K% & BB% for Pitchers
K% and BB% for pitchers have been added to the player pages and leaderboards in the “Advanced” sections.
They are both calculated using Total Batters Faced. SO/TBF and BB/TBF.
David Appelman is the creator of FanGraphs.
K% and BB% for pitchers have been added to the player pages and leaderboards in the “Advanced” sections.
They are both calculated using Total Batters Faced. SO/TBF and BB/TBF.
David Appelman is the creator of FanGraphs.
YEEESSSSSSS!!
Huzzah!
Awesome
And SIERA too??? Amazing!
Speaking of that, whatever happened to the SIERA article that showed up in the RSS feed put was pulled from the site?
Any chance player pages will ever allow calendar selection on splits (i.e. I want to compare May SwStr% to July SwStr%, etc.)?
HOORAY!
Now can we get K% for batters to have a PA denomination?
Cause otherwise year-to-year analysis, player-to-player analysis is apples-to-oranges based on BB%
This was done earlier in the week. http://www.fangraphs.com/blogs/index.php/k-change/
Thanks David.
This is great. My only other slight gripe is HR/FB over HR/OF because of the skill involved in PU%….but i’ll leave that fight for another day!
(and yes, I’m aware of your article from last year dismissing the need to use HR/OFFB and poo-pooing IFFB skills as substantial, BUT while it may be “good enough” overall, in fantasy, you want that “Extra 2%” wherever possible, and those players “at the margins” who dont fit are best served with HR/OFFB. Then again, if you popularize it here, I do lose one of the big reasons I win fantasy pitching stats and pitching trades every year….)
I thought IFFB/FB was a matter of luck.
Nope. Line Drives are pretty much luck/scorer bias, but popups are a mild skill, perhaps a real skill for just a few pitchers. It skews the data for extreme pitchers. For example, Volstad has a 3.67 xFIP, but 4.10 FIP if you use HR/OFFB
That’s not true. His xFIP with IFFB removed is 3.75 instead of 3.68.
You have to change the baseline. At least in 2011, the most any pitcher with more than 50 IP is different is .23 and that’s Tyler Clippard. He’s not going to sustain a 24% IFFB rate either. He’s the outlier and other than that you get about a .15 change max.
The biggest changes are always going to be from pitchers with unsustainable IFFB rates. I don’t know why we’d want to factor that into an ERA estimator when we’re trying to weed things like that out.
The articles he’s referring to are:
http://www.fangraphs.com/blogs/index.php/are-popups-a-skill/
http://www.fangraphs.com/blogs/index.php/infield-fly-balls-and-xfip/
I’m not saying I can’t be convinced to switch it, but I need more convincing on this one. The data I’ve seen actually points to leaving IFFB out of it.
Dave,
That assumes, however, that the scale for xFIP with OFFB (let’s term it exFIP) is on the same scale. I also include LD-normalized HR/OFFB, so sorry, you are correct as to the new present value using just HR/OFFB over HR/FB.
My opinion of the matter is this — LD% is the non-skill, not the IFFB% rate. What I do personally, and I’ve got some people testing the validity of this at the moment, is normalize batted ball data in a single season based on either a 19% LD rate or a pitcher’s career rate. This gives me xGB, xOFFB, xIFFB, xLD. From there, I calculate the yearly or 3-yr (depending on what I;m doing) lg HR/OFFB, multiply that by xOFFB. Then I also multiply .008*xLD. Those collectively form “xHR”. I then just do the simple FIP formula, though I have a formula for expected outs (xIP):
C+(3*BB-2*K+13*xHR)/xIP
I’m not advocating the use of that formula in place of xFIP because it does sacrifice a lot of simplicity, but just saying that I think that one of the problems with xFIP is that it ignores IFFB (arguable), but not LD%
*David. Sorry I always say “Dave” just because 2 of the 3 Dave/Davids go by Dave. My apologies
Again, not trying to be abrasive here. I think all the work here is great. Just trying to propose ideas to help make the model approximate best. As a IFFB can never be a HR, it just seems strange to include it in the xFIP model
Have you reviewed Dave’s article:
http://www.hardballtimes.com/main/article/never-swat-an-infield-fly/
Another reason I bring it up is because I’m 90% certain Dave’s original formula used HR/OFFB, not HR/FB. Not sure if there was mis-communication when it was brought to FG, but I’m just advocating bringing it back to as originally conceived
And I do, by the way, recognize the power of simplicity in these formulas. If you want to get complicated, you can better approximate values with base runs, etc. For example, Brian Cartwright has a formula for wERA that converts a pitcher’s events into an ERA, which is nice because a pitcher creates his own runs environment (as he points out, a walk by Pedro Martinez does not have the same expected effect as a walk by Daniel Cabrera). But still, if we’re going to do any normalization, just hoping to find a closer approximation.
I’m also trying to get Runs Created by Event data to work a line drive-normalized tRA formula using my xWHIP calculator. I wonder if doing that will have a valuable “future predicting” effect.
Obviously, we’re working with small samples here. With large data, all the metrics do a pretty similarly good job and predicting future value. Even ERA is not as “bunk” as once thought. But in small samples, part. 1 year, thats where xFIP-type stuff tends to shine most
Dave’s original formula definitely excluded IFFB, but for some reason I wasn’t clear on that when I brought it over here.
Calculating xFIP using “expected” HR data sounds interesting. Do you have an article on that somewhere?
At some point there’s going to have to be an entire section of ERA estimators!
Dumb question but what exactly is an IFFB? Is it a ball that is literally caught in the infield or can it be a weak pop up caught in the outfield grass? Or is it just considered any flyball that is caught by infielder? Thanks
On FanGraphs It’s tracked as any fly ball that is closer than 180ft from home plate.
David
Can you send me an email to the email address listed in this specific comment? We can discuss in more depth off-forum, but the original basis for expected home runs comes from calculated expected WHIP (xWHIP) formula, which has been validated. I have two stats people who are going to validate eFIP (xFIP using expected home runs) in the coming weeks, so I can send that your way when it comes out.
here are some basics on it:
http://www.hardballtimes.com/main/fantasy/article/xwhip-and-efip1/
If you are interested in including xWHIP on players pages, I’d be willing to share my formula with the site. I have a web developer working on an online interface for the xWHIP calculator so that people dont need to use excel to calculate xWHIP.
Also,
On the topic of HR/OFFB, can we do the same for hitters? A popup is NEVER a HR. It needs to either be HR/AO or HR/OFFB
Biggest risers and fallers, when going from ranking starters by K/9 to ranking them by K% (out of the top 25 or so):
Verlander jumps from 10th to 4th
Anibal Sanchez drops from 4th to 8th
Garza drops from 5th to 11th
Beckett soars from 39th to 20th
Obviously, on both sides you’re looking at some pretty fine gradations separating individuals starters. Still, interesting to consider.
Yup. Beckett was the first I noticed. Low BABIPs mean you’ll climb up the ladder with this switch.
SIERA (Y)
Love the additions. How did you get SIERA, though? I thought it was a proprietary stat of BP . . .
Sssssssssshhhhhhhh.
Clutch post by appleman!
Also, is this reflected in FIP now/will it be implemented over K/9 and BB/9?
Thank you!!!!!
Awesome!
Another pitching stat I’d like to see added is RA (or RA9, as Tango likes to call it), which many people prefer to ERA. But for something so basic, it’s surprisingly difficult to find, as it doesn’t seem to be on either this site or baseball-reference. Just a suggestion.
Anyway, keep up the good work!
And SIERA too! Momentous! No more BP custom reports every time I want to see those numbers.
About time.
I mean…awesome!
David,
Do you have an examples off hand of a player’s 2010 to 2011 K/AB going down, but K/PA increasing due to fluctuation in BB/PA inter-year
what exactly is siera
similar to xFIP measure — stands for Skill Independent Earned Run Average. It’s one of BP’s lesser “made up” (i say that facetiously, btw) proprietary stats
I was curious to see how this is affected Greinke, cuz Greinke is facing extra batters with all those hits he’s letting up. He would still lead anybody on that leaderboard with a K rate of 30.8%. His walk rate is still only 5%, which would be 13th. Here’s to hoping his ERA starts to reflect it.
Is there a way to make the site full screen width? The ‘advanced’ pitchers leaderboard is just incredibly tightly packed now.
I like the SIERA addition.
According to BP, the formula for SIERA is
6.145 – 16.986 * (SO / PA) + 11.434 * (BB / PA) – 1.858 * ((GB –
FB – PU) / PA) + 7.653 * ((SO / PA) ^ 2) +/– 6.664 * (((GB – FB –
PU) / PA) ^ 2) + 10.130 * (SO/PA) * ((GB – FB – PU) / PA) – 5.195
* (BB / PA) * ((GB – FB – PU) / PA), where +/– is a negative sign
when (GB – FB – PU) / PA is positive, and where +/– is a positive
sign when (GB – FB – PU) / PA is negative.
Is this the version Fangraphs is using? Does it include the park adjustments?
Luvs it
What about HR/OFFB (FB%-IFFB%)? This stat seems to be far more reliable from year to year (while HR/FB changes with IFFB rate)