ABA: The New WHIP
March 19th, 2012 | by Ray Flowers |
When is one not one? No, I’m not operating in some alternative universe outside of the Matrix with Morpheus trying to wake me up to the truth (if you don’t get that pop culture reference you need to start seeing some movies), I’m simply asking when is one not necessarily equivalent to one? If you are interested in riddles, or just want to know what the hell I’m talking about, please read on.
AVERAGE BASES ALLOWED
Average Bases Allowed, henceforth ABA, is an innovative way to look at pitcher’s effectiveness and is designed to replace WHIP (Walks + Hits / IP), though I would settle for it to be used alongside WHIP until it catches on (I’m so amenable aren’t I?). What spawned the idea of ABA? Consider the following simple comparison.
Pitcher A allows one hit and one walk in two innings. Therefore his WHIP is 1.00 (two base runners in two innings).
Pitcher B allows one hit and one walk in two innings. Therefore his WHIP is 1.00 (two base runners in two innings).
So, according to WHIP, both pitchers have performed the same. However, does that mean that they were equally effective? What if we added a bit more depth to our example?
Pitcher A: Allowed a walk an a single in his two innings.
Pitcher B: Allowed a walk an a home run in his two innings.
Therefore…
It is reasonable to posit that Pitcher A had an ERA of zero. After all he gave up only two bases in his two innings. However, Pitcher B’s ERA was at least 4.50. Why? If Pitcher B walked a guy and then gave up a home run to the next batter he would have allowed two runs in two innings – hence his ERA would be 9.00, an even if it was a solo shot it would have still plated a run leading to a 4.50 ERA. So as you can plainly see, while the hurlers may have the same WHIP, the actual result of their performances in the real world would have been drastically different. Because of this simple yet often overlooked fact, I went about trying to set up a way in which I could analyze a pitcher’s performances in a more equitable way. Instead of using hits and walks as does WHIP, I decided to use total bases allowed and walks (because WHIP leaves out things like hit by pitch, I made the decision to do the same with ABA). Why replace hits with total bases?
Is it more important to know how many batters are allowed to reach base or is it more important to know how many bases they received when they reached base?
Here is the formula for ABA.
ABA = (TBA + BB) / IP
Does it not stand to reason that the pitcher who allows fewer bases to those batters who do reach base would have a better chance of limiting the amount of runs that score? Let’s take a look at a concrete example to illustrate.
In 2011 Tim Lincecum and Colby Lewis had identical WHIP’s of 1.21. Does this fact mean that they were equally effective hurlers in 2011 at limiting hitters ability to produce bases and runs? Let’s use ABA to investigate to see if we can form a more nuanced opinion between the two hurlers who had the the same WHIP last season.
Lincecum: 111 singles, 48 doubles, two triples,15 homers, 86 BBs in 217 IP
Lewis: 112 singles, 35 doubles, five triples, 35 home runs and 56 BBs in 200.1 IP
Remember, according to WHIP both pitchers were equal with a 1.21 mark. This is not the case according to ABA.
Lincecum: 273 total bases + 86 BB in 217 IP = 1.65 ABA
Lewis: 337 total bases + 56 BB in 200.1 IP = 1.96 ABA
As you can see, if you were only looking at each pitchers WHIP columns last season, you might miss the fact that Lincecum did a much better job at limiting baseball runners last season (this is also reflected in the ERA – 2.74 for Lincecum and 3.38 for Lewis). Thanks to ABA we can state that, despite equal WHIP marks, Lincecum was easily the more effective pitcher last season. All told there were five pitchers who threw at least 160-innings and posted a WHIP of 1.21. Here are the ABA mark for all five.
1.61 – Madison Bumgarner
1.65 – Tim Lincecum
1.80 – Jeff Karstens
1.83 – Hiroki Kuroda
1.96 – Colby Lewis
As you can tell, WHIP really doesn’t tell the whole story. ABA may not either, but it certainly is a much more accurate gauge of how a pitcher has performed.
Speaking of that, how in the heck to read ABA? Glad you asked. The lower ones ABA the better, but it doesn’t read the same was as WHIP. Whereas the average WHIP last season was 1.32, the league average ABA of all pitchers in 2011 was 1.86.
Here is a rough key you can employ for ABA.
Below 1.50: elite level performance
1.50-1.70: All-Star level
1.71-1.89: Solid major leaguer worthy of counting on in fantasy
1.91-2.10: Barley holding on to an role as a fantasy starter.
2.11 and up: Might as well line up a pitching machine
So there is my brief explanation of Average Bases Allowed, or ABA. Now that you know what it is, you’ll have to read PART II where I will take a look at the hurlers who threw at least 40-innings in 2011.
To sign up for your baseball league this year make sure you check out Fleaflicker.
By Ray Flowers
Tags: ABA, Average Bases Allowed, Colby Lewis, Hiroki Kuroda, Jeff Karstens, Madison Bumgarner, Tim Lincecum
By nokwurst on Mar 19, 2012
Excellent concept Ray, can’t wait to see who are the ABA studs & duds!
By Polka on Mar 19, 2012
I can name you 11 guys in my main league who hear ABA and think of Dr. J and the A-Train!
By mother on Mar 19, 2012
Good stuff, Ray. Where can one find the batted ball data? Thanks!
By Craig D on Mar 19, 2012
Ray,
With the breaking Soria news – Holland is the better arm than Broxton, but Broxton has the experience. I usually go with the talent, but who do you think the Royals with go with?
Given the opportunity does Broxton flame out? And, the job ultimately falling to Holland? How do you see it working out?
Thanks
By Matt on Mar 19, 2012
Ray,
Great idea. I was curious to see if anyone else has picked up on this yet and it looks like a guy named Alfred P. Berry came up with an ABA stat in 1951. They equate it to Opponents’ Slugging Percentage as well. While not the exact same thing, it probably correlates pretty well.
I’m enjoying your site..
By Ray Flowers on Mar 19, 2012
Matt- Thanks for the note.
I’ve been pushing ABA since 1951. Never heard of Alfred P. Berry, but it’s not a shock at all that someone would propose something similar – it’s not like it’s a complicated theory like Win Shares or Total Player Wins.
Trying to get some traction for ABA… been slow going.
By Ray Flowers on Mar 19, 2012
Mother – for batted ball data, fangraphs is great:
http://www.fangraphs.com/leaders.aspx?pos=all&stats=pit&lg=all&qual=y&type=2&season=2011&month=0&season1=2011&ind=0&team=0&rost=0&players=0
By Ray Flowers on Mar 19, 2012
Craig D – I almost always go skills over role. Doing the same thing here. I’d say Holland who is coming off a dominating season, versus Broxton who really hans’t been Borxton since 2009. He might get the first shot cause of his history though, certainly possible.
By Reed on Mar 19, 2012
Love the concept of considering the disparate impact of XBHs vs singles on a pitcher’s overall effectiveness. However, recognizing that no stat is perfect, it seems that ABA has a bit of a shortcoming. Compare these two scenarios:
A. Pitcher walks batter A, walks batter B then allows a 2 run single to batter C in one inning. His ABA would be 3.00. His ERA would be 18.00. WHIP is 3.00.
B. pitcher allows a triple. His ABA would be 3.00. His ERA would be 0.00. WHIP is 1.00
It seems to me that there should be some recognition that allowing a larger quantity of runners is detrimental, compared with isolated XBHs. Might it make sense to augment the formula to incorporate the disparity between these scenarios?
By Cliff Prince on Mar 20, 2012
Ray, I do like the idea quite a bit. I use a similar formula to rank hitters. Combining total bases + walks + stolen bases – caught stealing gives a nice balanced view of a hitters overall value in one neat stat.
@Reed, good point. That is certainly an exception to the rule. A bit more tweaking is needed it appears.
By Ray Flowers on Mar 20, 2012
Reed – It’s all about sample size. You give up a run in the first inning your ERA is 9.00. You pitch five more scoreless innings and your ERA is 1.50. You go 3 for 3 you are hitting 1.000. Seven hitless at-bats drops you back down to .300. You can pretty much skew any number if the sample size is small enough. ABA is no different than that. If you look at one inning, or one game the number could be crazy. Increase the sample size and things will normalize.
By Ray Flowers on Mar 20, 2012
Cliff Prince- See my comment to @Reed. I don’t think any tweaking is needed. We just need to be looking at a sample size that’s large enough to give a true picture.
By Reed on Mar 20, 2012
I hear you, but regardless of sample size a pitcher who allows more baserunners is worse than a pitcher who allows isolated XBHs.
IMO, a stat that equates a triple with 3 walks is lacking something. 3 walks is a far worse result than 1 triple.
I’m thinking something like ((HR+BB+H)+(TBA+BB))/IP. that penalizes the pitcher who allows a lot of baserunners, as well as the pitcher who gives up HRs. in this formula, the “points” each event adds to the numerator would be as follows:
BB: 2
1B: 2
2B: 3
3B: 4
HR: 6
As you can see, this compresses the ratio between singles/walks and XBHs, reflecting that a single/walk is more than 1/2 as bad as a double.
Also, in support of ABA you use an example that was limited to an analysis of a single inning
By Ray Flowers on Mar 20, 2012
Reed – The single inning example was just to clearly illustrate what ABA is. It wasn’t meant to be a definitive example of how valuable the tool can be in small sample sizes.
In regards to the theory itself…
I’m not trying to develop a definitive model. I’m merely trying to improve upon WHIP. Is WHIP fantastic? If it was, there would be no reason for ABA. However, everyone knows/understands WHIP, so making a simple “play” off that was my goal. There are much more complicated measures that do a way better job of computing pitcher’s performance than ABA. There is no doubt about that. The sole purpose though of ABA is to simply improve upon WHIP. Once you start throwing too much data at people either lose interest or get lost (I bet over half of baseball fans have no idea how to compute OBP).
Trying to keep it simple was my goal here.
By T.Hunt03 on Mar 21, 2012
Thanks for the tip: It could be a great stat, but
what about a walk and a 2B or a walk and a 3B? I
cannot see any absolute symmentry in this equation. Most of the time this works, calc.=90+/-%. Something to keep working on. Not a
Stat…
By Ray Flowers on Mar 22, 2012
THunt – I could keep working on ABA, of course it could be “better”. However, the whole point is to improve upon WHIP but do it in a simple way that doesn’t turn over the apple cart. If it’s too fancy, too involved the masses will tune it out. There’s a fine line between turning people off an offering something they would use.