Liriano's outstanding season has been crucial to the Twins' success. On a smaller level, it has served as a shining example of why the BABIP statistic is one worth considering whenever a pitcher's numbers don't seem to reflect his performance.
Back in mid-July, when Liriano's ERA had ballooned to a season-high 3.86 in the wake of a thrashing from the Tigers, I remarked that the left-hander is "one pitcher you don't need to worry about." I pointed out that Liriano's .361 BABIP (batting average on balls in play) was unsustainably high for any pitcher, particularly one throwing as well as Liriano. He was doing what he needed -- getting ground balls and strikeouts while avoiding walks -- but had unfortunately been victimized by substandard defense and plain old bad luck.
As usual, I took some backlash for suggesting that Liriano's poor results might not be entirely his fault. Many believe that BABIP is a farce, and that if hitters are running up a high batting average against someone it's more than likely the pitchers fault. One commenter snarked, "Sometimes I wonder if you bloggers watch the games." Yet, over time, BABIP has proven to vary randomly for individual pitchers while always remaining steady for the league as a whole. (Major-league pitchers this year have a .298 BABIP. Last year: .299. The year before that: .300. And so forth.)
Liriano's BABIP since I wrote that article? It's .306, right in line with the league norm and drastically lower than the figure he posted over the first-half. Consequentially, Liriano has gone 7-1 with a 2.84 ERA since the All-Star break (compared to 6-7 with a 3.86 ERA prior), and it's not necessarily because he's pitched any differently.
It should not, however, cloud our view of him as one of the game's very best pitchers, and a guy who will give the Twins an advantageous pitching match-up in the first game of the playoffs whether he's facing CC Sabathia or David Price.
22 comments:
I'll give you that batting average on balls in play is an important statistic. But i will draw the line at saying that Liriano was even close to being in contention for the Cy Young award. He wasn't that good.
Tim,
He WAS that good, but as shown, had some bad luck and not the greatest run support. This isn't the first person to suggest that he would have been among the top pitchers for the Cy Young. Voters are still too fixated on wins which aren't nearly as important as they used to be; see Santana v. Colon a few years ago. Put the Yankees behind Liriano with the run support and he probably adds 5 wins and is contending for the award if not the leading candidate.
I don't personally believe BABIP should be a major consideration when reviewing a player's season or judging his Cy Young candidacy. Bad luck happens, it's part of the game.
However, if we're talking about who the best pitcher is, or trying to predict how someone will perform going forward, it's very useful. If healthy, he should enter next season as the Cy Young favorite in the AL. With better luck and run support, he could easily win 20 games with a sub-3 ERA.
BABIP has holes, as do many advanced statistics.
Does Liriano give up ground ball rockets down lines? Rockets up the middle? In other words, do a lot of his ground balls, which he induces at a high rate, find holes that only hard hit balls find? Without analyzing every ball in play, we're not going to know, are we?
One hole Liriano has is fielding his position. With his wild follow through that has his back to the plate and his body half way to third base, he's not in good position to field balls that well. A part of his BABIP flop (albeit a small one) could be his ugly fielding mechanics.
Still, if he's on his game, there's no other Twins pitcher I'd rather have out there. He's earned the right to start game 1 and has the ability to dominate the very best lineups the AL (hello, Yankees!) has to offer.
BABIP is not pure chance. That much should be obvious. Using rough estimation, Liriano had about 180 outs in the first half of the season on balls in play. Give or take.
So if the BABIP truly should be 0.300, then after 180 outs the further the number is from 0.300 then the less likely it is for it to be that far off.
Each ball in play is an event with 30% chance of being a hit. But after a whopping 180 of these the result is 36%. Do you know what the odds are of being that unlucky? And then in the second half, basically the opposite happens. His BABIP is so low that it drops the overall average down to 30%.
Time to go get a lottery ticket is all I'm saying if you truly believe that is all luck.
Truth is he's been better in the second half. Balls in play are in general not hit as hard and therefore have been less likely to avoid an out. Grounders can't escape the infield. And hitters aren't able to hit the ball to the part of the field they want as often. Guy tries to pull one and it squirts to the first baseman.
He's just been better. Not luckier. Just better. Same with Blackburn. He still isn't getting strikeouts but half the damn grounders now are hit so soft that their only chance is an infield hit. Earlier on everything was hammered. Grounders, flies, liners. Now things are tapped.
Good pitchers induce ground balls, weak contact, and strikeouts. Liriano has made less mistakes second half and therefore hitters are squaring on the ball less frequently.
Reading this site I'd swear most of our pitchers are the luckiest or unluckiest people on earth.
I don't think I ever said BABIP is pure chance. That's not something I believe. What I said was that a .360 BABIP in the first half for a pitcher like Liriano was fluky. That mark is 30 points higher than Nick Blackburn's first-half BABIP, and if you think Liriano was getting hit harder than Blackburn you weren't watching the games.
You said yourself that a pitcher's job is to "induce ground balls, weak contact, and strikeouts." Liriano did all of those things much, much better than Blackburn. That's a fact. He still allowed a higher percentage of hits on balls in play. Fluke.
Baseball is a crazy game. When the ball is put in play, a lot of different things can happen that aren't necessarily under the control of the hitter or pitcher. It shouldn't be all that shocking that over the course of a half-season, or even a full season, a player's results often don't accurately reflect his performance.
It makes sense that BABIP will even out eventually. For weakly hit balls that find holes in the infield, you'll eventually have some rocket shots that are hit right at infielders.
What about the fact that he's only given up a handful of home runs all year? That seems to be indicative of how well he's pitched, even if some of it can be marked under good fortune.
Anon,
There is about a 5% chance of this happening. While this is a pretty low chance to happen to any specific pitcher, there is a very high probability it would happen to SOMEONE!
Basically, assuming it is completely random, if you take a sample of 20 pitchers, you would EXPECT this to happen to one of them.
Comparing this to a lottery is absolutely ridiculous.
Yeah anon's nonsense about babip was a poor effort. Obviously pitcher skill factors into babip but the fairly large variance in yearly babip indicates its not particularly independent of nonpitcher skill factors. And even if babip was all chance a season of pitching wouldnt be close to a large enough sample to be meaningful.
I also think youre hung up on semantics with regard to luck. I think people throw around the word luck to suggest that whatever they are referencing is likely to regress, not because it is super accurate. I think fluky is probably a better word. Pitchers pitch better or worse during stretches typically for specific reasons, too many hrs, lots of gbs etc. Pitcher skill certainly went into generating those specific results, but for small sample sizes theres a good chance the results are representative of the pitchers actual ability, and you end up with an unsustainable "fluky" result.
Nick/Tim,
Aside from his BABIP, Liriano's 6.3 WAR is 3rd in the league (Behind Cliff Lee and Roy Halladay). His FIP is a full point below his ERA (2.44 - 3.44). His FIP and FIPx are both good for 2nd in the league. All of these numbers points to not only Ace material, but CY Young material.
So this ties into what Nick is saying. Liriano's BABIP(Which he holds the 2nd highest in the league) along with the stats presented above show that Nick is right. Liriano has been quietly dominate, the only question I have about him? His nerves.
Babip is certainly as important statistic for measuring pitchers, however, the most alarming statistic is that Liriano has only given up 5 home runs and he doesn't have the greatest whip at 1.25. If he has the same whip next season, his ERA likely be around a 4 era. Look at Sabathia, Verlander, Hernandez and others. They give up on average 18 or so homers a year. Even if Liriano throws 200 plus innings next year, he will most likely give up 10-15 more home runs next year. You can't say he is an ace if he hasn't reached 200 innings in his career.
I am a complete doofus with stats but I find your conversation very interesting from what I can follow. All I know is what I see. When Frankie looks relaxed on the mound, staring down the batters with an expressionless face, throwing a good mix of pitches and not overthrowing the ball, he is devastating. That is when I love to watch him pitch. The pitching coaches just need to practice a little Zen with him, get him to deep breathe and meditate to calm his nerves. Be the ball, Frankie, be the ball...
well whip and babip are pretty closely tied. If his babip drops his whip will drop. Its true hes been fortunate with hrs, his batted ball luck has more than matched in the bad luck category.
well whip and babip are pretty closely tied.
Exactly. The reason Liriano's WHIP is so high is because he's given up so many hits. In 2005, his walk rate was essentially the same but his WHIP was 1.00, because he held opponents to a .205 average (aided by a not-so-unreasonable .283 BABIP).
Nick,
Do you even watch the games? Watching 30 or so Little League games a summer makes me recognize BABIP even more. In fact, I shouted it out in one instance. Liriano has lapses in concentration, but is still a top 10 pitcher. I almost went off the road when I thought he was hurt.
I suspect I was the guy with the snarky comment in July. I still say that when Liriano has bad games, and he hasn't had many the 2nd half, it is because he is not commanding his fastball. He has great stuff, really all the time. In games when he gives up bunches of runs, it is generally always because he can't command his fastball.
He gets behind in the count, people can lay off his slider(which nearly always ends up outside the strike zone) because they are ahead in the count and can lay off marginal pitches. When that happens he has to groove his fastball and he gets hit hard.
The biggest difference in the 2nd half is that he nearly always has had good command of his fastball. The few times he didn't he was hit hard, just like the first half.
I don't doubt he may of had better "luck" and perhaps better defense in the 2nd half, but good pitchers like Liriano pitch around bad luck, if they have their command.
BABIP is a great stat for forecasting fantasy stats. You bank that the pitcher you have some faith in is good enough to not deserve the BABIP he holds. It then allows you to draft bargains. This is the real thing. We aren't doing a playoff re-draft, so BABIP is not going to help much. This is about picking a game 1 starter.
Nick, good call forecasting Liriano to bounce back using BABIP. This was a savvy observation and you deserve credit. As far as game 1 starter, lets look at how well he is pitching right now. BABIP is a long range forecasting tool, and I bet you can make a case for Liriano without it.
Dave, that's silly. The reason BABIP is good for forecasting fantasy stats is that it can tell you how the guy is likely to pitch going forward--not just "long range," but in his very next start. There's absolutely no reason to draw a distinction between its usefulness for next year's fantasy draft and for Game 1 next week. Same thing.
Bill, I believe your comment is silly.
Look at what Nick was saying earlier in the season. Liriano's stats were skewed because his BABIP was inflated compared to his norms. Did Nick say that he would all of a sudden shed his BABIP and return to the norm right away? No! Instead of that he said as the season progresses his skewed stats will trend down because his BABIP will trend down as well.
BABIP in isolation is meaningless. Of course after getting rocked a pitcher will have a high BABIP for that game (except for really bad control issues). Does that help you determine if he will get rocked the next game? Does that help you determine if there are mechanical issues the pitcher needs to work on or if they were just unlucky? The rub with BABIP is that bad pitchers have higher BABIP's (and conversely good pitchers will have low ones). The value in BABIP is that every pitcher will have a career average BABIP to compare their current season BABIP with. It is this comparative aspect that makes it valuable. As with any aggregate statistical measure, the larger your sample size the more reliable it will be. The same goes with the stats you are trying to predict. Your prediction is much more likely to be accurate if you are predicting for a longer period of time. If you are predicting for one game, a more localized statistical analysis is required.
I used fantasy as a good comparison, but BABIP has valid long term forecasting uses in real baseball as well. Contract negotiations, trades, promotions, roles; all of these can benefit from a long term BABIP analysis. I don't think you will find even hard core sabrmatricians that believe BABIP is a single game performance predictor.
Dave, I'm 100% certain that Nick didn't mean to imply what you seem to be suggesting -- that Liriano's stinky BABIP would gradually get less stinky, on a game-by-game basis, until it was finally back at "the norm." That is, the implication wasn't that with a .361 through mid-July, he'd have, say, a .340 BABIP for his 6-7 starts in August and a .320 for September and maybe get back around .300 in time for next season. The implication, rather--and stop me if I'm putting words in your mouth, Nick, but if so we're going to have to have a serious talk :)--is that in any given game, Liriano's as likely as anyone to keep his BABIP right around the league norms, but that his season number would "trend down" because it was already at such a high level that it was going to take a lot of time at the league norms for it to come back down to earth. In fact, as Nick's fifth paragraph above indicates, that's exactly what the actual results show happened. Liriano's BABIP for his next 8 starts after Nick wrote that piece was .301.
The only reason BABIP isn't a good "single game performance predictor" is that there's no such thing; there are way too many variables that go into a single game (or even a few of them) to make it worth trying to predict exactly how a pitcher will do. Which I guess is what you're saying with your second paragraph above. But you seem to be jumping from Obvious Fact A--that Liriano's BABIP isn't a guarantee that he's suddenly going to stop giving up so many hits in the playoffs--to Crazy Conclusion B--that we should ignore BABIP altogether and focus on "how well he is pitching right now" (which you can't actually determine without considering BABIP). So in other words, of course there's a huge sample size issue in predicting his actual performance in that one game--Liriano could go out and throw a no-hitter in the playoffs, or could give up twelve hits in three innings. But the justification for taking his flukey BABIP into account is exactly the same for planning your playoff rotation as it is for planning your 2011 fantasy draft.
His fastball is too hittable. For a guy who can at times light up the radar gun and who has a great pitch(slider) a good changeup its really strange. Frankly he really scares me opening up the playoffs.
I disagree. You can look at bullpen sessions, recent pitch movement evaluations, recent pitch command indicators, and how well he is hitting his spots in recent games. Key word recent. The best way to predict the weather is to predict the same weather that occured the day before. BABIP is a way to see if a pitcher is pitching to his norm and allows us to see beyond base statistics (ERA, WHIP, W, L). I am in no way saying rely on these obviously skewed stats, but rely on a more detailed analysis. BABIP is the opposite of detailed.
Post a Comment