Edge%: Another Fantasy Tool
Bill Petti and I have been looking at pitches thrown at the edge of the strike zone over the last few days. Pitchers who throw a higher percentage of pitches on the edge perform better than those pitchers who throw a lower percentage. Edge% is not going to change the core traits of pitcher talent, high strike outs and low walks, but it is a nice tool to help differentiate between two similar pitchers.
Let’s examine at two pitchers, CC Sabathia and Dan Haren. My gut reaction when seeing the pair is to go with CC especially after Haren struggled through 2013. By looking at their valuations, Edge% makes the choice a little harder.
When ranking pitchers, I use a formula I created last year which only uses strikeouts and walks to get a pitcher’s valuation. The pair have almost identical projections when using CAIRO projections. Looking at their numbers it is easy to understand why.
Name: K%, BB%
Sabathia: 21.3%, 6.2%
Haren: 20.9%, 4.7%
The extra strikeouts Sabathia gets is offset by his walkss. In a draft or auction, I historically keep a few other numbers besides my Fantasy Valuation. The first extra stat is age. Try to always go with the younger pitcher. This pair are both going to be 33, so no advantage to either one. The second is K%. A pitcher is more likely to improve their walk rate than strikeout rate. Take the higher K% pitcher when given the choice. The pair is close, but Sabathia get a small edge.
Another factor is GB%. Sabathia gets a nice boost with an average GB% of ~47% over the past two seasons while Haren is closer to ~41%. The final stat I used in previous seasons was health. Both pitchers come with risks. They are both are past 30 and spent time on the DL last season which will be shown in their 2013 DL chances (2012 link). Injury concerns also exist because both pitchers saw in a drop in their velocity from the season before.
Average Fastball Velocity (mph)
Name: 2011, 2012
Sabathia: 93.8, 92.3
Haren: 90, 88.5
Looking at just the previous values, I would take Sabathia. With Edge%, I have a new value. The value correlates decently from season to season, so it is predictive. Here are the pairs values from the past 3 seasons
Edge%
Name: 2012, 2011, 2010
Sabathia: 15%, 16.2%, 17.7%
Haren: 17.4%, 16.8%, 14.6%
Haren has been throwing on the edge more in recent season while Sabathia is going there less and less. Finally a mark to put up in Haren’s favor. In the end, the advantage in Haren’s Edge% is not enough to overcome Sabathia’s ground ball rate.
Edge% is another tool to help differentiate pitchers with similar rankings. With Sabathia and Haren, it was not enough advantage for Haren to push him over Sabathia, but it put the pair’s value closer together.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
While I do like this edge% idea, I’m gonna have to go with my gut if I had to choose between the two. And I think you know who I mean when I say ‘gut’.
Did you know that your gut has more nerve endings than your brain?
By going with your gut, I can only assume you mean Haren. Your gut will be much happier with some food in it.
Ha! Ha! Must be a beautiful thing to be a multi-millionaire, pitch at an elite level and be able to chow down on boxes of Cap’n Crunch with reckless abandon.
A stat that correlates from season to season is always worth looking at, but I think you need to define the real-life correlates a little better. What is the argument that says a higher Edge% is better? Compared to 2011, Sabathia had a higher K% with a lower Edge% last year, and Haren had a lower K% with a higher Edge%. Which would you rather have?
It could be that some pitchers always have a higher Edge% (maybe it’s the same pitchers that always outpitch their FIP, like Hellickson and Glavine). Or Edge% means a pitcher is forced to ‘nibble’ more, and dominant pitchers can live without the hitting the edge. Or any of a thousand other possibilities that I’m not smart enough to think of.
Whatever it is, it’s a really cool stat, and I hope it sparks some great ideas.
Very interesting.
I watched about 15 of Dan Haren’s starts last year and I can confirm that he was definitely pitching on the edge of the strike zone all year. However, I felt that most of those pitches were called balls. Fortunately, his BB/9 remained under 2 which is good, but all of those called balls on the edge of the strike zone resulted in high pitch counts and forced Haren to throw very hittable pitches in the strike zone late in hitter counts which the hitter would crush (see HR/9 of 1.43). I know this is anecdotal, but I have a strong feeling the numbers would support the theory in which case I think Edge%, at least for the 2012 season for Haren, would not be a positive indicator of how a pitcher performed well. However, I could be completely wrong (read: Jeff Zimmerman is a hell of a lot smarter than I).
Regardless, I am hoping for a rebound for Haren in 2013. Last year was my first year owning him in my long-term keeper league after coveting him for a few years and I was very disappointed to say the least.
I wonder if this is an easily evaluated claim. I have no idea how, but I hope somebody does.
I decided to take a look at this myself, the data I put together from Fangraphs* and BaseballHeatMaps* can be found here: https://docs.google.com/spreadsheet/ccc?key=0AgcNiR1EMwPhdGFJeHRNcGJqWTg2VjZrWGJhU1ZyZ0E. All I did was compile the data and perform the eyeball test. Someone else with the time and ability may be able to actually complete some sort of statistical analysis with the data.
Just looking at the data, it appears that Edge% for Haren does not have a direct effect on what I saw during the season. In 2012, he did have more pitches called for Balls (and subsequently less for Strikes) than he did in the past with his IP/GS dropping to below 6 for the first time in the past five years also. However, his Edge% has been relatively consistent between 2008 and 2012 (read: no wild swings) and it does not seem to correlate with BB/9, K/9, IP/GS, or HR/9 via the eyeball test. Those numbers seem to have bottomed out in 2012, but I don’t think one can definitively concluded that his increase in Edge% lead to more called Balls which led to his poor 2012.
*The number of pitches thrown differed between the two sources. Therefore, I took the Edge%, Heart%, and Outside% values from BaseballHeatMaps with everything else taken from FanGraphs.
(For whatever reason, my comments don’t post from my work PC, so I have to add this line so I don’t get a duplicate comment error.)
A cool stat, but how much of an impact does an Edge % difference of 2.4% (CC vs Haren) have? Over the course of a 100 pitch game that equals… about 2.4 more “edge hits”.
How does one hit the edge of the zone more, but have a lower strikeout rate? Could it could be that Haren was the victim of a stingier collection of umpires and more of his edge hits were miscalled as balls while CC got the benefit of the doubt? But then why is Haren’s bb% lower? Maybe Haren faced tougher hitters? Maybe CC had the better pitch sequence? Maybe Haren’s dip in velocity made him more hittable? So many variables but I’m left still thinking that this bodes well for a correction in Dan Haren’s favor.
I would be more inclined to see Haren’s rising edge% as tightly and inversely related to his declining velocity… and thus a harbinger of bad things, not good… and push him down a few pegs.
Haren: 2012, 2011, 2010
FA velo: 88.5, 89.8, 90.5
Edge%: 17.4%, 16.8%, 14.6%
meanwhile, CC is still humming that fastball (though granted, he was down a tick last year)
Sabathia: 2012, 2011, 2010
FA velo: 92.4, 93.9, 93.6
I’m taking CC. Hands down.
I decided to take a look at this myself, the data I put together from Fangraphs* and BaseballHeatMaps* can be found here: https://docs.google.com/spreadsheet/ccc?key=0AgcNiR1EMwPhdGFJeHRNcGJqWTg2VjZrWGJhU1ZyZ0E. All I did was compile the data and perform the eyeball test. Someone else with the time and ability may be able to actually complete some sort of statistical analysis with the data.
Just looking at the data, it appears that Edge% for Haren does not have a direct effect on what I saw during the season. In 2012, he did have more pitches called for Balls (and subsequently less for Strikes) than he did in the past with his IP/GS dropping to below 6 for the first time in the past five years also. However, his Edge% has been relatively consistent between 2008 and 2012 (read: no wild swings) and it does not seem to correlate with BB/9, K/9, IP/GS, or HR/9 via the eyeball test. Those numbers seem to have bottomed out in 2012, but I don’t think one can definitively concluded that his increase in Edge% lead to more called Balls which led to his poor 2012.
*The number of pitches thrown differed between the two sources. Therefore, I took the Edge%, Heart%, and Outside% values from BaseballHeatMaps with everything else taken from FanGraphs.
(For whatever reason, my comments don’t post from my work PC, so I have to add this line so I don’t get a duplicate comment error.)