The World Series of Power Versus Finesse
Only three teams threw the ball faster, on average, than the Royals this year. Not surprising when you’ve got youth like Yordano Ventura, Greg Holland and Kelvin Herrera throwing fire on the regular.
Only one team threw the ball slower, on average, than the Giants this year. Not surprising when you have distinguished gentlemen like Tim Hudson, Ryan Vogelsong, and Jake Peavy stepping on the rubber three out of every five games.
This difference in velocities has ramifications for pitch mix, of course. The Royals threw fastballs more often than the Giants. The Giants threw breaking pitches more often than the Royals. In fact, the Giants threw more breaking pitches than anyone in baseball.
Is one team better equipped to handle the strength of the opposing team?
We have pitch type values on our leaderboards, and they provide a great way to look at this question. Though you can run into some strange places when you use the values to judge pitchers, some of those limitations go out the window when you use them to judge hitters.
For example, R.A. Dickey has sometimes shown a plus fastball by pitch type values. Over his Mets career, it was regularly a top pitch, and in 2011, you could have used the number to say he had a top-15 fastball. That year, he got 2.9% whiffs on his fastball, so it probably wasn’t a top-15 pitch.
He did get an above-average 45.6% ground balls on his fastball that year, and that points to another flaw with the pitch type values. They measure what happens on that particular pitch and can be subject to the luck of the bouncing ball. He had a .278 batting average on balls in play on the fastball in 2011, a year after that number was .405 on the same pitch.
If you use these to judge a hitter, you do rid yourself of the first problem. Unless every pitcher is sequencing the slider in the same way to the hitter, you can probably get a good sense of a hitter’s ability to hit a particular pitch from his pitch type values. And if you take a bigger sample than one year, you might be able to mitigate the second flaw by gathering a large amount of balls in play.
So let’s not point too heavily at this leaderboard that suggests that the Royals can’t hit curveballs, are below average against sliders, and that the Giants can handle fastballs. That might seem like a large sample until you realize it’s just a collection of small samples, made murkier by how the balls bounced for each player in particular.
Let’s instead look at the respective lineups and see how they’ve fared against fastballs and breakers over the last three years. Now we’re increasing the balls in play sample and reducing the effect of sequencing. By using the ‘per 100’ values (/c), we’ll also reduce the effect of pitch and plate appearance volume on the outcome.
| Royals | wFB/c | wSL/c | wCB/c |
|---|---|---|---|
| Alcides Escobar | 0.04 | -1.88 | 0.38 |
| Nori Aoki | 0.45 | 1.61 | 1.89 |
| Lorenzo Cain | 0.40 | -0.18 | -0.83 |
| Eric Hosmer | 0.51 | -1.14 | 0.00 |
| Billy Butler | 1.27 | -0.96 | -0.58 |
| Alex Gordon | 1.05 | 0.19 | -0.70 |
| Salvador Perez | 0.25 | 0.01 | -0.87 |
| Omar Infante | -1.04 | 0.50 | -0.61 |
| Mike Moustakas | -0.53 | -0.76 | 0.35 |
| Average | 0.27 | -0.29 | -0.11 |
As a team, the Royals like fastballs. That’s too bad, considering the Giants’ reluctance to throw fastballs. As a team, the Royals hitters have not faired well against breaking pitches over the last three years. Once again, not a great fit considering their opponent. Maybe we’ll see some hits from Nori Aoki, and Alex Gordon (sixth-spot hitter) could pair some ability with the slider with more power.
But that’s two guys. The rest of the team is average or worse against breaking pitches. And, with a minimum of 1000 plate appearances, you’ll see that Cain, Perez, and Gordon are bottom-30 against the curveball. Escobar and Hosmer are bottom 20 against the slider over the past three years. They might have a tough time with guys like Madison Bumgarner (35% sliders), Tim Hudson (23% sliders), Ryan Vogelsong (19% curves), and Sergio Romo (52% sliders).
The Giants’ hitters have done better in general, but *how* they’ve done so is also important.
| Giants | wFB/c | wSL/c | wCB/c |
|---|---|---|---|
| Gregor Blanco | -0.05 | 1.31 | -0.25 |
| Joe Panik | 0.36 | 0.58 | 1.75 |
| Buster Posey | 1.47 | 0.37 | -0.03 |
| Pablo Sandoval | 0.56 | -0.57 | 0.98 |
| Hunter Pence | 0.38 | 1.15 | 0.81 |
| Brandon Belt | 1.12 | 0.11 | -0.21 |
| Travis Ishikawa | -0.16 | -1.13 | 1.17 |
| Brandon Crawford | 0.46 | -1.71 | -0.27 |
| Michael Morse | 1.04 | -0.79 | 0.04 |
| Average | 0.58 | -0.08 | 0.44 |
There are some hitters here that are vulnerable to breaking pitches. Crawford has problems with sliders, apparently, as do Sandoval and Ishikawa. The Giants are great against curveballs as a group, and that’s semi-relevant given the fact that only six teams threw more curves than the Royals. Panik, Sandoval, and Pence are ready for the yakkers.
But what the Royals do differently than anyone else is throw fastballs. And it does look like the Giants are well-suited to take that strength on. Posey was 19th-best against the fastball this year, even, and Crawford also found himself in the top 40 in the cumulative version of this statistic. There are some great fastball hitters on this list, and then there are some good ones.
By the pitch type values, at least, it looks like the Giants’ hitters are equipped to handle the fastballs of the Royals’ staff. And the Royals’ hitters may have trouble with the Giants’ breaking balls. These are general statements, hopefully made more believable by focusing on a larger sample. How Posey will do with 98 mph cheese from Ventura in particular is why we watch.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
nice article. The other thing that the Giants have is a handedness match. 3 out of 4 starters, and all the core relievers of the Royals, are righties. The G have a heavily left-handed line-up (6 out of 9, with S being classified according to split). The Royals have 3 out of 9 lefties.
The only thing I don’t like is Petit not starting and MadBum starting a game at home. His large reverse H/R splits are real. I would have preferred Petit starting Games 1 and 4, and MadBum Games 2 and 5.
Interesting. And you’d do that even at the further exclusion of Madison pitching in a potential game 7? Not saying you’re wrong, just wondering if that went in to your preferences at all.
I think that Bochy goes with four pitchers. Not sure of Madison’s effectiveness on three days of rest after a 250IP season. Right now Game 7 belongs to the all too hittable Hudson.
I’ve thought about this too (the strategy of pitching Bumgarner 1-5 vs. 2-6), but my guess is having Bumgarner as a relief option in Game 7 if he pitches 1-5 is probably more significant (especially since he’s left-handed and the Royals lineup has a lot of lefties).
I don’t necessarily believe Bumgarner is better away than he is at home (small sample size this year, he was better at home last year, and for his career, it’s very close), but there’s potentially an argument to be made that he doesn’t really benefit from AT&T as much as other pitchers do. He pitches primarily to his glove side (as the Fox announcers keep reminding us), and AT&T’s large right-center isn’t much of an advantage for him (in essence, it’s a comparative disadvantage).
Petit over Bumgarner to start game 1? Wow. I can’t even… Wow.
Thanks for your suggestion, Glib. Think I’ll stick w/Bochy on this one.
Game five is still a home game, and none of the Giant starters will be throwing on three days rest.
I’m pretty sure Glib meant 1 and 5 vs. 2 and 6.
“The only thing I don’t like …”
That’s actually two things — unless we’re using Michael Palin’s Spanish Inquisition Count System. But, of course, then we’d have to add AN ALMOST FANATICAL DEVOTION to keeping tiny Tim on the roster for no clearly articulated or plainly apparent rationale … or superstition, or sentiment, or maybe he thinks the Royals might possibly go into meltdown out of FEAR, or astonishment (or paroxysms of laughter — that’s 4 things just in the subset), BUT ANYWAY …
Note:
1. Only Bumgarner will start on his normal rest of 4* days, presumably twice, and he’s a horse: more durable than any other starter actual or potential on either team.
2. Against the Nats, Vogelsang got into the mid-90s on extended rest, and has 9* days before his only start in Game 4.
3. Also against the Nats, Hudson was as effective against at any time all season on extended rest, and he too has 9* days before his (first, and possibly only) start in Game 3.
4. Even Peavy gets 9* days before his first start in Game 2, and 5* days between that and his second if he gets one — and if there’s a Game 6 and someone other than these 4 starts it, then 6 days between his first and second starts.
The only question is about 1 game, if it happens, being Game 6 (or quite a bit less likely, Game 7). If Petit isn’t needed in any of Games 1-5, it’ll only be because the Giants have already wrapped up the Series win before Game 6. If Petit is used in one of Games 1-5, it’ll likely be a long relief outing – which makes a start for him in Game 6 or Game 7 highly unlikely. Same result if Petit is used more than once in shorter relief.
So it comes down to this: if the Series rolls out to 7 games, will Bochy see Peavy on 5* days rest and Hudson on 4*, as preferable to Lincecum (yes, Lincecum) in Game 6 and Peavy on 6* days rest in Game 7?
[*The Giants have been in 3 World Series’ since the earthquake in the one against Oakland. And AFAIK they still play their home games in SF. I’m just sayin’: from Game 1 on, all calculations of rest before next starts are necessarily tentative.]
Consider that, historically, Bochy’s decision-making has been tinged with a certain amount of hunch-ridden sentiment. Not entirely, of course: for example, he hasn’t called on Lincecum since the last series of the regular season — in late relief in a pretty-much nothing game against the Padres where there was neither big risk and down side.
However both Bochy’s sentiment and his huncheology lies heaviest with 2 pitchers: Peavy and Lincecum.
We don’t know the answer, and may never know. But … if Hudson gets hit hard in Game 3, and there’s a Game 6, there’s a non-trivial chance of Lincecum starting it.
Indeed, given the probabilities, including that the scenario of getting to Game 6 without Petit being called on is so unlikely, it’s more probable that Lincecum starts Game 6 than Petit, and also more probable though somewhat less so than Peavy (though I grant that Lincecum starting Game 6 remains less probable than the combined chances of it being one or the other of Peavy or Petit).
That’s at least 3 things not to like, in the sense of keeping one’s evening meal from repeating while watching.
Last year Bumgarner had better splits at home than on the road and in 2012 they were MUCH better. It’s more likely his home struggles this season are a result of sample size. Advanced stats and splits are wonderful, but don’t overthink the obvious. Going with Petit in Game 1 over Bumgarner is overthinking the obvious.
That was the Orioles problem, they had mostly RHB’ers and the only LHB’er with any power was off the roster. I think the Giants will fare much better and expect the Royals bats to go back to where it usually is, which is not very good. They will still play great defense and run, but it probably won’t be enough. Gianst have been there before and the Royals have not, and that’s a big advantage as first timers tend to try and do too much
Because a particular hitter or group of hitters is good or bad against fastballs, it does not necessarily follow that he is good or bad against a pitcher or group of pitchers that throw predominantly fastballs. At least I have never seen any research that indicates that we can draw this kind of conclusion. The reason it may not hold is two-fold:
One, if a batter is particularly good at hitting fastballs, ALL pitchers, whether they be predominantly fastball or not, will throw fewer fastballs. Two, if a batter is good versus fastballs in general it does not necessarily mean that he is good against the fastballs that predominantly fastball pitchers throw, which are generally above-average fastballs.
More importantly, I am skeptical that what we see when we see that individual hitters are better at fastballs or off speed is anything more than random fluctuation. Here’s why.
If we know that a batter prefers fastballs, we will throw him fewer and fewer until he is not so good at them anymore. In fact, game theory tells us that the value of every type of pitch to every batter should be the same, given the same count,score, badecrunners, etc. Otherwise we would simply stop throwing that pitch altogether.
Now because of those “individual and unique situations” (count, score, etc.) we cannot guarantee that the value of all pitches to any particular batter should be the same, but it should be pretty close.
So, I am not buying this kind of analysis. Never did.
While I strongly agree with the game-theoretic concept for pitching, I am doubtful that in many cases that “we will throw him fewer and fewer until he is not so good at them anymore.” In some cases, we might just not throw them at all, unless they are our best (or only) pitch.
For example, with Bonds you just wouldn’t throw certain pitches. If those were your only good pitches? You’d walk him instead, but it’s not like he got worse at hitting those balls. Quite the opposite: only pitchers who were very good at those pitches even attempted them.
Since not all pitchers have enough plus pitches to hit certain weaknesses, sometimes they will still need to play strength-on-strength. In those cases, you’d get some signal: the batter would still hit his favorite pitch well because he could expect it coming, despite the fact that it is his favorite.
I mean, otherwise the same argument could be made that batters shouldn’t hit better in different parts of the strike zone. After all, if they hit well in those parts, the pitcher would just avoid them. But, much like placing pitches, some pitchers just can’t “avoid” certain pitch offerings without being entirely predictable. In that case, you throw the hittable pitch only enough to prevent complete predictability (despite the fact that the batter is more likely to generate runs on that pitch, on average, it is an investment in maintaining the value of your other pitch).
But we’d still expect this effect to be small for viable major leaguers. After all, guys like Wily Mo Pena who never learn to lay off breaking balls don’t tend to have very long careers.
“I mean, otherwise the same argument could be made that batters shouldn’t hit better in different parts of the strike zone.”
That is exactly 100% correct. The reason you don’t see that is two-fold:
One, you have to use the right metric, and that MUST include balls, strikes, and HPB in that zone.
Two, the Nash equilibrium only applies to the exact same situation for each batter. If you just look at all pitches in each area across all situations, you will NOT find that they all have equal value.
Three, desired location and actual location are two different things. Nash equilibrium only applies to intended location. To do a proper analysis, you would somehow have to infer intended location, which is not so easy.
And of course since Nash equilibrium only applies, as I said, to each batter and each unique count, score, inning, runners, etc., the sample size for each of those unique situations would be so small that there would be so much random fluctuation in the pitch values that it would be very difficult to conclude that the true values were actually the same.
In addition, during the last weeks of the regular season the Giants faced several pitchers who threw NOTHING BUT fastballs, relying almost entirely on location and making strikes, and didn’t fare at all well. In particular, two of the three games they lost to the Padres in San Diego in the second last series they had with them this season, squashing their hopes of taking a late run at the Dodgers, were to pretty much 100% fastball pitchers.
I’m not a statistician, and even if I was I wouldn’t be as good as MGL, so I am wondering if he could answer my wondering about two parts of this idea.
“If we know that a batter prefers fastballs, we will throw him fewer and fewer until he is not so good at them anymore. In fact, game theory tells us that the value of every type of pitch to every batter should be the same, given the same count,score, badecrunners, etc. Otherwise we would simply stop throwing that pitch altogether.”
Doesn’t this not take into account pitch mix? For example, lets say you have a pitcher with a fastball and a slider. He is facing a batter who is bad against sliders and good against fastballs. For the sake of ease, we will say nobody is on and nobody is out. We will also assume that Mr. Pitcherson has only those two pitches, maybe he is a reliever or something. At the base idea, you would never throw the fastball, right? But if the hitter knows this, presumably because he never gets thrown fastballs, then he will just sit slider, and presumably he would then get better results against the slider, while he might do not as good against the fastball (As he is gearing his swing for a slower pitch). So you would throw a fastball at times to help counteract this, right? Which might explain why, for example, people have not stopped throwing that pitch altogether.
In addition, doesn’t this ignore the skill of the pitcher, especially control? For example, if the pitcher has really bad control of his slider (Maybe that’s why he is a reliever and not a starter) but good fastball control, then he might still throw the fastball, especially in 2-0, 3-1 and 3-0 counts, because he can get the fastball in the zone better, which would show why we didn’t stop throwing that pitch. For example in a 3-0 count, as you mention in the next bit about situations, the fastball should be worth a lot more and the slider a lot less due to the slider being much more likely to cause a walk. Couldn’t that be a big difference.
Finally…could it be that the only reason we haven’t stopped throwing the pitch is that people haven’t figured out we should do that? Not saying that is true or even likely, but that could be an explanation.
Also, why would game theory say every pitch is the same? Shouldn’t it change based on how good the pitcher is? Plus, I could be wrong, but don’t we actually see pitchers throw fewer pitches to players who are good at specific pitchers, even if it isn’t to the extent of never throwing it? Presumably due to players very good throwing (For example, Aroldis Chapman will always use his fastball), situations or even just pitchers making the incorrect call or due to pitch sequencing. So we would see a drop off but not necessarily all the way.
Eheheh…okay, I talked too much and this is probably way too long for you to answer as I imagine you’re a busy person MGL, but I found this idea fascinating so I threw a lot of probably stupid questions out. If you take the time to read this, then thank you. 🙂
Ever heard of Finnegan glib? No BUENO? I thought this article was going to be about the Royals’ small ball and speed opposing some decent power from the Giants. You talk about off-speed stuff being the main contributor, I however, prefer looking at something more simple and meaningful. H2H. Royals in 6.
sorry, I do not follow other teams as much as I should. But I can see that Finnegan is not pitching in the postseason.
LH pitcher Brandon Finnegan not on the Royals’ roster for this World Series? You’re wrong on that.
lord I have to proofread better . is not pitching MUCH. 4.1 out of 80 IP, ERA 4.15. I just don’t see handedness as anything but a Royals disadvantage. Affeldt-Lopez 8.1 IP, ERA 0.00.
Aoki and Panik (both batting 2nd) devour curveballs. Interesting.
So what does KC do against the curve? Try and take even more pitches? Move up in the batters box to try and get to the curves early?
Cool to see in numbers what we’ve seen for the last few years as Royals fans: a lot of them hack at breaking balls, often ones in the dirt or the opposite batter’s box.
I have a hard time believing Sal Perez is an average slider hitter. I know he can poke them out to left (see WC game walk off), but recently, they are so far out of the strike zone that they’re nearly impossible to hit, and he’s still swinging at them. I would be interested in his numbers from this year compared to previous, as it’s apparent to the eye that his discipline and pitch recognition have deteriorated.
I think the Royals will need to keep their walk rate, which has jumped dramatically in the postseason, at that much-improved level if they want to score more than 2 runs a game against the Giants pitching/defense.
I’ll keep it short. Giants in 4 or 5 games. The Royals bullpen will only get one or two save opportunities, and they will blow one of them.
The Royals ( and perhaps everyone else ) ought to treat Panda like Bonds in ’02. It’s their only hope.
Just a great article. This sort of analysis leads to so many interesting discussions.
My first post on fangraphs. Awesome.
While I think these stats give insight into how a batter fares against a pitch in general terms, I really don’t find it useful. As a softball pitcher in college, I’d like to share what I’ve learned playing the game. I think it’s important to recognize a few factors. 1) The pitcher’s style… No two fastballs are alike! 2) Mistakes and how the pitch was thrown. Some breaking balls end up flat and thus hittable. 3) Location, location. If it’s thrown over the plate, you can say goodbye to the ball. I think location is actually more important than type of pitch. You can jam a batter inside or make them go after a pitch outside. 3) Game-situation and the count. If behind the count 3-0, the pitcher is going to throw a strike with their best pitch and that becomes predictable for the hitter to get a good bat on the ball. The Royals seem to have matured into an aggressive hitting team with attacking first strikes. With the exception of Giants’ Pence and Sandoval.
What I think would be more productive in this case is to compare a batter against the pitcher, not a general pitch thrown by a variety of pitchers in MLB. i.e. Butler is 3-2 against MadBum and Infante is 10-4. Pretty good numbers. Meanwhile, Shields pitched a shutout game against the Giants earlier in the season. We’ll see if that’s any indication for what will happen in the opening World Series game tomorrow. Should be fun!
Intresting..I think the Giants will fare much better and expect the Royals bats to go back to where it usually is, which is not very good.