De-Lucker! 2.0: Hot, Fresh, New xBABIP
Fare thee well, father, mother. I’m off
to de-luck the f*** out of this s***.
Let us delve once again into the numbers.
With this All-Star break forcing to watch so little baseball, we now have a moment to drink up the frothy milkshake of statistics from the first half. So, you and I, we shall dissect the stats and find out who has been lucky, unlucky and a little of both.
Combining Fielding Independent wOBA (FI wOBA for shortsies) and slash12’s xBABIP, we can get a specific wOBA calculation that strips away unusual luck, whether good or bad. It is important to remember these are both regression-based calculators, so they are backward-looking, not forward looking. Please do not pester me in 3 months when “[X Player] didn’t suck like you said he would! lulzlulzlulz, ur dum!”
…
Is BABIP the only stat that has a lot of luck in it? No. Stuff like home run rates can be wild early too. And moreover, BABIP is many parts skill, several parts luck. But there is more luck (or random variation) in BABIP than probably any other hitter stat out there. That is why it is worth focusing on it here.
Praise be to Jeff Zimmerman and Robert Boden (slash12) for updating the best xBABIP tool out there, and making it even better. They detected what I myself detected in my previous De-Lucker! rendition — that infield shifting and decreasing offense league-wide have rendered previous xBABIP tools, shall we say?, off.
These new De-Lucker results should be most splendidly accurate! Oh my!
(For the following calculations, I used their 2009-2011 constants.)
SORTABLE! FILTERABLE! DOWNLOADABLE!
(Note: Minimum 45 PA, presently filtered to show only 250 PA or more.)
Reactions to the De-Lucker results:
- • The FI wOBA sluggers of note? With only a 45 PA minimum, our top five FI wOBA hitters are:
1. Matt Kemp
2. Anthony Rizzo
3. Salvador Perez
4. Yunel Escobar
5. Drew Stubbs
With our 250 PA limits, we get:
1. Josh Hamilton
2. Joey Votto
3. Ryan Braun
4. Jose Bautista
5. Adam Dunn
- • And the laggardboard? With 45 PA min.:
448. Ryan Raburn
449. Danny Valencia
450. Miguel Cairo
451. Johnny Cueto (congrats on getting 45 PA, pitcher — lolololol)
452. Brett Hayes
And with the 250 PA threshold:
401. Brandon Crawford
403. Willie Bloomquist
408. Brennan Boesch
413. Sean Rodriguez
429. Clint Barmes
- • Dee Gordon and Cliff Pennington rank as this rendition’s winners of the Fear Thy Regression as both players will still be well under .320 wOBA if their BABIP regresses but their peripherals don’t improve.
- • Austin Jackson ranks out our biggest Regression, So What? Award candidate, as his wOBA and De-Luck’d FI wOBA are a whopping 61 points apart. Still, FI wOBA says he will settle around .348 wOBA, which is still great, but that is presuming Ajax will slip to a .319 BABIP — which is some 60 points lower than his career rate.
If he meets his career number (.378 BABIP), his expected regression slims down to 24 points at a .385 FI wOBA.
- • I wouldn’t put too much stock in Adam Dunn’s .414 FI wOBA. Unless he meets his .337 xBABIP — which would be about 40 points above his career BABIP — he won’t sniff anything much higher than a .384 wOBA.
- • Jose Bautista and Adam Jones — both playing under their established BABIPs, both sporting the peripherals of much higher BABIPs. Watch out world. They are become death.
You can De-Luck! too!
Click this button to download the above Excel (.xlsx) file:
Encircled and enarrow’d.
Then use this custom leaderboard and paste your results and limitations and queries into yonder first tab — so titled “De-Lucker 2.0 Full.” And pesto sauce! You’ve got your own, updated De-Lucker! results!
ALSO NOTE: Downloading the file will allow you to see the BABIPs and xBABIPs. They simply cannot fit in this narrow column.
This is sweet! Random thing, but I couldn’t find AJ Pierzynski in there. I tried sorting by White Sox, to no avail. I’m curious how much his hot-fire heat is for real.
I see him having .362 wOBA, .387 FI wOBA with a .025 positive difference.
His great numbers this year have been fueled by an uptick in HR% — he’s actually got a lowish BABIP. I would say his HR-rate may be a bit flukish, but it’s much less common for HR-rates to go this far haywire. Who knows? Maybe he Bautista’d?
Oh word, that’s really interesting, Thanks a lot!
Wow, so the Mariner’s have one guy outperforming expectations (Michael Saunders, -0.001), another at just under his projected performance in Kyle Seager at 0.006, and then their other FIVE listed regulars are all at least 0.028 in the hole, averaging out to ~0.037 points for those five. Smoak, Ackley, Montero, Ichiro, Ryan. Is that the Safeco effect in action?
That very well may be. FI wOBA, for all its majesty, cannot account for ballpark adjustments.
Oh, that wasn’t intended as a dig at FI wOBA, my thought was: are 5 of the Mariner’s best hitters all being extremely unlucky at the same time, or is there some other effect in play? It’s felt as though they have been extremely unlucky (hard-hit balls right at people, warning track shots, etc.) this year, but it’s difficult to count myself as an ‘objective observer’.
Didn’t think I could feel any worse about Adrian Gonzalez’s season…
i’m surprised that he has actually been LUCKY this season. that’s just… are we sure this is Agon? i thought his power was supposed to get better this year now that the shoulder surgery is in the rear view…
what exactly is keeping moustakas so low? his BABIP and HR/FB aren’t crazy. is it the low LD% or the high amount of popups? aside from infield hits, his batted ball data is almost identical to edwin encarnacion, who isn’t ticketed for a similar regression.
And Moustakas has been a candidate for regression since this column first came out several months ago, but his numbers have held up.
He’s got a VERY low xBABIP.
Not quite sure why.
Moustakas is 3rd of 155 qual batters in both FB% and popup %, which murders xBABIP.
mike moustakas
FB%: 50.2 %
IFFB%; 17.6 %
de-lucker diff: -.036
edwin encarnacion
FB%: 50.6 %
IFFB%:15.5 %
de-lucker diff: +.009
so 2% if IFFB% is enough for .45 of wOBA? i’m skeptical.
.045 rather. you know what i mean.
Encarnacion also has a much higher home run rate. I suspect Moustaka’s might join his soon, but 5.2 point difference right now puts a world of hurt on Mous’s xBABIP.
Trying to show Drew Stubbs does not stink is a waste of time.
Is this better than just using the batter’s career BABIP to project future BABIP?
No. But it can be more current and more easy.
In all matter, I would say refer to their career BABIPs. Their xBABIPs are backwards looking, saying what kind of BABIP they would have given their batted-ball types. But, if a BABIP veers abnormally from a career BABIP, then the player (or the pitchers) will likely make adjustments to return to their career BABIP.
If they don’t fix their mistakes, then they will likely end up in the minors. If pitchers don’t react to their positive changes, then the hitter will end up in the HOF.
Very disappointing seasons for Rickie Weeks and Desmond Jennings, who have been .300 wOBA hitters without being unlucky.
ok, so my apologies, but can someone help me with how does one use this data. IE, Billy Butler, could someone explain what each of the number-set tells us and what we can use that to predict? My apologies on my ignorance, just learning some of these more advanced metrics/applications.
Butler has a wOBA of .364 but a FI wOBA of .383. The author is saying that Butler has been unlucky and we should see his wOBA increase.
I wouldn’t take this to heart as there are many reasons why a particular player would be over/under-performing his peripherals. Also, it might be his batted ball data that changes rather than his wOBA.
See: Dan’s comment.
This data indicates Butler had a better first half than his numbers indicated. If, upon inspecting his numbers yourself, you find that to be an accurate statement, THEN TRADE FOR BUTLER IN YOUR FANTASY LEAGUE!
Butler’s difference is positive, so that means that assuming he maintains the batted-ball rates that he has put up so far this season (which is more reliable than say, just his OPS or wOBA), he will perform better than he he has up to this point.
I didn’t pore into the methodology, but I wonder if this might be skewed a little to favor LH hitters against whom defenses play extreme shifts? I assume that the system thinks that when Adam Dunn hits a soft liner to short right, it should be a hit. In reality it shouldn’t be, because there’s a reason the fielders are playing there. On the flip side, there is nothing unlucky going on when David Ortiz slaps a soft grounder to third and it becomes a hit.
Again, no idea how this is factored in, and it probably doesn’t make a huge difference. Just curious.
Yes, that is correct. The inherent problems with all xBABIP calculators are chiefly: (1) league BABIP is constantly changing and (2) pull hitters (especially left-handers, for now) have their BABIPs reduced by the shifting.
Until we can find a way to systematically account for that, we will just have to keep those limitations in mind while looking at the data. (And always refer to career BABIPs in the case of oddities.
I thought I remember Dave saying that shifted plays are disregarded for UZR – why can’t we do that for BABIP?
And UZR and Fangraphs made the right call on ignoring the shift for dWAR calculations. B-Ref’s DRS is doing the opposite: not only using these plays but giving extra credit to plays made in the overshift by third basemen, since when these batted balls are hit to the right side other third basemen are, you know, actually standing in the vicinity of 3B. Brett Lawrie’s ridiculously high dWAR has turned him into the AL WAR leader according to B-Ref and has made him a scapegoat for dWAR inefficiencies.
Is there a way to do this for pitchers?
Hmmm… That’s an interesting thought. The whole FI wOBA concept derives from the FIP formula. I guess there’s still the possibility that BABIP is affecting FIP to a significant degree (such as an unusually low BABIP increasing the K/PA and such), but I think the benefits might be marginal.
Maybe I’m looking at it from the wrong way, though.
A quick question to help clarify the application of the lovely De-Lucker: How many PAs does it take for batted ball profile to stabilize? I thought it was curious to see some serious fly ball hitters faring so poorly compared to their xBABIPs (namely Tex and Granderson), because I thought it based it on batted ball profile rather than league average BABIP. After further digging, both those guys are hitting way fewer fly balls this year compared to the last two years. If that trend continues, they’re in for some positive regression in AVG. If, however, they revert back to their fly ball ways, the downtick in xBABIP might wash out their expected gains. Any insight on whether we should believe their new LD/GB/FB line, or bank on it breaking down closer to their marks over the last few years?
These are good questions. Personally, I believe batted ball rates normalize quickly (maybe 50 to 100 PA) but they are also a moving target, like what you are alluding to. So in general, we expect players to come close to their career norms, but they also go through periods of adjustment where they are struggling or excelling and pitchers change how they’re pitching them.
As far as a study that offers some sort of mathematically precise means knowing when batted ball data is “real,” I do not believe there is such. I think there have been attempts, but the truth is a bit more complicated: We just have to use our best judgement in matters of significance, say I.
First off, I really love this continuing series.
Second, I don’t understand why you are calling out Dee Gordon and Cliff Pennington as “Fear Thy Regression” candidates. Their FI wOBA are both just a shade over .300, but they’re both shortstops. They’re not supposed to slug like a first baseman. Speaking of which, Ike Davis is my candidate for “Bad Beyond Luck”. The De-Lucker gets him up to just .322. Not exactly the savior the Mets thought they had from 2011.
Good points all around.
so bautista and adam jones are due to have a big 2nd half?
Not necessarily. This is only suggesting that their batted ball data so far indicates that they should have been more productive, and may have been unlucky. BUT, it is entirely possible that the batted ball data they have accumulated so far is unsustainable. Adam Jones, for examples, has a HR/FB rate far above his career norm, which is why ZiPS projects him at a lower wOBA in the second half than in the first. If he stops getting such a high percentage of his fly balls to go over the wall (21% is awfully hard to keep up), then he may be worse.
Thank you for the response, re: Bautista and Adam Jones. In this case, then Moustakas has been very lucky with the production he has put up? In this case, we should expect a decline in production for Moustakas unless his batted ball profile changes?