BABIP Driven Starting Pitcher K/9 Changers
We all know by now how to best utilize a pitcher’s BABIP data during and after the season. For the most part, a pitcher with a BABIP significantly below the league average is probably going to see that metric rise, while a pitcher with a mark above the league average will likely post a better one moving forward. But that’s not the only use of BABIP. It also affects a pitcher’s strikeout and walk per nine innings rates. This is why many now prefer to use strikeout and walk percentage.
When you dig into each of the metrics, it is easy to see why. The denominator of the strikeouts per nine innings rate is innings pitched, while strikeout percentage uses total batters faced. That means that if a pitcher gives up lots of hits in an inning, he will have additional opportunities to strike a batter out, and therefore increase his K/9 rate. He may even up striking out the side, but maybe it took him 12 batters to do so. That would give him a K/9 of a perfect 27.0. Compare that to the pitcher who faces just three batters and strikes out two of them, while getting the third to ground out. His K/9 is a less impressive 18.0 and based on this example, would appear to be a worse strikeout pitcher. However, the first pitcher’s strikeout percentage is 25%, while the second’s is 67%.
So another use of BABIP, aside from perhaps determining who has benefited from good fortune and whose luck has been poor, is to identify which pitchers have had greater or fewer opportunities to strike batters out. Since some fantasy leagues use K/9 and most others use straight strikeouts, this type of analysis is useful as it could uncover who should see an uptick or downturn in strikeout rate over the rest of the season.
First, I looked at all starting pitchers with at least 50 innings pitched in a season from 2003-2012 (n = 1,770) to formulate a regression equation to turn a pitcher’s strikeout percentage into an expected K/9 (xK/9). That equation is:
xK/9 = 0.5022 + (K% * 35.5724)
R-squared = 0.976
I then calculated the xK/9 for every starting pitcher who has thrown at least 20 innings this year and sorted by the difference between K/9 and xK/9. The group of pitchers with a K/9 greater than their xK/9 had an unweighted BABIP of .323, which is exactly what was expected. The group with a K/9 lower than their xK/9 had an unweighted BABIP of .275. Boom. And for fun, the small group (just 18 pitchers) with a K/9 the same as their xK/9 had a .296 BABIP. What’s the league average starting pitcher BABIP? .295. So this population is a perfect illustration of the concept that BABIP has a great effect on a pitcher’s K/9. Furthermore, the correlation between BABIP and the difference between K/9 and xK/9 from this population was .66, which is rather significant.
Of course, the K/9 surgers won’t necessarily experience an increase in fantasy value, nor will the decliners suffer from a reduction in value. The surgers might surge because their BABIPs jump toward league average, which means more hits, hurting their ERAs and WHIPs. So the increased strikeouts will be offset by worse ratios. The opposite would be true of the K/9 decliners, as the concept works in the other direction. These pitchers should actually experience better BABIP luck, so the improved ratios should offset a drop in strikeout rate.
Below are two tabs of a spreadsheet that are separated based on whether the pitcher’s xK/9 is below or above his actual K/9. The first sheet is the potential K/9 decliners and the second one the surgers.
Mike Podhorzer is the founder of ProjectingX IQ, an advanced fantasy baseball analytics platform that transforms projection data and in-season performance signals into actionable intelligence. He is the 2015 Fantasy Sports Writers Association Baseball Writer of the Year and three-time Tout Wars champion. He is the author of the eBook Projecting X 2.0: How to Forecast Baseball Player Performance, which teaches you how to project players yourself. Follow Mike on X@MikePodhorzer and contact him via email.
Another reason to love Iwakuma?
Also, even Cashner’s xK/9 is disappointing. What’s up with him this season?
Well, you also have to taken into account my recent xK% equation! Lots of factors at play. So if Iwakuma’s K% is real, then his K/9 should be higher. However, I don’t think the K% is real in the first place.
Cashner’s slider is bring thrown with less velocity and its SwStk% is way down. Still holding out hope his strikeouts return, but given that there’s an explanation behind the downturn, I’m not very confident.
Also, the same logic should apply to Iwakuma’s (and other low-BABIP pitchers’) BB/9s and HR/9s. As BABIP regresses upward, all 3 of the per 9 metrics should increase. I guess the important take away is that an increased BABIP should always hurt a pitcher’s performance on the aggregate, but that the increased K/9 is at least a silver lining.
Nice work. Good to see the impact of BABIP on % vs. per 9 stats quantified.
Couldn’t we get an even better xK/9 equation by factoring in BB% and HR%? A low BB% or a low HR% would lead to an decreased K/9 relative to a pitcher’s K% just as a low BABIP does. Whether it’s a low BABIP, BB%, or HR%, fewer hitters are reaching base safely in between the strikeouts. Actually, you ought to be able to almost perfectly calculate all the per inning metrics using BABIP and the per plate appearance metrics. I think the only events that would be left out then would be things like double plays, caught stealings, and errors.
This is true, but the assumption is that BB% is a skill that should remain fairly constant. We know BABIP isn’t, especially in tiny sample sizes, so there’s a good chance a pitcher’s BABIP will move up or down over the rest of the season. This is not the case for BB%, so K/9 wouldn’t be affected.
Yeah, I think I explained that poorly. What I meant is that you could have an equation for xK/9 as a function only of K%, BB%, and HR%. If you assumed that the % stats would stay constant, then the only other factor would be BABIP. I would expect, for example, that Iwakuma’s xK/9 using this equation would be somewhere in between the one you calculated and his actual K/9. If he keeps his BB% this low, then his K/9 will always be somewhat depressed relative to his K%, even after BABIP regression.
Cashner is primarily a 2 pitch starter, his breaking ball is his worst pitch and catches the zone too much. Doesn’t air out his fastball as often as he’s a starter now. I’ve been impressed by his control improvements but again, needs to pitch effectively outside the zone.
Mike, when you do these…what do you call them? scrollboxes? can you freeze the headings so we can read them when scrolling to the bottom? Or if they can’t be frozen, put them in the title under “BABIP Driven K/9 Changers” (or in place of it)?
They are spreadsheets! I just copied and pasted from Excel into Google Docs. I tried to freeze the top row, but it doesn’t actually remain frozen when embedded into the post. If I check in on Google Docs itself, the row will appear frozen, not sure why the functionality no longer works when embedded.
any way we could get some analysis on homer bailey’s strange home/road splits this year?
Small sample size – meaningless. If you check xFIP, he hasn’t been that much worse away. It’s just a high BABIP and low LOB% hurting him, while he has yet to allow a homer at home, which won’t continue.
“He may even up striking out the side, but maybe it took him 12 batters to do so. That would give him a K/9 of a perfect 27.0.”
Or as it is more commonly referred, The Manny Parra.
Oh man, what memories…Parra was one of my favorite sleepers back in the day!