Unlucky R-Ho
Considering most people look primarily at the previous year’s statistics when drafting a player, one place to gain an edge on the competition is to attempt to put last year’s statistics in context. By this, I do not necessarily mean the run-scoring environment the player played in, but rather why their statistics seemed to vary from their usual numbers. To do this, I will start with a series of good luck and bad luck players. The luck comes in the way their batted balls in play are turned into outs or hits. Peter Bendix and Chris Dutton did some great research on defining how we should look at BABIP and what drives it.
Over the next week or so, I’ll go through some players that were on the extreme ends of the luck spectrum with the idea that those players who were very lucky last year will probably come back to earth and those that were unlucky are due to breakout. This can be very helpful when deciding to draft one player or another. The first player in the series is Ryan Howard.
Last year, Ryan Howard had arguably his least productive year as a Phillies player (don’t tell the BBWAA). His BA was a down to a career-low of .251. His OPS was also the lowest of his career at .881. He maintained his ability to hit for power and drive in runs (48 HR and 146 RBI), but his average made him a slightly better version of Adam Dunn to fantasy players. He usually more closely resembled a top 10 fantasy player.
Looking at his new xBABIP and his BABIP we see where some of his batting average was lost. He seemed to be incredibly unlucky on balls put into play.
His BABIP of .289 was .019 below his new xBABIP of .308. He was robbed of roughly 7 hits, and his average should have been closer to .262. While this difference doesn’t seem all that great, it really is. Every single has a large effect on OPS as they are double-counted in the formula because your OBP is increased by one more time reaching base and SLG increases by one more additional base gained. The SLG is even more affected by these missed hits because if they turn into 2B or 3B (pretty unlikely with Howard) they could add 2 or 3 bases.
In Howard’s case, if you add these 7 hits broken down the way his hits were over the season his OPS goes up to .906. His slash line would increase to .262/.349/.557. He might have also seen an increase in his league-leading RBI totals. If you play in almost any league Ryan Howard is a good player to have on your fantasy team. He may be able to had for a more reasonable price than in years past because of his “down year” last year.
Yet he is also lucky as he gets to play home games at Citizens, probably nullifying those 7 hits IMO.
He’ll still be there next year. He’ll also still have all those table-setters in the lineup.
Lucky how so? CBP has graded out as a slight hitter park according to Park Factor most years. It’s reputation as Coors Field East isn’t exactly deserved.
Definitely a solid lineup…
How do you adjust BA for xBABIP?
You find out how many hits they would have had with the xBABIP. To do this you just rearrange the BABIP formula:
H=(BABIP)*(PA-HR-K-BB-SF-HBP)+HR
Although, after reexamining their use of BABIP it seems they don’t include SF in their BABIP calculations. In future posts where I reference Bendix and Dutton’s xBABIP I’ll likely use that version because it will make the differences more stark.
Thanks! I always wondered how you guys do that.
OK so it seems Bendix and Dutton do not include SFs in their BABIP formula. I was wondering why they had different BABIPs on the year from the numbers we have here. Using BABIP=(H-HR)/(PA-HR-K-BB-HBP), Howard’s numbers look even better. Controlling for his typical breakdown of hits in play, R-Ho’s slash line actually would have looked like this:
.265/351/.561/.912
I dont fully understand the whole BABIP luck theory. Here are his seasons BABIP and BA
08: .289 .251
07:. 336 .268
06: .363 .313
05: .358 .288
I dont see any significant correlation between the two, in fact the R squared value is .8073, not exactly a strong indicator.
It seems high K players are systematically going to have lower BA than BABIP bc they dont put a lot of balls in play (via the K) that still effects their BA.
If someone can statistacally explain why luck caused Howard hit .313 in 06 and mid .200s in 07 and 08 Id like to hear it. Personally I think it has nothing to do with stats and everything to do with skill, or lack there of. Howard has big holes in his swings, he misses a lot of balls, and the stats you should be using to describe REAL LIFE skill are plate discipline. Howard has too high of O-swing and too low of Z-contact, a bad combination for BA.
You are right in that high-K players will generally have lower BAs than BABIPs. The role of “luck” is determining whether or not the BABIP was “deserved.” Traditionally to get xBABIP you merely added .12 to LD% to get their xBABIP (what their BABIP should have been). The article I linked above of Dutton and Bendix’s work found a formula that predicted BABIP much better than the traditional method. This method incorporates speed, LD%, plate discipline, and a variety of other factors in composing xBABIP.
Here is Howard’s BABIP and xBABIP for the last 4 years:
YEAR BABIP (actual) New x-BABIP
2005 .352 .313
2006 .347 .316
2007 .322 .320
2008 .282 .308
As you can see, his true “BABIP-skill” varies very little from year-to-year. The luck, though, is what plays on this skill and contributes to fluctuations in his actual BABIP.
Hope this helps and check out the article it’s a great read.
I’m not sure what kind of correlation you’re looking for, but an R-squared of .8073 is actually a VERY high value.
I’m a hardcore sabermetrician, a fundamental believer in firing Joe Morgan, and I hate this saying, but clearly I have to pull out the “I watch the games” card. Did you write this article after watching him play?
The reason that Howard’s BABIP has gone down so much is due to the shift. Teams have not only been employing the shift on him more so than ever, but they are also doing it better. Compare his BABIP when there are runners on, especially runners on second base, and when there are bases empty. The former case limits the defenders ability to put three infielders in shallow rightfield where Howard hits an obscene number of line drives and groundballs.
His BABIP is likely to go up but the .352 and .347 from 2005 and 2006 are things of the past.
I have watched Howard play and Papi and Pena and the vast amount of other players who get these extreme shifts. To think that the shift is the sole cause for him under-performing is giving it too much credit I believe. In 2007 Ortiz’s xBABIP was .302 and his BABIP was .352. The article that I linked addresses this specifically:
“For example, if Jacoby Ellsbury hits a ground ball in the hole between short and third, he has a higher chance of getting a hit than if Bengie Molina hits the exact same ball in the exact same place. Anecdotally, this is how Ichiro manages to get so many hits every year. And fans of the Red Sox, Yankees and Rays can tell you that David Ortiz, Jason Giambi and Carlos Pena have been robbed of many a base hit because of the extreme defensive shifts used against them, whereas Dustin Pedroia, Derek Jeter and BJ Upton have gotten more hits because of their batting eye and their ability to use the whole field. Surely, these factors contribute to whether or not a batted ball becomes a hit.”
If you take the time to read the article you may gain a better understanding of the statistic. This is not an end-all-beat-all it is merely another tool in the toolbox. If you want to ignore the role of luck in a batter’s statistics feel free, but I’ll gladly take every advantage I can get when trying to guess what a player will do next year.
Don’t argue with a straw man. Argue with me. It’s harder but more fun. I have a strong belief in BABIP being a luck statistic and have conducted a plethora of analysis into how much is explainable and how much is luck. I appreciate you linking that article which I missed at the time, and I have reached a lot of the same conclusions in my own research as well. So don’t try to argue with me pretending that I’m claiming that Howard’s BABIP isn’t unlucky. It is. But as I wrote the first time, the days of his .350 BABIPs are over. I actually think .308 is probably fair, but I think there’s a reason that “spray” was an important statistic, and you need to cite that if you’re going to discuss his performance and it’s misleading to post his old BABIPs without acknowledging the reason for the change. Check out Howard’s BABIP splits depending on where runners are. I’ve made the mistake of thinking he could repeat significantly above average BABIPs, but he simply won’t have the same BABIP on groundballs that is necessary to repeat that elite performance.
I completely agree that we likely won’t see BABIPs of .352, but I would also say that it’s unlikely we see a BABIP of .250. That is the essence of BABIP and the luck that’s built into it.
I didn’t point to the spray effect in the post because I wanted people to read the article. It’s incredibly interesting and insightful, and it really helps to put this series of posts in context. Their model is great at saying how a given player should perform, so when one varies so far from that ideal we wouldn’t expect them to repeat those results. THIS is the idea behind the posts. Not that they will rebound to their higher-than-normal BABIPs, but that others will view them as being luck neutral and if you put them in the context of luck it can be quite helpful.
As far as his BABIP-splits, I see what you are saying. The one thing that worries me about those is the whole SSS issue. Just because it fits the way we expect it to come out does not necessarily mean it’s true. It does seem to confirm our idea, though.
MattS, good post, they need ppl like you to be “editors” or “reviewers” of a lot of sites SABR articles bc many cherry pick stats and dont follow real life happenings.
Just wondering how much Howards numbers vs league avg, vs “rbi” men, and I think I might punch anyone who says rbis arent a legit stat, compared w/ RISP vs 1st only vs bases empty….
Some things just cant be explained by stats, they can only be explained by observing real events.
Ryan,
You disprove the notion of using BABIP as an indicator in your article…
BABIP = (H-HR)/(AB-K-HR+SF)
You talk about players ability to hit to all fields, have good batting eyes, not face defensive shifts, have speed. Where are these factors in the BABIP equation??????
To me, a lot of these stats are found backwards not forwards, meaning a bunch of numbers, when manipulated, come relatively close to a known number. To me FIP is the epitome of a made up stat, FIP=C1 * [C2*HR+C3*(BB+HBP-IBB)+C4*K]/IP. To be truly a stat that disregards defense, shouldnt there be some sort of account for the defense? Lets say the rays had 8 gold glove position players while the yankees had 0, obviously, a pitcher of equal abilities will have different stats due to the defense behind them no their HR, BB, K, IP
Where is speed in H/AB? Where is defensive shifts in TB/AB? These statistics don’t take any of these “factors” into account yet they do have a hand in them. IF YOU READ THE LINKED ARTICLE YOU WILL SEE HOW THEY FACTOR IN xBABIP. I’m not just going to parrot what it says. Do yourself a favor, and read the article. Then, if you have questions I’ll be more than happy to address them.
RE: FIP
The point is not to account for defense the point is to disregard defense. So, it looks at the only outcomes a pitcher truly has control over.
A couple interesting notes about FIP:
1. It correlates from year to year much better than ERA, so it is thus regarded as more “predictave” than ERA.
2. If you take a team’s FIP, UZR, and wOBA and figure the total win-values (or just add up the ones on the player pages), you get a great predictor of W-L record. Better than pretty much any out there.
Sorry about before, I thought you meant read your article, not the hardballtimes article… the variables for xBABIP seem to be more realistic.
Before I start, let me say I believe predicting/projecting/statistically evaluating pitching must be 100x harder than batting bc there are a lot of factors and they can change rather quickly.
FIP only includes HR, BB, K, IP, but why would you “punish” a pitcher, who has a good defense behind him, who doesnt pitch away from contact? Isnt that being smart and utilizying the ‘tools’ available to you? If you know you have great defenders behind you, and you dont try and strike batters out but rather have faith in your defense and thus allow yourself to pitch deeper into games be a plus not a negative? What about GIDP? if you can get two outs with one pitch, shouldnt that be a sign you are a good pitcher?
I have yet to form an opinion on using FIP for teams, but using it for individuals seems too convenient. If a team has bad defense, their pitchers should have FIP’s lower than their ERAs
here are the ERA/FIP of the 10+ GS pitchers from FLA last year-the worst defensive team IMO
Scott Olsen-4.20/5.02
Nolasco-3.52/3.77
Miller-5.87/4.00
Hendrickson-5.45/4.76
Johsnon-3.61/3.37
Volstad-2.88/3.82
Sanchez-5.57/4.87
If anything, FIP suggests the marlins have a good defense, but I dare you to find one person who actually watches the baseball to make that claim
I’ll address this point first:
FIP only includes HR, BB, K, IP, but why would you “punish” a pitcher, who has a good defense behind him, who doesnt pitch away from contact? Isnt that being smart and utilizying the ‘tools’ available to you?
How would you account for that? What statistic do you see that shows how different pitchers pitch differently based on who’s behind them? Additionally, let’s pretend a pitcher does “pitch to contact.” If they do, then wouldn’t their walks and HBP decrease along with their Ks? This would help their FIP as the loss of Ks hurts it. A great example is Derek Lowe. look at his K-rates and BB-rates. He is the ultimate “pitch-to-contact” pitcher, yet he still has very good FIPs year after year.
What pitching stat do you prefer? You are anti-FIP, but offer no solution. What you rather I use?
“You are anti-FIP, but offer no solution. What you rather I use?”
Thats my whole piont about some of these stats, they arent any good. Just bc I dont have the time to develop an alternative to FIP doesnt make FIP a good stat to use. How about you not use FIP at all?
Quick math lesson, which shouldnt be necessary for stat ppl, FIP=13*HR+(BB+HBP-IBB)*3-K*2 / IP. Lets assume HR stays constant (15), K/BB ratio stays constat.
YR 1, pitch for Ks-50 BB+HBP-IBB 150 K, 180 IP
YR 2, pitch to contact, 40 BB+HBP-IBB, 120 K, 190 IP-the increase in IP is bc less pitches are thrown, allowing one to stay in the game longer.
YR1 FIP=13*15+3*50-2*150/180=3.45
YR2 FIP=13*15+3*40-2*120/190=3.59
As you can see, bc BB+HBP-IBB is multiplied by a different factor than K, which is not related to K/BB, the effect on pitching to contact vs pitching for K isnt linear.
If someone developed a stat that accounted for the defensive players defensive abilities behind them, ie range, errors, miscues… then Id be interested in a fielding independent stat, till then, FIP is fundamentally innacurate IMO. How can you have a fielding independent stat that doesnt calculate the effect a defense has???
How do you explain FIPs numbers that say the majority of the florida marlins SP benefitted from the defense that was being played behind them when clearly their defense was horrendous????
Howard was a lot less productive in ’08 because he drew half as many intentional passes — 18 less. That pretty much would have brought his walks total comparably close to previous seasons.
Yeah, he was a little unlucky, but the main problem is that pitchers are challenging him now instead of giving him the Barry Bonds treatment.
His PA against LHP has also gone up each season since ’06.