Henderson Alvarez and Being Worse Than a Coin
You don’t know how long I’ve been waiting for this to happen. Sure, probably nobody else noticed that it happened, and Henderson Alvarez himself probably didn’t notice that it happened, but, let’s not be judgmental. You have your things, and I have mine. This is one of my things.
Let’s break hitting down to a level that’s so simple it barely even applies to real hitting in the first place. As a hitter, you want to hit the ball hard. To hit the ball hard, you want to maximize your swings at hittable pitches, and minimize your swings at less hittable pitches. The strike zone mostly captures the hittable pitches, and a pitch taken in the strike zone will count against you for a strike. So to make things excessively simple: you want hitters to swing at strikes, and you don’t want hitters to swing at balls. Generally, a swing at a strike is a good decision. Generally, a swing at a ball is a bad decision. The most disciplined hitters in baseball will swing at a lot more strikes than balls.
Conveniently, we can establish a discipline baseline. What’s the worst discipline one might ever expect? That would be an even blend of swings at strikes and swings at balls. That would suggest zero discipline at all, and that’s the approach we’d expect if swings were determined by flipping a coin. If everything were 50/50, a hitter would have an O-Swing% of 50% and a Z-Swing% of 50%. To somehow perform worse than that would hint at the existence of anti-discipline, which I don’t even know how I would explain.
Henderson Alvarez is best known for being a baseball pitcher. Because his employer’s in the National League, he also sometimes has to be a baseball hitter. This past season he came to the plate 67 times. Henderson Alvarez’s discipline was worse than a coin’s.
We’ve got PITCHf/x data stretching back to 2008. That’s not that long ago, but it at least gives us seven solid years of information. Over that era, there have been nearly 4,000 player-seasons of at least 50 plate appearances, 2014 Henderson Alvarez included. So let’s talk about that. Alvarez just posted the highest O-Swing% on record, and he beat second place by a full five percentage points. He beat second place in just this season by nearly ten percentage points. Yet Alvarez’s Z-Swing% — his rate of swings at pitches within the strike zone — places in the lowest 20th percentile. Against strikes, Alvarez was unusually passive. Against balls, Alvarez was unusually aggressive. When you put those things together, magic happens. Bad actual hitting happens, but statistical magic happens.
This is a table. I’ve created a statistic which is simply Z-Swing% – O-Swing%. Here are the ten lowest-ranking player-seasons, since 2008:
| Rank | Season | Player | Z-Swing% – O-Swing% |
|---|---|---|---|
| 1 | 2014 | Henderson Alvarez | -9.3% |
| 2 | 2012 | Alberto Gonzalez | 2.7% |
| 3 | 2012 | Lucas Harrell | 3.5% |
| 4 | 2014 | Bartolo Colon | 4.2% |
| 5 | 2009 | Jorge de la Rosa | 5.2% |
| 6 | 2009 | Jon Garland | 5.9% |
| 7 | 2014 | Brett Hayes | 7.2% |
| 8 | 2009 | Eliezer Alfonzo | 7.4% |
| 9 | 2009 | Paul Maholm | 7.9% |
| 10 | 2012 | R.A. Dickey | 8.0% |
Alvarez is the first to pull it off. He’s the first player to have a decision-making process that underachieved a literally random decision-making process. And Alvarez didn’t simply finish in the negatives — he had a difference of more than nine percentage points. This is a man with an O-Swing% of 66%, and a Z-Swing% of 57%. Alvarez offered at balls twice as often as Dayan Viciedo. He offered at fewer strikes than Christian Yelich. In retrospect, the Marlins would’ve been better off making Alvarez’s decisions with a quarter in the dugout.
Courtesy of Baseball Savant, here is Alvarez’s 2014 swing/no-swing map, with his strike zone represented by a box:
There are swings close to the zone, and there are swings not close to the zone. For the record, I’m folding bunt attempts in with swings, because you can always pull back on those. There’s no reason to try to bunt at a ball. According to Brooks Baseball, Alvarez faced 60 pitches away off the plate, and he swung at 40 of them. He swung at 12 of 29 pitches in the zone, over the middle third. Alvarez this year exhibited anti-discipline. That’s inconceivable to me, but it happened.
What are some representative Alvarez plate appearances? We can go back to April, when he batted twice in a game against the Nationals. Plate appearance the first:
Plate appearance the second:
Here’s how that second one ended up:
Henderson Alvarez struck out, then reached, then was called out on batter interference. It’s not quite Hunter Pence hitting the baseball three times with one swing, but this was one of the weirder individual plays all year long, and it’s fitting it happened at the beginning of one of the weirder batting seasons on record. No matter how you choose to analyze him, whether as a hitter or as a pitcher, Henderson Alvarez has a lot of weird going on.
What is Alvarez as a hitter? This season, he tied for fourth in strikeouts on missed or foul bunts, with four. That’s bad, and the discipline also suggests bad, but then Alvarez batted .193, which is good enough for a pitcher. Last season, by some miracle, Alvarez batted .300 and slugged .500, and the man has a swing that isn’t embarrassing:
Alvarez hit a home run there, against the Cubs in 2013, and here are a couple quotes:
“As soon as I hit it, I knew it was gone,” Alvarez said through translator and Marlins third-base coach Joe Espada. “I was up there looking for a good pitch I could hit and I got it and put it in play.”
…
“It was a cutter, [Wood] was trying to back-door a cutter and pulled it across the middle of the plate,” said Cubs manager Dale Sveum. “That guy can swing the bat as good as any pitcher in the game.”
“That guy can swing the bat as good as any pitcher in the game.” For all I know, that’s not incorrect. Alvarez does have a good swing and he hit that ball against the Cubs nearly 104 miles per hour, which is substantial. The issue is that hitting has two parts — you need to have a good swing, and you need to know when and where to apply it. Used to be Mariners fans would hear every month about Greg Dobbs‘ picture-perfect swing mechanics. Apparently he played baseball this season, but he’s 36 years old, and he’s sitting on a career wRC+ that isn’t much higher than double that.
It’s not simply that Alvarez likes his swing so much he uses it all the time. If that were the case, he’d have a much higher Z-Swing%. He’s somehow simultaneously over-aggressive and over-patient, or at least he was this season, and this season’s numbers aren’t getting updated anymore. Alvarez officially did something no one has done. Ultimately, it wasn’t that bad for the Marlins — Alvarez as a hitter was worth 0.2 WAR. Ultimately, he didn’t under-perform a coin — a coin wouldn’t have hit 11 singles. But a coin, at least, would’ve had a better process. Alvarez’s ends justify the means, but the means just don’t make any sense.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
Awesome. This is why I read FG.
This is wonderful. It actually makes me giddy.
I suspect this might happen fairly easily if you often take the first couple of pitches on principle — and pitchers know it. Combined with that “automatic” no swing on pitches highly likely to be in the zone with terrible discipline and a possible tendency to chase breaking balls when behind and…viola!
The problem is his very high first-pitch swing rate!
i love the analogy, but coins have (or have appx) equal chance of heads or tails. Not the same with in-zone and out-zone pitches.
The league zone% was .488. If a hitter swings half the time (expected coin flip), then Z-swing would be .244 and O-swing would be .256. The Z-swing minus O-swing is -.012. So, yes, Henderson Alvarez did do worse than a coin.
the Z-swing would be 50% and the O-swing would be 50% and the Z-swing minus O-swing would be zero.
Given a 45% zone rate, and a coin that says swing or don’t swing BEFORE every pitch (not after you already know the results), you’d actually swing at ~62.5% of the strikes and 40% of the balls or Z%-O% of 22.5% so everyone on the list did worse than a coin.
kdm – your math is implying that the coin flip occurs after you know if the pitch is in the zone, which is a really weird way of interpreting the thought experiment.
d_i – if you assume 45% zone rate, z-swing is .225 and o-swing is .275.
@palehose – you’re doing the same thing. You’re assuming they swing at half the strikes and half the balls (as if that’s known before the pitch) rather than half the PITCHES before they are thrown. Here’s a table:
balls throw 5000
strikes throw 4000
total thrown 9000
zone % 44.44%
strikes swung 2500
balls swung 2000
z swing % 62.50%
O swing % 40.00%
swing % 50%
Z%-O% 22.50%
Pale Hose – These numbers aren’t necessarily my forte, so someone correct if I’m wrong, but it seems like the zone rate shouldn’t impact these numbers. I believe you’re using the wrong denominator for the z-swing/o-swing numbers.
A 45% zone rate suggests that a sample of 1,000 pitches should result in 450 pitches in the zone and 550 pitches out of the zone. If one were flipping a coin to determine the swing in advance, one would expect that he would swing 225 times at strikes and 275 times at balls.
To get a z-swing of 225, you would need to be dividing swings at strikes by total number of pitches seen, not number of balls in the zone seen. Instead doing 225/450 gives you z-swing of 0.50, with 275/550 also obviously giving 0.50.
Using your example:
1. Flip coin (heads=swing; tails=take) probability of swing = .5
2. Probability of in zone = .4444
3. Probability of out zone = .5555
4. Zone swing = .5*.4444 = .2222
5. Out of zone swing = .5*.5555 = .2777
steex – This is the definition of Z-Swing in the glossary.
Z-Swing%: The percentage of pitches a batter swings at inside the strike zone.
To me that implies swings as the denominator.
steex is right, I had my rows referencing the wrong things. Sorry for adding to the confusion.
I agree with you that’s what it says, but even in that very post there is discussion that is never address/concluded with regard to what that really means. Elsewhere, I see z-swing more clearly defined as the number of balls swung at inside the zone divided by the number of balls pitched inside the zone.
Having someone’s z-swing impacted by the total number of pitches they’ve seen is illogical – if two players both are thrown 500 in-zone pitches and both swing at 300 of them, they should have the same z-swing for the stat to be useful. The fact that one of them may have seen 600 pitches out of the zone and the other 800 out of the zone should be immaterial to z-swing, but would not be if total pitches is used as the denominator.
If swings is the denominator, then every player should have z-swing% + o-swing% = 1.000 (as that accounts for every swing), and that is definitely not true.
so did we conclude that i’m right?
Yeah, my understanding of Z-Swing was incorrect. That is a confusing definition, though.
If you swing at 50% of pitches, you’re gonna end up swinging at 50% of all strikes, 50% of all balls, 50% of all fastballs, 50% of all sliders, 50% of all pitches in the dirt, 50 % of any further classification! (on average, of course).
I love math but it kind of overcomplicates the concept here.
No, that definition implies percent in the zone is denominator, and that’s the only way the stat makes any sense.
there might be an obvious explanation. swing and miss pitches get him more pitches later in the at bat. swings in the zone are more likely to generate outs, swing and miss are more likely to lead to another swing.
Since he is the only one to do it in 8 years… there is likely no obvious/easy answer.
I would have liked to hear more about the few position players on this list… not so good for them.
Might just be that he prefers to hit breaking balls, which are slower and tend to be out of the zone, rather than fastballs. Could, in fact, even check that hypothesis.
That is some weird, wild stuff. Great piece. The conclusion was great.
I’ll be honest: I was starting to lose faith in Fangraphs after He-Who-Shalln’t-Be-Named decided to abort the perfectly cromulent embryo that is Notgraphs. But this reignites my passion.
And so I say: Make Jeff Sullivan the adoptive parent of Notgraph’s slowly decaying carcass so that he can also breathe life back into Banknotes Harper, Dick Allen inserted into Literature, and that post with the team of Adam Dunn’s, Koji Uehara’s, and Dee Gordon’s.
#KeepNotGraphs
Poor team Bartolos Colon, getting completely overlooked.
Also, poor, poor Dee P. Gordon. :'(
His walk and K-rates both improved this year!
Don’t look. You don’t want to look.
Seems to me, this could have much to do with trying to bunt when everyone knows you are bunting. When everyone knows you are trying to bunt, the pitcher is unlikely to throw you a strike. If you square up and miss, then it is a strike. So if you’re a modest bunter, but come to bat in many obvious bunting situations, then you could easily get these numbers.
Stop it. You don’t EASILY get these numbers. You just don’t. It is possible NO ONE else has ever had these numbers. You guys rag sportswriters for coming up with silly, easy, false narratives, and then every time Jeff comes up with an article like this, there are tons of attempts to explain it away…
Alvarez was signed as a 3rd baseman by the Blue Jays. He is really good athlete…He is incredibly weird though. Thats probably why his nickname is “The Entertainer.” Its fun to watch him every 5th day lol
This is a thoroughly fun piece to read.
Is it just me or did Alvarez get jobbed on the batter’s interference call? It seems to me that once a fielder (the catcher) has touched the ball it’s not that batter’s fault if the ball bounces into him.
These umpires
https://www.umpire.org/vb/showthread.php?t=8864
agree with me; once the catcher drops the ball, the batter is a runner, and the runner’s interference rules which require intent to interfere should apply. I suppose the relevant rule is 7.09(a), “It is interference by a batter or a runner when —
(a) After a third strike he hinders the catcher in his attempt to field the ball,” but even if “hinders” doesn’t require intent it’s hard to see how Alvarez is hindering the catcher when the catcher hasn’t popped out of his crouch yet.