Don’t Blame This on Dusty Baker
It’s safe to say that Dusty Baker is not exactly the most progressive manager in baseball. Baker has always been an old school guy, and relative to what is becoming the new normal, he’s definitely a traditionalist. So as we live-blogged game five of the NLDS last night, one of the common questions in the early innings is what mistake Baker would make that would cost his team the game. The expectation was set in advance; if the Nationals lost, Baker’s old-school philosophies would be part of the reason why.
Of course, the Nationals did lose, 4-3, in one of the most intense playoff games you’ll ever see. And as is the case in any one run loss, there’s always second guessing of decisions the next morning, wondering if things would have turned out differently if other choices would have been made. But in reality, Dusty Baker’s big expected mistake last night never came, and we shouldn’t blame him for last night’s loss.
The primary decision being second-guessed this morning is Baker’s decision to remove Max Scherzer from the game following Joc Pederson’s game-tying home run to leadoff the seventh inning. To that point, Scherzer had been absolutely dominant, and the decision to remove him led to a procession of pitching changes, with various pitchers unable to hold the Dodgers offense down; by the time the seventh inning ended, the Dodgers were up 4-1. But in this case, the data supports Baker’s decision to trust his relievers over a tiring Scherzer.
At the point he was removed, Scherzer had faced 24 batters, and he was attempting to get through the Dodgers lineup a third time. You’re certainly familiar with the times-through-the-order penalty by now, and Scherzer, like pretty much everyone else, gets worse the more often hitters see him in the same game. Like most high-quality pitchers, his drop-off is smaller than the league average, but it’s still there; his career .303 wOBA allowed on the third time through the order is 17 points higher than the .286 he allows the first time he faces hitters in a ballgame.
But it wasn’t just the fact that Scherzer was going through the lineup a third time; he was also going to have face left-handed hitters a third time. And Scherzer, as great as he is, still has some weaknesses against lefties. His career wOBA allowed against RHBs is just .264, but spikes up to .321 against LHBs; the gap was even more pronounced this year, at .206/.323.
And when asked to face left-handed hitters a third time through the order, Scherzer is just simply not a great option. In 1,101 career plate appearances against LHBs a third time in the same game, those lefties have hit .276/.350/.448 against Scherzer, good for a .348 wOBA against him. When he’s tiring and faced with left-handed hitters, his strikeouts go down (23.4% K%), his walks spike (9.4% BB%), and his home run rate (1.17 HR/9) becomes a real problem. He’s also allowed a .345 BABIP to LHBs when facing them a third time through the order, so there are just no real positive indicators here.
In the seventh inning, the Dodgers were sending up Pederson (LHB), Yasmani Grandal (switch-hitter), and Andrew Toles (LHB), and if a rally started, a pinch-hitter for the pitcher, then Chase Utley (LHB) and Corey Seager (LHB). Scherzer, who started the inning at 98 pitches, was facing a gauntlet of lefties, and while he’d looked good in the sixth inning, history suggests that you don’t want to leave Scherzer out there to face a bunch of lefties if things start going off the rails. Given his dominance and the fact that he was throwing a shutout, asking him to try and get through the seventh was reasonable, but at any sign of trouble, Baker should be ready to remove Scherzer and go with his deep, quality bullpen.
Which is exactly what he did. Scherzer gave up a leadoff home run to Pederson, so Baker went and got him, turning things over to a bullpen group that was terrific during the regular season. Of the relievers the team carried in the postseason, they combined for a 63 ERA-/79 FIP-/87 xFIP-, pretty similar to the 71/71/82 marks that Scherzer himself put up during the regular season. On a per batter faced basis this season, the relievers in the Nationals bullpen were about as effective as a rested Scherzer.
| Pitcher | wOBA allowed, 2016 |
| Mark Melancon | 0.226 |
| Shawn Kelley | 0.264 |
| Max Scherzer | 0.267 |
| Sammy Solis | 0.272 |
| Blake Treinen | 0.280 |
| Mark Rzepczynski | 0.304 |
You could probably quibble with going to Rzepcczynski first, rather than Solis, given the run of lefties due up and the likelihood that Dave Roberts would force Scrabble out of the game by pinch-hitting some righties — but he did get Solis in the game to face the team’s top two hitters in the lineup, which caused Roberts to send Carlos Ruiz to pinch-hit for Chase Utley.
This was, of course, the big at-bat of the inning, but if you had to honestly ask the Nationals and Dodgers front office guys who had the advantage in a Solis-Ruiz matchup, I bet everyone would admit that was a better spot for Washington, not LA. Ruiz might have had the platoon advantage, but he’s still a slow-footed backup catcher facing a quality pitcher in a situation where a double play ends the rally. And, while the results went against the Nationals, Solis got exactly what he wanted; a weak (85 mph exit velocity) groundball that, if it was hit a few inches to the left, would have been the double play the Nationals were looking for.
The ball hits Anthony Rendon‘s glove. This is about as textbook a “game of inches” play as you’re going to get, but because Ruiz’s grounder was just out of the third baseman’s range, the Dodgers took the lead, and the rally continued. Baker, again going with the data, wisely ignored Justin Turner’s reverse-platoon split, and went after the Dodgers best right-handed hitter with the team’s best right-handed setup guy, Shawn Kelley. Kelley held RHBs to a .173/.203/.336 line this year, a paltry .224 wOBA, and has been death to right-handed hitters his whole career.
Turner won that battle, though, doubling off the wall and bringing in two more runs. And that hit proved to be the decisive blast. But you can’t fault Baker for turning to Kelley against a right-hander; that was also a good match-up for Washington. Turner just won the battle, but questioning having Kelley on the mound there is using hindsight at its worst.
Despite expectations, there was no monumental mistake from Baker last night. He didn’t send Jayson Werth around third base when Corey Seager was already holding the ball on the relay home, after all. He took Scherzer out perhaps a batter too late, but that was at a point when no one was calling for Scherzer to be removed, given how dominant he had been up to that point, and he certainly didn’t take Scherzer out too early, as some have suggested since. He got good match-ups with good relievers against weak pinch-hitters, but had to watch as a seeing-eye single found a hole. That’s baseball.
And even as old school as Baker is, he still hit Bryce Harper second all series, which isn’t something he would have done in prior years. Under his watch this season, the Nationals shifted 758 times this year, up from the 216 times (second-fewest in MLB) they shifted under Matt Williams in 2015. Baker isn’t a perfect manager — no one is — but he’s not the scapegoat for the Nationals this year. The team had a successful bounce-back season and then played a really good division series to as close as a draw as you can get.
It’s easy to draw up a narrative around Baker’s postseason losses, especially since he doesn’t exactly endear himself to the analytics crowd with some of his comments, but that narrative is no more fact-based than the one about Clayton Kershaw not being able to live up to big pressure moments in October, and hopefully we watched the death of that silly story last night.
Baker’s not the modern-day archetype of manager bucking traditional wisdom, but he did just fine this year, and the Nationals didn’t lose last night’s game because of any decision that came out of the dugout.
Dave is the Managing Editor of FanGraphs.
I won’t blame a singular game on Dusty, but I absolutely will blame Strasburg being hurt on him, along with every other Nationals pitcher on the squad that has to deal with his abuse.
This is an antiquated myth.
As a Nationals fan, I have seen a manager abuse pitchers. That was Matt Williams, who frequently got relievers “hot” and then didn’t use them. By the end of his tenure, much of the bullpen was exhausted (Craig Stammen has pitched 4 MLB innings since 2014) not only from their actual game appearances but from Williams not realizing he was tiring guys out by warming them up multiple nights in a row, or warming them up multiple times in one night.
Is the myth about how he handled Mark Prior and Kerry Wood?
How was Strasburg abused? That sounds like nonsense.
Wrong…
Strasburg unravelled taking the ball both the Sunday before and the first game after the all-star break. He had been on the DL before that with a back issue, and the arm was supposedly good to go. Strasburg has always been a bit of a mental case, struggling most when the routine is disrupted (baserunners, rain delays, long innings)so the organization made the call to get him back on 5 days rest and he sat out the ASG. Unfortunately, there was an underlying elbow issue that hopefully resolves itself.
Dusty leaves every team he manages in shambles after a few years…every time. I suspect the Nats wont be any different.
The new even year mantra is in DC. 2012 2014 2016, one and dones. Extremely frustrating in a sports town that needs anything.
Hard to pinpoint blame on Dusty here. So many mistakes cost the Nats. Collective team failure. Three times not scoring the run from 3rd with less than two outs. Bob Sendley making the worst send home I’ve ever seen. And Espinosa being unable to lay down a sac bunt. Winners don’t do that in elimination one run games. Like Mike Rizzo’s mantra says. There’s always next year.
Following your logic, there’s always 2018.
Agreed with the three Ks in the three ABs with a runner on 3rd and less than two outs.
But my biggest culprit in this failure, is Scrabble’s 4-pitch walk. Nats traded future MVP Max Schrock for a 4-pitch walk in the deciding game of the NLDS. 🙂
I’m surprised there is no mention of the RE change of the Espinosa bunt decision, and/or its necessary break-even point
Regardless, agree with your overall point. Baker managed the game much better than his reputation would have you believe.
How can you say he wisely ignored Turners reverse split? Turner has been for a couple years now one of the best hitters against right handed pitchers and was been well below average against left handed pitchers. I don’t care what the pitchers stats against eighties are unless it’s somebody with a long track record of it you don’t ignore something so obvious especially when you already have a lefty in the game. And this is not hindsight 20-20 I said right as he went and got him that it wasn’t going to end well and what happened? A two run triple. I don’t usually fault managers but when I a fan knew that you have to ask how a guy that’s been in the game 50 years didn’t know that.
Because it takes something like 500 to 1000 PA’s for most offensive stats to stabilize. See http://www.fangraphs.com/library/principles/sample-size/
And, by the stats that stabilize more quickly – K% and BB% – Turner’s platoon splits against lefties look pretty good. The main thing that stands out is a lower BABIP against lefties than righties.
Don’t some stats like ISO or SLG stabilize much earlier? Turner has had over 700+ PAs in his career vs. LHP and, for whatever reason, at this point, he hasn’t exhibited the same power against lefties as he has against righties.
True, with men on 1st and 2nd, a BB or a single would’ve been a bad result for the Nats anyway–but the XBH hit really hurt. And, since the Nats brought in LHP Oliver Perez to retire the next batter in AGon anyway, maybe they could’ve just brought in the lefty first to face Turner and kept a quality reliever like Kelley for later in the game instead of burning him on one batter.
Not that preserving Kelley would’ve mattered much–can’t think of a spot later in the game where using him could’ve changed things in the Nats favor. But if you hold Turner to a single or walk and then retire Gonzalez, only 1 run scores and Heisey’s bomb probably makes it a 3-3 tie. Then having a quality reliever like Kelley to use in extra innings might matter.
The pool of major league hitters has a lot of variation in ISO (in “true talent”). It has not so much variation in platoon split amount, including ISO platoon split amount.
I blame everything bad that happens to his team – and especially his poor pitchers -on Dusty Baker.
I disagree. Dusty’s bullpen usage was poor independent of the decision to remove Max after the home run last night. Using Leverage Index, Rzepcyznski and Kelley were tied for highest pLI, 1.93, (Lefty and Righty) followed by Treinen, Solis, Perez, Melancon and Lopez for the series. So, he used his best reliever in very unimportant situations, his 2nd best lefty in more important situations than his best lefty and his best righty, also his best reliever, in less important situations than inferior righties. Looking at gmLI (when they enter the game), it paints the exact same picture.
Writing it in much simpler terms, Melancon didn’t throw an important pitch in games 4 or 5, games that were decided by the bullpens. That can’t happen.
This sort of hindsight analysis is trivial.
Okay, what am I missing here? I’m really asking, because the down- and up- votes for these two comments make me believe there must be something I’m not seeing. Since gmLI does indeed measure the leverage when pitchers enter the game (and not in situations they themselves create afterwards), and since the lesser relievers were brought in in more important situations than the better ones, how is it hindsight? Isn’t that kind of exactly how we should judge a manager’s bullpen usage?
There’s a bunch of people here who downvote comments based on something the author said in another comment section. Ignore the numbers.
Well I didn’t vote but one potential line of reasoning I see is that he was assuming they would maintain the lead and thus creating an even higher leverage situation later in the game. It would then make sense if you trust your guys. For me to believe that he used his relievers wrong I’d need more than five games to think it was more than just bad luck
Strongly suspect that what you’re missing is the extreme weakness of the correlation between up/downvoting and quality of comment here. The big brightly colored visible numbers just don’t correspond very well at all to who’s making a good point and who isn’t.
Agreed. Down votes and up votes have devolved into meaninglessness on this site. Write comments for yourself and for the edification of others. Look forward in hope of substantive responses. Ignore the green and red candy at the bottom of each comment. It’s a trap!
I didn’t see any red candy so… Oh, there it is! I found it! =)
Hah, fair enough. I tend to think there’s consensus when the numbers are extremely lopsided (they aren’t now, but they were when I posted above). It would seem in this instance I was fooled by an early small sample size.
“Given his dominance and the fact that he was throwing a shutout, asking him to try and get through the seventh was reasonable” is said after a very good summary of exactly why Scherzer should have never come back in after the 6th. I was stunned when I looked down to see him come out to warm up for the 7th. As you say, not great against LHH, even more so this year, even more so third time through, even more so after many pitches, and LAD had quality LHH after LHH, etc., etc.
So how could the quotes sentence come right after? They had nine outs left to get and had a one-run lead. Why in the world is the answer to say sure, take him out if he gets into trouble, but go ahead and risk that trouble despite all the evidence that says a) he is likely going to be mediocre at best as an option while b) a bullpen full of likely better options are sitting there waiting to get those nine outs.
It was absolutely a terrible decision, one that should squarely be blamed on Dusty. And that has nothing to do with his past, or the reputation he carried with him from it, anyway.
I’m confused. You spend a lot of time quoting 2016 and some career stats to justify certain decisions, rather than using projections and sabermetric principles, and then all of a sudden you get to Turner’s splits and despite his large career reverse splits you summarily dismiss them and say that it was a “wise decision to ignore them.”
So which is it? Should we use past performance to evaluate decisions or projections built on sound sabermetric principles? You can’t have it both ways and use whichever one suits your narrative otherwise you have no intellectual credibility.
For example you cited Scherzer’s own L/R splits to justify taking him out. Why does that not apply to Turner.
You also cited Scherzer’s own TTO penalty to also justify taking him out. What if he had had no personal TTO penalty? Could you then make the argument that he COULD have stayed in because HE would not have suffered a drop in talent the 3rd time through the order despite the league average drop (of around .35 runs per 9)?
These seem like excellent points.
Rekt
I agree with everything you said. Just ignore his stats? What? I’d understand it if it was just a one year anomaly, but that’s not the case. We’re nearing 800 PA’s for mr. Turner against lefties and he’s just a lot better against righties. Why are we ignoring this? Because it doesn’t make sense? Is that how this works?
Well, look at all RHB who have had much higher performance (wOBA or OPS or BA – whatever you want) in around 800 PA and then see what their splits look like in the future. Get back to us when you do that. Or I can give you the answer if you want…
Just a side note on the game. Urias’ pickoff move is an absolute thing of beauty. Was on the first base side at Nats park and had a perfect view to compare his leg lift on throws home vs. 1B. The speed of his leg left on his pickoff really helps disguise it. Contrary to some Nats commentators, absolutely not a balk. His gunslinger move is also quick and deceptive. Like the previous Fangraphs article opined, Urias is going to pick off a LOT of baserunners in his career.
Clear balk and I hope the umpires call it correctly in the future. Has no place on the diamond.
Open to correction (Rule 8.05), but his right foot did not pass the back edge of the pitcher’s rubber (objective judgment), and the umpire’s subjective judgment is that he stepped “directly” to 1B. Rich Hill was worse than Urias if you believe the subjective “directly” rule was violated. Also, his leg kicks were distinctly different between going to home vs. 1B.
Pretty much all lefties balk imo.
The runner thrown out at home to end the 6th inning was a very bad send. It looks even worse considering the Nationals started the 7th with a walk, HR and single.
The one thing that I would’ve definitely done different as Dusty is brink in Melancon sometime in the 7th. Bring in your best RP in a situation that is guaranteed to be high leverage.
I blame the fact they did not have a good plan on Dusty and not the guys in the pen being jerked around.
He made the correct call bringing in the guy that gave them best chance to get out of the inning with the lead intact first (Scherzer). The analysis you provided above is good justification to go get him at the first signs of struggle. But they did not go to the guy who gave them the best chance to get out of the inning. If they believe that they did (and Scrabble hasn’t been THAT bad against righties, just can’t strike them out), why not commit him to more than one batter? By panicking with each failed out, they gave the Dodgers the upper hand. Sticking with the warm arm would give them the best chance to get three outs.
Nobody wants to be seen as Grady Little, but the last couple nights (and playoffs in general) have seen the opposite extreme, where the decision that a cold arm out of the pen is better than the fresh warm one already in the game.
That said, Solis probably should have been the first out because he gave the Nats the best chance to grab 3 outs. The Nats may have still had to face Turner in that inning, but the revolving door and head games would not have started and both Kelley and Treinen would have been effectively warm and mentally prepared.
Any time 6 pitchers pitch in an inning, it is on the manager… unless they are the 2003 Tigers or something.
Dave this contrarian bent is not a good look for you. Dusty tried to tiptoe through the tulips with his multiple reliever changes and ruined his chances for better leverage later. He should have continued to trust Sherzer because he was having a great evening, clearly dialed in. Instead, death by bullpen.
Also, in the 9th paragraph you misspelled Rzepczynski.
Sherzer
I cannot agree that Baker made the correct call in having Kelley face Turner in the 7th inning. To say that the he went with the data is absurd. This was not a situation which needed an entire statistical breakdown of all reverse splits. There are reverse splits and there are REVERSE SPLITS! This was Justin Turner, he of the 209/303/630 line against LH’s and 305/356/868 line against RH’s in 2016. Even if Baker was bringing in Mariano himself, he was making a crucial mistake and he paid the price.
See my post below. And if you still don’t believe it look at all batters in history who have had similar reverse splits for 500-1000 PA and see what their splits look like in any future time period. When you do that, get back to us…
All batters in history is the fallacy of your argument. Just because some other batters had certain results in past confrontations means nothing. Statistics can be carried too far and in the bell curve of these pitcher-hitter situations, some banjo hitter with a sausage bat is not Justin Turner. He is very unusual in this trait but it must be respected. Baker ignored it and lost the game because of it.
The big problem with the arch-conservative attitude of “there can be nothing new under the sun” is that sometimes something new and different actually appears. I had an argument about this the other day with a Famous Author who was maintaining that exact viewpoint (“We have never seen this before, therefore it cannot exist”) and he was shocked to see proof that not all people behave the same way.
It would have been a mistake to bring in Mariano Rivera against Turner in that situation? That’s positively absurd and would have undermined your entire argument if you hadn’t already done so by your reliance on Turner’s supposed “REVERSE SPLITS!!!!!!!!!!!!!!!!”
I used Mariano only as an example that bringing in any right hander, even the best ever, to face Turner in that spot was a grievous error by Baker. What do you mean by supposed “Reverse Splits”? I guess you haven’t seen the truly remarkable reverse split that Turner has for his entire career.
What’s remarkable is that you’re still using history rather than projections as the basis for your argument. What matters is what happens next, not what happened yesterday.
I’m confused. I’m easily confused though. In another post you mention an example of gamblers fallacy then I see, “What matters most is what happens next, not what happened yesterday”.
In Blackjack, isn’t it most important to know what happened “yesterday”, (counting the cards of all the previous hands that have been exposed to keep the count) so you can bet accordingly today? (“next” hand)
Like usual, I probably didn’t word that right. Hopefully someone understands what I meant and can help me out explaining it better because, I thought the best way to predict future behavior is to look at the past behavior?
FWIW, my projection for Turner’s true platoon splits, which is what should be used to make any decisions about whom to bring in to face him, is 12 points of wOBA which is less than half of the average platoon split for a RHB.
There is no such thing as a RHB with a reverse platoon split (there may be an occasional batter like that but would have no idea who it is).
So basically Turner is expected to be a little better against a LHP.
But, those numbers apply to pitcher with normal splits. For example, if you bring in a RHP with large splits, then all of a sudden Turner acts like a normal RHB with normal RH platoon splits. Similarly, if you bring in a lefty with small splits, then he essentially is a RHB with little to no split. IOW, the split of the pitcher gets combined with the split of the batter.
Shawn Kelly is an extremely good RHP with a slightly less than normal platoon split (estimated). Against any batter, RH or LH, he is excellent. Against a RHB like Turner with a small platoon split he is equivalent to a league average closer or better, so that is certainly an excellent “matchup” for WAS. I put “matchup” in quotes because, as I said, Kelly is expected to be excellent against both RH and LH, albeit over .4 runs per 9 IP better v a normal RHB than LHB.
In order for a LHP to be a better matchup versus Turner, even with Turner’s small platoon split, that pitcher would have to be someone like Chapman or Miller, one of the best LHP in baseball.
Scherzer, the best we can tell, has a true platoon split of right around average for a RHP. A LHB will add around .2 run per 9 inning to his RA and a RH batter will subtract around .2 runs.
Given that he is overall around 75% of league average in preventing runs, a LHB makes that 80% and if you add in the third time through the order penalty, a LHB makes him around 89% in run prevention which is around equivalent to a league average set up reliever.
Just some numbers to play around with in terms of analyzing some of these decisions.
What is this small platoon split you keep bringing up? It is huge and has been very consistent throughout his career. To quote you “there may be an occasional batter like that” well, if you could look past your numbers you would see Justin Turner.
There is nothing “small” about Justin Turner’s reverse platoon split. It is over 60 points of wOBA for his entire career, and well over 100 on the season. There is small sample noise, and then there is Justin Turner. That said, the reverse platoon split is only a small part of the problem. The bigger issue is that in addition to extreme reverse platoon splits, Justin Turner has also been the best hitter in baseball against sliders for the last three full seasons.
So when you bring in an extreme fly ball RHP who is a two pitch reliever who throws his slider 45% of the time to face an extreme reverse platoon split RHH who is better than anyone in baseball at hitting sliders, I think you’re needlessly tempting fate.
To be clear, I wouldn’t have had a lefty face Turner there either, I’d have brought in Melancon. Melancon’s repertoire is diverse enough to neutralize Turner’s splits and that was the “save” situation right there.
In science, when something appears in real life that doesn’t fit our current thinking/hypothesis, we change it, because we are professional servants of truth. Statisticians, on the other hand, aren’t concerned with truth, and wave the existence of things that don’t fit their models away by using various methods of dismissal that would get a scientist reprimanded or fired.
Absolutely spot on. The reality that Turner is a radical outlier and destroys their little statistical world is impossible for a statistician to accept.
Kelley’s basically as good as Melancon.
How so, Dave?
Finally somebody sees past the numbers and at the reality of who Justin Turner is. Your comment on his ability to handle sliders had already been shown clearly when he battled Scherzer, as tough as there is, through the 13 pitch AB earlier. Turner knew what was coming and decided the game.
“There is no such thing as a RHB with a reverse platoon split” (there may be an occasional batter like that but would have no idea who it is).
Very, very rare (like you mentioned) but here are a few of the few with 1000+ PA: Adam Jones 95 wRC+ vs (L) & 112 vs (R), Dan Uggla 99 vs 112, Kevin Pillar 103 vs 111 and with 750+ PA: Justin Turner 96 vs 138, Mike Sweeney 105 & 123.
Turner is the most extreme example I saw in a limited amount of time. His 750+ PA was dismissed by Dave as not a large enough sample. Either Craig or Sean cited thinking that around 1000 PA was the threshold. I can’t see Turner making up a gap of -42 wRC+ in his next 250ish PA against (L) just to pull even with his same side wRC+.
He doesn’t need to pull even; what’s done is done. MGL is saying that the future expectation for Turner [i.e. his projection] should be that he will be marginally better vs lefties.
In betting terms he’s saying that he’d happily take even money that he [or another of your still-active examples] has normal platoon splits next year.
Why does he not have to pull even? He would have to pull even by the 1000 PA threshold to be able to say that he doesn’t have reverse-platoon splits right? It happens, it’s rare but it happens. I understand MLB as a whole hits better against opposite handed pitching. Not all players fit the same mold, there are outliers in everything. If I was betting on Turner, I’d wager he ends his career with a reverse-platoon split. Since 1000 PA is the stabilizing point, when he reaches that mark and is still well below the norm, is it safe to say he hits righties better than lefties? We can’t say that now though right? We have to wait for a sample of 250 PA to tell us what 750 already has.
* 96 vs 132 -36 wRC+ for Turner
I agree with the sentiment that it is possible for a right-handed hitter to have a true reverse platoon split; however, when both the majority of historic results and common sense indicate that the norm is for there to be some platoon split, you need to kick the tires pretty hard.
So how could Turner have such a large reverse split in his career-to-date numbers without having a true reverse-platoon split? Here are two possible culprits.
1. Batted ball luck: one of the primary sources of fluctuation of outcomes vs. skill is simple batted-ball luck. Versus righties, Turner has a career BABIP of .333 and a 12% HR/FB ratio. Against lefties, those values drop to .277 and 7%. However, his Hard% against LHB is higher (30.7% vs. 28.9%) while his Soft% vs. LHB is lower (13.7% vs 16.5%). It would appear that batted-ball luck has played a big part.
2. Lurking variables: we also have to keep in mind that Turner has gone from being a part-time player with middling offensive numbers in his pre-Dodger days to being a full-time player with very good hitting stats. What this means is that his career PA vs. RHB is split 68% “new” Justin Turner vs. 32% “old” Justin Turner, while is career PA vs. LHB are split 55/45 (most of this came from lots of platooning his last two years with the Mets). This difference would further skew his career stats to look more favorable from the right-hand side.
Thank you for your response. Possibly…?
1. BBL( you use career numbers): .277 & 7%: He has a 21.7 LD% vs lefties compared to 25.3% against righties(lowers BABIP). He hits more fly balls vs lefties(lowers BABIP). He hits more fly balls to the opposite field vs lefties(possibly, lowers BABIP and w/average power, lower HR%). And yes, luck. Baseball.
2. LV: I’m being critiqued for using a small sample size(773 PA) and you are telling me to use even less? I get he is performing a lot better with regular playing time and on the up vs lefties but all i can look at is what he has done every opportunity he has had vs a lefty, ever. I’m not even arguing that he is bad vs lefties though. He is about average(so far). He just happens to have performed much better against righties. I’m interested to see if the gap is large enough that, when he retires, the statistics will show if he hit better vs righties than lefties. I’d bet a 1$ on it. And yes, luck. Baseball.
On the second point, I’m not really asking you to use a smaller sample size, but to rule out possible bias coming through a hidden (lurking) variable, in this case usage. As an example of how this might skew results, consider a hypothetical RHB with roughly average platoon splits (.015 higher wOBA vs. LHB than RHB), and no fluctuations in batted-ball luck. For the first two years of his career his offense is lousy and he mostly platoons, with a .225 wOBA in 200 PA vs RHB and a .240 wOBA in 200 PA vs LHB. The same guy then magically figure out how to hit over winter ball, wins an everyday job, and hits for a .375 wOBA vs. RHB in 400 PA and a .390 wOBA vs. LHB over 200 PA. So what do his career splits look like:
Against RHB: .225 wOBA in 200 PA and .375 wOBA in 400 PA = .325 wOBA in 600 PA
Against LHB: .240 wOBA in 200 PA and .390 wOBA in 200 PA = .315 wOBA in 400 PA
Huh? This looks like a reverse-platoon split, when really it’s a by-product getting most of his at-bats against RHB after he improved as a hitter. Splitting his PA into smaller subsets doesn’t increase the variability of the result (since we’re still trying to explain his career splits) but exposes a possible bias that’s independent of random variation.
Now, having said all this, we could take things to the next step and try to dive into how the guy (or Justin Turner for that matter) made his big improvement in the first place. For example, if Turner suddenly learned how to pick up the “dot” on a slider when he joined the Dodgers, that would help him more vs. righties than lefties, but would obviously be a huge help for both. But again, to drill down into this potential deeper-lurking variable, you’d have to subdivide the data further (Turner against sliders vs. non-sliders as a subset of each of the four subsets before).
Bottom line is, it’s hard to either prove or debunk Turner’s possible reverse split. All we can do is note that it looks like a huge outlier that would need further investigation.
I get that the gap vs (L) is not as large as the numbers suggest. He is a better player today than he was pre Dodger days. I wonder if he dug himself into too much of a hole though from his early years to be able to surpass his career split vs (R). That’s all this is about to me, will he join that very, very, rare group of RHH that finish with a reverse-split?
Say he doesn’t completely close the gap but hits better vs (L) than (R) going forward. Do we say he was better vs lefties or do we say he was better vs righties?
I say all of this with the upmost respect towards you. Thank you very much for not attacking me because we don’t quite see eye to eye. That is becoming uncommon here.
I appreciate the kind words. I wonder why folks jump too quickly down each others throats on here sometimes. No problem when you call someone out for being rude or for obvious trolling, but other than that what’s so wrong with agreeing to disagree. After all, greater minds than ours have been way wrong on many of these things. Why should we pretend to have a better grasp of things?
The concept of “pulling even” is a classic example of the gambler’s fallacy.
I was stuck on this 1000 PA thing. I think Turner hits righties better than lefties. I thought, I was being told I could not say that until he had at least 1000 PA against lefties. I looked at the gap (-36 wRC+) between how well he hit righties and how poor he hit lefties. That’s what I meant by pulling even by 1000 PA. This line of thinking is probably wrong, but it’s honestly what I meant.
The answer, as always, is Mike Trout.
According to BBRef, after 3558 PA, Trout is 310/405/567 against RH pitching, and 293/405/529 against LH pitching. I have no idea if it is statistically significant, and of course he’s seen a lot more right handers than left handers.
Trout being Trout, if he saw nothing but left handed pitching, he would probably soon be feasting on them.
I think it’s because all but the best lefties either call in sick or have a drone accident when they know they have to face Trout.
The one I question was the double-switch which took Rendon out of the game.
Before and after Scherzer giving up a bomb to Pederson, how much would you change your expectation of his performance? Either he shouldn’t have started the inning, or he should have stayed in, you could argue, but there’s not much window to argue the middle.
Jetsy, I felt the same. You can’t look for “signs of trouble” in Scherzer and hope to head off the damage — the first sign is likely to be a homer. In the last 2 years, 56% of runs against him came from homers. (What did we see in Game 1?)
Putting in Zep to face Grandal (after Pederson’s HR) was either a batter too late, a couple too early, or just the wrong guy entirely. Pederson was the inflection point of that inning — the guy most likely to hurt Scherzer, and most likely to fail (or get pulled) against a lefty. By going to Zep right after the homer, to face the switchy Grandal, all Dusty gained was a fresh arm. The ensuing walk was no surprise, given Zep’s 15% massive walk rate to righties.
And the Roberts move to Kendrick for Toles triggered Dusty pulling Zep after one batter, a righty. That was a small thing in the whole panoply of what happened, but it was so pointless and predictable.
I thought maximizing his LHRPs was the biggest key and challenge for Dusty this series, along with daring to pull his ace at a point that would expose him to carping. And he wasn’t up to the task. He didn’t lose this game — hey Trea, how about trying a Granderson face-plant next time you’re in a do-or-die spot? — but he did get outmaneuvered.
Just go back and look at the bottom of the 3rd in Game 4 and Game 5.
Dusty Baker has a history of abusing bullpens. The Giants, the Cubs, and the Reds all sent him along down the line partially because of this mishandling. This is not a myth of any sort.
I am not a Baker fan. But I have to say that Baker in ’16 is not the same manager as the Baker of the past. He did a pretty good job given the circumstances of the ’16 Nationals. The Nats are who they are and did the best they could. Unfortunately, they lost.
The bigger question? Maintain this team for another year or start rebuilding for 2 or 3 years using their great minor league farm system.
I actually questioned Dusty for leaving Scherzer in too long, and even that was more hindsight than anything, although there were signs he was tiring. No way he should have been left in IMO. But whenever you lose, hindsight allows one to second guess any decision. Pulling Scherzer didnt work. I would question his going to so many RP’ers in the 7th inning though. Maybe this matchup thing is overdone. Let a guy face more than 1-2 batters if he has proven himself over the course of a season. But then, maybe the outcome is even worse. But it didnt work for Bochy and it didn’t work for Baker
Roberts actually would have had more explaining to do if things did not go right. I thought he was a bit quick to pull Hill, which led to Jansen being gassed and unable to close it out and having to bring in Kershaw on 1 days rest after starting on short rest. Assuming neither Kershaw or Jansen suffer any ill effects then the outcome makes him look like a genius. But bad decisions don’t always lead to bad outcomes, and good decisions can lead to bad outcomes. Thats baseball
Even without the reverse split concern regarding Turner vs Kelley there is a lot to dislike about that matchup for the Nats. Kelley is a strictly fastball/slider (55%/45%) reliever with EXTREME fly ball tendencies and Turner is the best hitter in the majors against sliders over the last three seasons (based on pitch value) and a fly ball hitter with an extreme uppercut swing. That alone makes the move questionable to me.
I also don’t think Turner’s reverse splits can be ignored. We aren’t talking ~20 points of wOBA, we’re talking 60 points over his entire career and over 100 points this season. There is more than just random stat noise there.
My theory is that Turner’s extreme leg kick and his uppercut swing path gives him a lot less room for error vs LHP working him hard in and soft away. He has to cover the entire plate vs LHP. Vs RHP, he splits the plate in half knowing he has the bat speed to get to hard stuff anywhere and he knows soft stuff starting in will typically have glove side run right into the meaty part of his swing and soft stuff starting away will be a ball or an unhittable strike. I think his heatmaps would back up this theory, but it could just be small sample.
All in all, it wasn’t a terrible move, but I think the better move would have been Melancon who has a more diverse repertoire. At that point, the likelihood of having a save situation was almost zero anyway. The “save” was the situation Kelley got thrown into. The only really questionable move was the reactive move of pulling Scherzer after the homer. I’m no Dusty fan but I’m not sure changing any of the moves really changes the outcome.
This is a well written article. I am no fan of Dusty; but his pitching moves all made sense.
Dusty Baker may be plain unlucky, but he’s proven that you can’t win with him.
Pinch hitting for Zimmerman and Rendon were dumb moves, and bringing Scherzer out for the 7th wasn’t too bright either.
That said, had Murphy taken his normal swing against Kershaw instead of trying to take him deep, the game would’ve been tied
Strasburg was brittle long before Johnny B came to Washington.
Dusty had a solid first year,nats fans should look forward to 2017.
Regarding Dusty’s decision to leave Scherzer in to start the 7th, am I being naive when I also consider Joc Pederson’s proclivity for hitting home runs? I don’t know his stats well, but IIRC he had very poor BA and OBP, and lots of strikeouts, but if he made contact, the ball is leaving the park. That’s been his MO since being in the majors.
When you consider that in conjunction with the home run being Scherzer’s bugaboo, Scherzer was nearly 100 pitches in, third-time-through, on a night when he clearly did not have his best stuff (dominating is a very generous term to describe his performance, especially if you watched every pitch and the length of the ABs), and last but not least, the fact that it’s a one run game so there is literally zero room for error, I felt the decision was unjustifiable. Especially when you had Solis warmed the previous inning, you have a stream of lefties coming up, and Solis’ splits versus R aren’t exactly ghastly, you could stick with him for a righty or two if you had to. It should have been a matter of get through the seventh however you have to, burn everyone you need to, and then go to Melancon for six outs.
But back to my original question. I’m no statistician nor do I know Joc’s stats particularly well. Am I wrong to give him credit as a serious homer threat considering the rest of the circumstances? I know he’s also a strikeout threat, and Max is a strikeout guy, but it didn’t look like a good matchup from my view in left-center field.