Kris Medlen Pitches to Contact the Right Way
February 13, 2013 at 10:00 am by Mark Smith under Atlanta Braves
Kris Medlen was awesome in 2012.
Once he gained a foothold in the rotation on July 31st, there was no looking back for the diminutive righty. Over his 12 starts, he threw 83.2 innings (7 innings/start), had a 0.97 ERA (2.22 FIP), struck out a batter per inning (27% of hitters), walked 1 per 9 innings (3% of hitters), and had a 1.95 GB/FB rate. If the goal of a pitcher is to strike a bunch of guys out, walk very few, and keep the ball in the yard (which it is, I think), Kris Medlen did an amazing job. Just to put a few things in perspective, the best FIP in the majors among qualified starters was 2.82 (Gio Gonzalez), the 10th best K/9 was exactly 9.00 (Clayton Kershaw), the best BB/9 was 1.19 (Cliff Lee), and the 12th best GB/FB was 1.93 (Adam Wainwright). With all of that, Medlen set a major-league record when the Braves had won in 23 consecutive Medlen starts (though 11 of those were from a previous season). In other words, Medlen was really, really, really good. But you knew that already. What we want to know more about is how he did it and, more importantly, if he can continue to do it.
Our first step is to look at something Medlen is known for – pitch location. Here are his pitch locations against league-averages.
Not surprisingly, Medlen’s pitch locations back up the consensus thoughts on him. He throws a higher frequency at the corners, especially the low corners, and the pattern is much tighter than that of the rest of the league. What this indicates is that Medlen has control (the ability to throw strikes) and command (the ability to locate specifically within the zone). Though it’s a little hard to see here, Medlen does a little better at also keeping the ball down, with a stronger frequency nearer the corners than the League. These certainly help explain the low walk rate and high ground ball rate, but what about the strikeouts?
One of things I’m curious about is pitch location with two strikes. My theory coming in was that Medlen’s command allowed him to pitch out of the zone more efficiently than other pitchers. Most pitchers, in my opinion, throw a lot of 2-strike pitches nowhere near the zone. This would get some strikeouts, but it would also lead to more balls (and thus more pitches and a deeper pitch count). My thought was that Medlen stayed nearer to the zone but still outside of it, inducing swings-and-misses, caught-lookings, or weak contact. Going back to the heat maps.
On 0-2, he definitely hugs one side of the plate (in to lefties and away from righties), but he stays in the zone or just outside. The League goes a little further out of the zone and hits more of the zone.
On 1-2, Medlen shifts away from the side of the plate and to the bottom of the zone, but again, he’s hugging the border of the zone with the League still going further out of the zone. This makes sense as the count favors the pitchers, and they don’t need to be in the zone. Back to the part of the zone aspect, I find it fascinating how Medlen pitches. On 0-2, he varies his vertical pitch locations, but on 1-2, he focuses on switching his horizontal locations. As the pitching philosophy goes, he pitches up and down, in and out.
Medlen is now all over the zone, but he still tries to stay away from the middle of the zone. The League is back within the zone (need to throw strikes now) but is all over the zone, including the middle.
For giggles, let’s look at 3-2. Medlen is definitely within the zone now, with the small sample size showing some red spots way out of the zone. The League is definitely within the zone now as they don’t want to walk anyone.
This confirms my hypothesis somewhat. Medlen does, in fact, do a better job of keeping the ball nearer to the zone with 2 strikes, but those were simply against the League. What about specific players in counts that strongly favor the pitcher (0-2 and 1-2)?
Scherzer, Kershaw, and Verlander are very good strikeout pitchers, but they tend to run up pitch counts occasionally. Again, going so far out of the zone could mean a few ridiculous swings, but it also ends in a lot of balls. Verlander, a man who throws a lot of complete games, keeps the ball closer to the zone in pitcher’s counts, but even he gets way out of the zone. Actually, Medlen’s best comp might be Cliff Lee.
This is what “pitching to contact” should mean. It shouldn’t mean just throwing the pitch over the plate. While throwing strikes is certainly part of it, it’s not a very nuanced understanding. The phrase should be “pitching so there’s a possibility of contact, though not hard contact”. All of these pitchers throw strikes in 2-strike counts, but Medlen shows the ability to throw the ball just out of the zone and closer to the edge of the zone with more precision and accuracy than other pitchers. Let it go and risk strike three. Swing and miss? Strike three. Swing and hit it, and you’ll hit it weakly. Medlen’s Chase% (swings on pitches out of the zone) was 32% (against League 29%) and was in the 81st percentile. Meds isn’t messing around, and it helped him get through innings faster (13.7 against the league average of ~16) while still striking batters out. Strikeouts are good things, and if you can get them quickly like Medlen, you’ve perfected the craft of pitching.
After all this, we still have one question left – can he sustain this. This is, unfortunately, not a question we can answer. What I can tell you is that Medlen has great mechanics that he can repeat, which helps his command. I can tell you he throws a two-seam and cut fastball that keep hitters guessing which way the fastball will break, which leads to swings-and-misses and weak contact. I can tell you that he throws an excellent change-up with sink and fade, and it is 9 mph slower than his fastball, which is very good and off-sets timing. And I can tell you that the pitching models we use (FIP, etc.) all think his performance was legitimately awesome last season, though not 0.91 ERA good because that’s just absurd for a starter. If Medlen continues to throw strikes and throw quality strikes as he has, the comps to Greg Maddux might be the first comps to Greg Maddux that are actually accurate.
Now we just have to hope that he’s healthy and that MLB hitters don’t find a way to adjust against him without Meds being able to adjust back.
Loading ...
Great article. I wish everything came with heat maps!
btw, ZIPS is out for the Braves today, an they seem to be really undervaluing the entire team. But, I guess they have always been the ying to the Bill James yang anyway.
I added up all of the WAR for the guys I think will make the 25 man roster, and it came out to be about 50 WAR. That’s really good. Considering most of those numbers look like the baseline, for the lineup at least, we should easily expect to be a 95+ win team. It just sucks that the Nats may be a 100 win team. Projecting J-Up at 3 WAR and Heyward at 4 COULD yield an extra 7 WAR between the two if Justin bounces back and Heyward follows his curve. Also, projecting Simmons as a 2 win player is extremely conservative, since he eclipsed that in 49 games in 2012.
I didn’t add em up. But the numbers (both counting and WAR) for the OF, Simmons and Freeman looked a little on the low side to me.
Mark, that is awesome.
Great post Mark. Do you guys plan on doing something similar with the entire rotation?
I think between me and mark, we will probably get most of the rotation, except Teheran, who hasn’t pitched enough to really have useful heat maps. It takes a pretty good chunk of data for heat maps to work properly.
Now that I know how the heat map thingy works, it’ll take me a lot less time than it did for Medlen.
I was literally going to look up these heat maps this morning before I checked CAC, and boom, they were already in front of me. Great work, and expert mind reading.
Another Maddux comp that I see in Medlen is his use of the 2-seam. He throws it in to LHB and backs them off the plate. They are either jammed if they try to pull it, or can’t hit it if they wait for the break. He throws it to the same side of the plate against righties, and it looks out of the strikezone all the way from the right side. They don’t really stand a chance. It also helps that his arm action doesn’t skip a beat on his change, and the similarity in movement between the change and 2-seam is almost unfair, given the velocity difference.
Another interesting facet of his game that I think helps him a lot is his pace. Since they started tracking the stat, he is about 3 seconds faster between pitches than league average, meaning the hitters’ calculations will be much less in-depth than the ones CAC has provided here.
As 2012 wore on, I could have sworn his ability to produce weak contact was a fluke, but as the sample grows and grows, he really is starting to look like Maddux.
Very good comment and I completely agree. Medlen throws that 2 seamer the way he does and it looks like a flashback to the old Maddux days. Also both Medlen and Maddux are similar body types too. Both guys being on the smaller side of league average for a pitcher which makes people question if they could handle the workload of a starting pitcher.
Well Maddux was one of the most durable pitchers ever. Maddux once joked that he thought one of the reasons he was as durable as he was during his career was because he had a little fat on his mid section which kept him from injuring muscles that a lot of players injure.
This is a great article…I just wanted to make one critique. I found comparing Meds to other know and successful pitchers much more useful than to compare him to the rest of the league. The entire league is an enormous data set of hundreds of thousands of pitches. You get all the good with all the bad so it is not surprising that the heat maps basically look symmetrical mainly centered around the middle of the plate.
I think you could have written this with Meds heat maps standing on their own as the league maps didn’t add much. I did really enjoy the second part where you compared Meds to pitchers that we know are successful because the heat maps tell us a lot more.
This is great stuff. It would also be kinda cool to see Tommy Hanson’s heat maps as it seemed to me that he had very little command of his fastball last year. I know a lot of his problem was reduced velocity (mechanics?) and lack of movement on the pitch, but even at 89mph, if Tommy could have established his fastball early in games his breaking stuff would have been so much more effective.
This is awesome. I would love to see the comparison of Beachy 2 years ago (high strike out/shorter outings) to last season (less strikeout/longer outings) to see how he may have changed his approach. I believe he was doing what Medlen is doing and that is nibbling around the corners all day long. Thanks for the article.
Great write up. But when comparing Meds to three other pitchers you have Meds heat charts for everyone.
That’s because they’re comparing Medlen to 3 different pitchers using the same data. It helps to see all 3 side by side rather than having to keep looking back at Medlen’s chart.
Eh, he just goofed with his phone.
Yeah I got all impatient and jumped the gun. Apologies.
Sorry. Turned my phone sideways and it showed the other maps.
So I have a question about how people discuss these heat maps. The “league average” is all of the pitches, correct? So it’s the aggregate? I guess I’m having a hard time figuring out exactly what comparing someone to the “league average” really means. It’s not what the “average pitcher” (which doesn’t really exist, does it?) throws. It’s just here are all of the pitches, and here is player X’s portion of those, right?
That’s the reason I started pulling specific examples. Either way, the heat map shows the frequency of pitches in a certain spot, so pitchers have throw a lot in those spots to show up as red as it does.
I guess it’s the phrases like “the pattern is much tighter than that of the rest of the league” that make me wonder what that is supposed to mean. The rest of the pitchers in the league aren’t necessarily throwing to wider targets individually. Medlen could have a spread that is wider than most other pitchers. Heck, he could be the most “all over the place” pitcher in the league. I’m not saying it is, only that this data doesn’t tell us anything about that.
Similar is the language about Medlen vs the rest of the league with respect to hitting more or less of the zone. Other pitchers aren’t necessarily pitching to a larger portion of the strike zone than Medlen on an individual basis.
I like the info provided and think it’s interesting to look at a pitcher’s approach in different situations. These heat maps do provide a great visual for that data and discussion. I appreciate the work you’ve put into this. I do, however, think it’s important to make some of these distinctions to be more precise with what the data means and the conclusions one can draw from it.
you’re right that this doesn’t represent an individual, but I think you are missing the usefulness of knowing what the overall league looks like.
I think what you’re saying is that perhaps certain pitchers pitch more to certain areas of the plate, and when you aggregate them it looks like they’re more all over the place than they really are. I think that’s true to some extent, but knowing where all the pitches in the league goes is still useful. Because it shows you were Medlen falls in the league spectrum. That’s what heat maps are best used for, to show you more the “how” than the “how good”. pure statistics are fine for showing us the “how good” aspect, we can get a very good gauge of medlen’s command by walk rates and strikeout rates. What the heat maps are more used for is to say “this is how he attacks hitters, and as you can see, this is more to this area than most pitchers.” You are right in that you should be careful when looking at just a heat map to say “this shows Medlen has better command than average” because we don’t know where those pitchers were trying to throw those pitches. But we can still see that Medlen lives more on the outside corner than most pitchers do. You can see where in the spectrum it would fall.
The biggest problem with pitcher heat maps is that pitchers tend to pitch lefties and righties differently. In some sense the most accurate point of comparison is going to be how right handers pitch, and then further break it down against LHB and RHB. But using all RHP is still a useful tool to get a point of comparison, so you can know if medlen pitches further outside, more precisely, or whatever, than the average. However, that then creates the issue of doubling the comparison heat maps you need, which may or may not be necessary for a given point you’re trying to make.
Thanks for the reply, Franklin. I’m still not sure I believe that some of these statements can be made based on this data. Hopefully I learn something here.
“But we can still see that Medlen lives more on the outside corner than most pitchers do.”
Can you say that based on the data in the heat maps, though? I am not seeing how the aggregate of pitches tells us what most pitchers do.
Say we make a heat map of some dart throwers. Four of them. A goes first. Throws 3 of her darts evenly spaced from 12 to 3 o’clock and 1 dead bullseye. B throws 3 from 3 to 6 and one bullseye. C: 3 from 6 to nine and one bullseye. D: 3 from nine to twelve and one bullseye.
Now let’s compare A’s heatmap to the aggregate. A doesn’t throw less bullseyes than anyone else, or the average, or than most. But the aggregate heat map is going to be hotter at the bullseye and A’s will be evenly dispersed from the bullseye up to her quadrant.
The aggregate heat map will just be hottest in the shared areas.
What I look at when checking these comparisons out is “are his hottest spots also a part of the hottest section for the aggregate?” If the answer is yes, then he is an unsurprising chunk of the aggregate.
Comparing the shape of the aggregate heat map for the league to an individual pitcher seems to me to tell us no information. Comparing the hottest spots to the hottest spots seems like the only thing that gives any info that can be used as a comparison.
Thanks for the discussion.
In my darts example… You could reasonably say that A uses the upper right quadrant at a higher rate than the aggregate. A’s heat map would be darker in that area than the aggregate.
But what I see with these heat maps is that Medlen’s heat maps are a normal chunk of the aggregate. His hottest spots are part of the hottest spots of the aggregate.
@bobby
Yes, the bullseye would show up as the hottest spot for the aggregate, however, the bullseye would NOT show up hotter for the aggregate at the bullseye than it would be for any of the individuals. Looking at the aggregate heat map and looking at the individual, you’d see both at the ‘heat level’ of 1/4 around the bullseye. Comparing the aggregate map to the individual thrower would simply show that he/she hit the bullseye at the group average rate, which is perfectly valid.
The thing to remember is that the colors on the heat map correspond to rates, not relative frequencies. So if it’s a given color, that means he hit that spot a given percentage of the time. I think that allays the concern you laid out above.
@Franklin
Thanks. My application of the rate vs amount was definitely off there. I’m still at a place where I think it’s a misnomer to talk about the “average pitcher” or “most pitchers” because it’s really the pitches and we don’t know how the pitches were split up. I think to do so makes assumptions about the distribution of the pitches over the pitchers.
Say one of the dart throwers, D, throws all four in the bullseye, leaving the 9-12 quad empty. Now you’ve got 7/16 in the bullseye. Now compare A’s heatmap to the aggregate. To say that A throws less bullseyes than the average or that she throws less bullseyes than most other dart throwers would be inaccurate. To say she did not throw “her share of the total” or the “league average rate” number of bullseyes is correct. League average rate and average individual/most individual language is what I have an issue with I think.
Thanks again!
What would really be awesome is to pull heat maps comparing Med to the pitcher we all think he reminds us of…Maddox. Would this be possible?
Pitch f/x data doesn’t really go back to the era that would be best to compare to Maddux. It goes back to 2002 in some doses, but it was remodeled in 2007 to be better.
Completely off-topic question: Do you guys think the Braves should bring back Tim Hudson for another year or two after this year? I was reading online where Hudson said he wanted to pitch a couple more years and finish his career with the Braves. If he has another decent year this year like last year, do you think the Braves should bring him back? I realize Hudson’s old (he’ll be 38 in July), but he seems to be aging okay. The fan in me wants to see Hudson finish his career with the Braves, so I’m glad I’m not in Frank Wren’s position here.
While I would not mind it there just really will not be room for him. There maybe next year but probably not after that. The rotation would basically be Medlen, Beachy, Minor, Teheran, and then perhaps either Gilmartin or Graham next year. The year after that though Gilmartin and Graham will definitely be ready so I just do not see it.
In regard to Huddy, it’s something that can be decided later. Let’s see what we have, and if the situation dictates needing to keep a pitcher, then we can look at Huddy’s performance to see if it’s worth keeping him.
And if I had my druthers, it may be better to keep Maholm.
They are totally unrelated, but I love how two pairs of best friends came up in the system, Francoeur / McCann and Hanson / Medlen. Frenchy and Hanson were the hyped stars and top prospects in the Braves organization, each supposed to be the new face of the franchise for years to come. Now they’re gone and both might not have a career in their 30s, barely holding onto their jobs while performing at the bottom of the barrel. We’re left with their sidekicks–one dopey chubby catcher and a 5’10″ that looks like a 12 year old–who are both about the best at their respective positions. Funny how that happened.
Just another example of how physical tools can be highly overvalued over mental tools.
I would not say that they(Hanson and Francoeur) were not mentally strong. Hanson’s shoulder just seems like it is decaying faster than many anticipated due to his awkward throwing motion. Francoeur seems to stuggle making adjustments year to year.
I strongly disagree about Hanson not being mentally strong. He just for the life of him could not keep the big innings from happening because he wouldn’t have the mental makeup to overcome something bad happening to him. He reminds me so much of a young Carlos Zambrano in that you can get to him early and rattle him and knock him out of the game. It’s why he has so many starts in his career that are 5 innings or less.
Look at his pitch count in those games too. He tries to be too fine and he gets crushed for it. How do you explain his game July 25th against the Marlins when he walked 7 along with the 7 SB allowed
To be fair, Hanson’s physical tools degraded because of injury, not he didn’t perform up to his physical tools. Frenchy didn’t live up to the physical tools he still has.
Also, Medlen has pretty good physical tools as well. The coordination to repeat the delivery as he does is quite impressive.
And the being too fine as you call it and running up his pitch count is a lack of a good mental makeup IMO. With the kind of stuff he had at the beginning of his career he should have throwing 7 and 8 innings every time out but he could barely go 5 innings because of pitch count. I can count how many times he would get 0-2 or 1-2 and then end up 3-2 with about 5 foul balls and before you know it he’s thrown 10 pitches to a single batter and it pisses you off.
That is more the lack of an out pitch after his velocity had decreased. The low 80′s change is not effective when your fastball sits high 80′s. He couldn’t command the curve well enough for hitters to swing at it with 2 strikes. His ‘refined’ mechanics severely subtracted from his stuff, which was really all he ever had.
Pitch f/x shows that he threw a good bit of two seamers in 2011. However, he has never registered more than 14 according to Pitch f/x in any other season. Considering how straight his change is, and how 12-6ish his curve is, you’d think keeping around a pitch with armside break would help. His stuff has just become so underwhelming.
Careful there…pitch f/x classifications aren’t perfectly accurate, so it’s more likely the same pitch was classified differently than Hanson randomly throwing a bunch of two seamers one year of his career.
That’s why I said at the beginning of his career when his fastball was in the mid 90s. He still struggled to throw 5 or 6 innings many times out because of guys able to foul pitches off at will despite him having a very good curveball.
I would be interested to see Hanson’s heatmap of 0-2 and 1-2 counts. I felt like in those situations with Tommy last year he struggled to really pound the zone and was just looking to beef up his K numbers.
I think if you looked at his heat maps, you would see a lot of blue in the zone and red everywhere outside the zone. It seemed like every time Tommy got ahead of a hitter, he would always get worked to 2-2 or 3-2.
What do you guys think about Bourn getting a 4 year $48 mil deal. I know the asking price was much higher when the Braves signed B.J Upton, but would you rather have Upton at 5 years $75 or Bourn for 4 years $48?
Is medlens small size still a cause for concern withi his ability to start and pitch through out the season? Are there any other starters with a similar stature?
Maddux was 6’0, 170 while Kris is 5’10 190. That’s close. I’ll trust him, he lowers his pitch count well because he’s so efficient. I think he’ll be ok.
I’m not convinced small stature really means that a guy can’t get through a season. Seems like a self-fulfilled prophecy – no one lets them try so of course none of them make it. Not sure there’s a lot of evidence behind it.
The real issue with small pitchers is not getting downward plane with the fastball, but because Medlen throws a two-seamer, it negates the issue.
Pedro Martinez was 5’11″ and 170 lbs. He turned out to handle starting from year to year pretty well.
I don’t think there’s much to the size-durability connection. Medlen’s delivery is so effortless, and knowing the athlete he is, I think he’ll be just fine. Little guys with a delivery and velocity like Lincecum truly blow my mind though. Plus, as the article outlines, Medlen throws fewer pitches per inning than the league average. So he may throw a lot of innings, but the true count of pitches should be lower than most as he continues to stay in and around the zone.
Great article once again. Given the quality of all the articles I would pay to read this stuff!