.

Computation
QOP is computed by integrating velocity (MPH), pitch location, and pitch movement. Pitch movement is defined as the vertical break, horizontal break, breaking distance, and/or rise. These variables are put together and assigned a number 0 - 10, where 0 is a very poor pitch and 10 is an excellent pitch. The MLB average QOP is 4.5 and median is 5.
Posted Image

Here is an example of QOP being used.
Posted Image
Validation
Wilson and Greiner have measured QOP against ERA, FIP, and SIERA which all produced a strong, negative correlation. That is, the better the QOP the lower the ERA/FIP/SIERA.
Posted Image
Furthermore, a search of the top 10 2017 QOPA leaders for pitchers who threw 1,000 or more pitches provides you with a list of some of the more effective pitchers in baseball.
Posted Image

Limitations
As with all stats, QOP has its limitations. From a mathematical perspective anytime we are averaging numbers together the data can be skewed by outliers, and QOP is no exception to this rule. To help minimize the effect of outliers Wilson and Greiner have created a guide to determine the margin of error depending on the sample size.

From a baseball perspective, QOP doesn't take into account of a pitcher who misses his spots. That is, if the catcher calls for a fastball high and inside but the pitcher throws it low and outside he could still get a high QOP score despite completely missing his spot. If technology exists for the location and break of each ball to be tracked, then I would like to see something developed that also accounts for the movement of the catcher's glove.

Author's Conclusion
Again, this post was solely meant to introduce you to this stat without diving into specifics on Twins pitchers. Personally, I look forward to using this stat and wouldn't be surprised if we start seeing it more and more in future posts by me or any other Twins Daily writer. Despite its limitations, I think it provides fans with a different, more insightful perspective than the traditional pitching stats (W/L, ERA, WHIP, etc.), especially when coupled with other SABR pitching stats.

I also wonder how well this stat can be used to predict future outcomes. I look at the list above and a couple names surprised me, but specifically Joe Biagini who was also a top 10 QOPA guy in 2016 under the same criteria. A quick look at his fangraphs page shows that he hasn't been great in 162.0 big league innings. Is this the sign of a good pitcher who has just had some bad luck early in his career? Or is he the poster child for how finding the average QOP can, at times, be a misleading statistic?

What do you guys think about this stat? Is this something you would look forward to seeing in future articles? What are your thoughts in the curious case of Joe Biagini?

  • Parker Hageman, jorgenswest, Oldgoat_MN and 2 others like this



Photo
Tom Froemming
Jan 08 2018 09:02 PM

I enjoyed this, thank you. I've been looking at this stuff some, but I'm not certain how to best apply it. For example, it makes me awfully suspicious that Kyle Gibson ranks ahead of Clayton Kershaw in QOPA. I always like to have another data point, but I just don't know how valuable it is.

    • Oldgoat_MN, mikelink45 and Matthew Lenz like this
Photo
Matthew Lenz
Jan 08 2018 09:27 PM

 

I enjoyed this, thank you. I've been looking at this stuff some, but I'm not certain how to best apply it. For example, it makes me awfully suspicious that Kyle Gibson ranks ahead of Clayton Kershaw in QOPA. I always like to have another data point, but I just don't know how valuable it is.

Interesting finding.This may be my next task at hand...to figure out some of these discrepancies.

 

Another limitation I thought of after I posted is that it doesn't take into account certain situations.For example, a pitcher ahead 0-2 could very likely throw a "garbage" pitch just to get the batter to chase.This pitch, although not bad at all, would probably result on a bad QOPV.Does that happen enough to make Gibson > Kershaw?Probably not, but could be a partial factor.Hope to dig in a little bit more here.

    • Tom Froemming likes this
I am nowhere near a metrics guy. But this one seems so flawed by its inability to consider location. Pitching is like real estate, location, location, location. While this seems a good tool to measure "stuff", it would sure seem to ignore control. While I am sure there are far better examples, Samuel Deduno comes to mind. Electric "stuff", but couldn't hit a barn door with it.
Photo
Matthew Lenz
Jan 09 2018 11:34 AM

 

But this one seems so flawed by its inability to consider location. 

I somewhat agree with you.It definitely does take into account location (mentioned in first sentence under "computation" heading), but I think the bigger flaw is that it doesn't take into consideration the situation.I mentioned this in my comment above to Tom...wish I would have included in the original post.But an 0-2 slider in the dirt gets you a bad QOPV although the location might be considered "good" due to the situation.Compare that same exact pitch to a 3-1 count and now we have a terrible pitch due to the situation and should earn a bad rating.This metric gives those two pitches the same rating.

 

At the same time, is there a statistic that gives you a single numerical value to measure a pitchers "stuff"?I am not aware of one, which directly from my post, "I think it provides fans with a different, more insightful perspective than the traditional pitching stats (W/L, ERA, WHIP, etc.), especially when coupled with other SABR pitching stats."

 

FYI, Deduno finished his career (assuming it's done) with a 2.3 WAR and a 4.44 QOPA. A 2.3 WAR is probably higher than you would have guessed based off of your post, yet a 4.44 QOPA is below the major league average.  

    • Muss0070 likes this
Photo
Andrew Thares
Jan 11 2018 01:27 PM

I have actually been looking into Joe Biagini as a potential breakout player in 2018. I think a large part of what lead to Biagini's 5.34 ERA in 2017 was his 61.5 LOB % and his 15.2 HR/FB %. These numbers were much worse than the 71-73 LOB % and 4-6 HR/FB % rates that he had from 2014-2016. While I don't necessarily expect that these numbers will go all the way back to his 2014-2016 numbers, I do see them regressing back to more of Joe Biagini's norms in 2018 thus making him a better pitcher.

    • Matthew Lenz likes this

Interesting post on the quality of pitches, however it is difficult to understand from a normal readers view.How does FIP or SIERA measure to determine an effective pitcher?I understand ERA is earned runs per 9 innings.How does a pitcher with a 2.00, 3.00, 4.00, 5.00 ERA fall into the QOP Scale if between 4-5 QOP rating is the average? 

 

I do not believe your line graph makes a whole lot of sense to me.According to the graph, as the “ERA/FIP/SIERA” increases (vertical y axis), the QOP Pitcher Effect (horizontal x axis) decrease by -0.5?So, if I am looking for the median of 5 QOP, which is the 0.0 x axis, the pitcher has an average ERA of approximately 4.75, FIP of 4.5, and a SIERA of 4.25?This line graph is difficult to understand. 

In my mind, a pitcher with an ERA of 4.75 would have a much lower QOP.I believe if you made the x axis scaling between 1-10 (and correlate to the QOP Scale that has QOP buckets of 0-2, 2-4, 4-5, 5-6, 6-8, and 8+), then this would give the reader a better understanding of how ERA/FIP/SIERA is truly affected by QOP. 

 

The line graph however depicts a different message to me, and is not easy to interpret for the average “layman non-statistics major” like myself.There is good information here, however I believe the message was not delivered clear enough of how having a QOP in the various buckets relates to a pitcher in the average bucket. 

 

1.Where does the twins starting pitchers fall into each bucket on the QOP scale? 
2.Where does the twins bullpen stand compared to a team with an elite bullpen (such as the Yankees or Dodgers)? 
3. How does Ervin Santana (Near Elite) compare to Clayton Kershaw (Elite)?
4. What was Fernando Rodney’s QOP in his successful saved games versus his blown saves in 2017? 
5. What about Matt Belize in 2017 (9 saves and 2 blown saves)?

6.Can this be translated to hitters as well?Such as, what was the QOP of the at bat that Joe Mauer hit a home run, versus the QOP of the AB's that Mauer slapped a single?

Photo
Matthew Lenz
Jan 11 2018 04:24 PM

 

Interesting post on the quality of pitches, however it is difficult to understand from a normal readers view.How does FIP or SIERA measure to determine an effective pitcher?I understand ERA is earned runs per 9 innings.How does a pitcher with a 2.00, 3.00, 4.00, 5.00 ERA fall into the QOP Scale if between 4-5 QOP rating is the average? 

FIP is fielder independent pitching.The stat is related to ERA, but it attempts to neutralize the pitcher from his defense.Essentially how many runs is the pitcher truly responsible for.A pitcher with a bad defense will typically have a FIP that is less than his ERA, because some of his earned runs were attributed to the poor defense.A pitcher with a good defense will typically have a FIP that is higher than his ERA, because the defense helped the pitcher save some runs.Here is an example...the Twins team ERA in '17 was 5.09 but FIP was 4.57.An indicator that the Twins defense was pretty darn good (which it was).

 

SIERA is Skill-Interactive ERA. Essentially, it is an updated (supposedly) better version of FIP.It's very hard to explain (especially in text)...even fangraphs sends you to a different resource to attempt to explain it.So if you're interested in more details follow the link.

 

I do not believe your line graph makes a whole lot of sense to me.According to the graph, as the “ERA/FIP/SIERA” increases (vertical y axis), the QOP Pitcher Effect (horizontal x axis) decrease by -0.5?So, if I am looking for the median of 5 QOP, which is the 0.0 x axis, the pitcher has an average ERA of approximately 4.75, FIP of 4.5, and a SIERA of 4.25?This line graph is difficult to understand. 

The x-axis on the graph above is a standardized value.A standardized value of 0 means the pitcher through an "average" pitch according to QOP's measures, which would be a QOPV of 4.5.A standardized value of +1 means he threw a pitch that is 1 standard deviation above the mean, which would be a QOPV of 6.75 (the standard deviation is 2.75 according to the founder's findings).A standardized value of -1 means he threw a pitch that is 1 standard deviation below the mean, which would be a QOPV of 1.75.You're right, this does take a college level introductory statistics course to fully understand, so I have include the same line graph but with QOPV scores rather than standardized scores.Hopefully that should do a more clear job of showing that ERA/FIP/SIERA all decrease as QOP increases.

Gu5DVXp.png

 

 

 

 

1.Where does the twins starting pitchers fall into each bucket on the QOP scale? 
2.Where does the twins bullpen stand compared to a team with an elite bullpen (such as the Yankees or Dodgers)? 
3. How does Ervin Santana (Near Elite) compare to Clayton Kershaw (Elite)?
4. What was Fernando Rodney’s QOP in his successful saved games versus his blown saves in 2017? 
5. What about Matt Belize in 2017 (9 saves and 2 blown saves)?

6.Can this be translated to hitters as well?Such as, what was the QOP of the at bat that Joe Mauer hit a home run, versus the QOP of the AB's that Mauer slapped a single?

As I said in my post, this was simply an introductory/overview of the stat. I intend to use it in the future and other writers on Twins Daily have stated they may as well.At this point, the QOP website only tracks year by year.Doesn't do splits or game logs.Hopefully something that will become more developed over time.

 

Photo
Hosken Bombo Disco
Jan 12 2018 08:43 PM
I'm glad this conversation is still going. I plan to dive in to your piece this weekend.
    • Matthew Lenz likes this
Photo
Matthew Lenz
Jan 12 2018 09:48 PM

I'm glad this conversation is still going. I plan to dive in to your piece this weekend.

would love your feedback!
    • Hosken Bombo Disco likes this
Photo
Hosken Bombo Disco
Jan 16 2018 06:28 PM

would love your feedback!

It's great to discover new statistics and new ideas. Thank you! It will be interesting to see what this stat tells us. Santana threw what seemed to be a quality slider to Aaron Judge in the 1st inning of the wild card game, Judge was fooled by it, lunged and swung at it anyway, and the ball landed softly and safely in shallow center for a hit.
Photo
Hosken Bombo Disco
Jan 16 2018 07:12 PM
But that was just one pitch. SSS alert
Photo
Matthew Lenz
Jan 17 2018 11:03 AM

 

It's great to discover new statistics and new ideas. Thank you! It will be interesting to see what this stat tells us. Santana threw what seemed to be a quality slider to Aaron Judge in the 1st inning of the wild card game, Judge was fooled by it, lunged and swung at it anyway, and the ball landed softly and safely in shallow center for a hit.

Yea, I wish it could be broken down more.I remember that pitch...it should get a good QOPV since the formula doesn't take into account the result of the pitch.