What makes movie action “good?”

There seems to be pretty uniform agreement that John Wick movies have great action, and simultaneously, there seems to be a lot of general “Marvel fatigue,” some of which is specifically about the MCU, but also some of which is about “action blockbusters” that seem to put a lot of focus on action and spectacle that a lot of people seem to be growing increasingly disenchanted with. See, for example:  Tomorrow War (2021) starring Chris Pratt, Extraction (2020) starring Chris Hemsworth, and to round out the Chris roster we also have Wonder Woman 1984 (2020) with Chris Pine.

Most of the discussion tends to focus on effects, and how making everything in a computer feels soulless while practical effects like the ones in John Wick and Everything Everywhere All At Once are superior, and there’s a certain degree to which blockbuster CG does look worse due to everything being computer generated. But I think a much bigger and more deep-seated problem is that blockbuster action is poorly structured.

There are a lot of great practical scenes in John Wick with choreography that is great not just in execution, but in concept: if you recreated them with computer animation, or hand-drawn animation, many of them would be just as fun. You could look at a storyboard with a director’s poorly-drawn pencil sketches and you’d still be able to follow the logic and understand what the scene’s impact is supposed to be. You can tell this, because people can describe these scenes, and regularly do exactly that, because they’re memorable. It’s what the legend of John Wick is all about. It’s not just the “top 10 most creative John Wick kills” listicles that remind us of that fact; people remark about it in-universe.

The thing is, John Wick doesn’t rely on the unique outlandish props for memorable kills. Most of John Wick’s kills boil down to “point gun, pull trigger.” And yet despite that most of his kills are with bullets, so many of them feel unique. Each scene feels like it has a tempo to it, a sense of flow, a chain of cause and effect. “That guy is hiding behind a pillar, but his toe is sticking out. So I’ll shoot him in the toe, causing him to bend over in pain, exposing the rest of his body so I can shoot center of mass.” There’s always the moment when John Wick runs empty and has to reload before delivering the killshot.

There’s a storytelling principle that’s often applied to plot structure. This was famously described by Trey Parker to a bunch of NYU students on MTVU’s “Stand In.” Trey Parker describes the writers room as containing a massive white board, split into 3 acts, where they write down ideas scenes and rearrange them. Each scene has to be entertaining by itself, but they also need to be connected by a coherent narrative through-line.

“You don’t want just one scene where, ‘well, what was the point of that?’ Take the beats of your outline, and if the words 'and then’ belong between those beats, you’ve got something pretty boring. What should happen between every beat that you’ve written down is either the word 'therefore,’ or 'but.'”

A happens, therefore B happens. Or, B doesn’t happen, but C happens, therefore D happens. Repeat.  You can use this to structure the plot for a 25-minute TV episode, or a 120-minute movie, or a 2-minute action scene. Follow it, beat by beat, and see if one leads to the next.

One of the evil henchmen comes swinging at our hero with a wooden chair raised over his head, but the hero dodges and the wooden chair smashes against the floor, shattering. Therefore, the ground is now covered with the broken remains of a wooden chair, so the mook picks up one of the long pieces of wood that served as a chair leg and begins swinging at the hero again. But the hero successfully dodges and the wooden club breaks and splinters in a way that causes it to become, therefore the mook starts trying to use the splintered end to stab the hero…

I’m not saying this is an amazing action scene, but it’s a competent one: each beat flows into the next. The chair becomes a chair leg, then the chair leg becomes a shiv. Each time the henchman comes rushing at the hero, he’s doing it with a different weapon. The scene has a logic to it: if you rearranged the shots, the scene wouldn’t make sense. You can’t start with a splintered piece of wood and then end on an intact chair.

So many mediocre action movies fail to deliver this basic sense of progression from one shot to the next. Instead, you get what Trey Parker describes as a series of “and thens.” The bad guy punches the hero. And then he punches the hero again. And then he tries to kick the hero. And then he punches in a different, cooler way. There’s no real sense in which each beat is a consequence of what followed it: if you cut the scene up and rearranged the shots, a lot of people might not even notice. (And in fact that sort of thing happens all the time in the editing bay.)

Tony Zhou self-deprecatingly describes this problem when critiquing one of his own videos, saying “This is a list you could put in any order. That’s why it’s so boring.”

For examples of action cinema where every beat feels like a consequence of what preceded it, watch any classic Jackie Chan movie (the ones that came out of Hong Kong, not Hollywood). Tony Zhou describes it like this:

“So how does Jackie create action that is also funny? First off, he gives himself a disadvantage. No matter what film, Jackie always starts beneath his opponents. He has no shoes. He’s handcuffed. He has a bomb in his mouth.“

“From this point, he has to fight his way back to the top. Each action creates a logical reaction. And by following the logic, we get a joke.” (Jackie is facing an assailant with a gun; Jackie has a gun, but it’s empty. Therefore, Jackie fakes surrender, handing his empty gun to the assailant. Therefore, the assailant is now holding an unloaded gun in his left hand. The assailant now thinks he has control of the situation, but reaching for the unloaded gun distracted him the fact that Jackie was entering a fighting stance and getting ready to kick: therefore, when Jackie kicks, he succeeds in knocking the loaded gun out of the assailant’s right hand. Therefore, the assailant tries to fire at Jackie using the gun in his left hand – which is empty, and he realizes it in a moment of surprise which Jackie seizes on by punching the assailant in the face.)

This is the joy of watching Jackie Chan films: much like the example of a chair (which morphs into a chair leg, which morphs into a shiv), a prop in a Jackie Chan movie is rarely just one thing. A ladder isn’t just a ladder; it’s a prop. And it’s several different kinds of props. Fighting with a ladder like this:

…is subtly different than fighting with a ladder after this happens: 

And if you flip it over someone’s head, it suddenly becomes a cage: 

…trapping one of the bad guys just long enough for him to look in surprise right before Jackie’s fist intersects with his face.

It’s great for action comedy, but it’s also great for straight action: sometimes, the “punchline” is someone getting defeated in a surprising way. John Wick is one of the few big franchises of recent years to reliably do this sort of thing well.

Often, John Wick accomplishes this by being clever.  But I think a big part of it comes down to the fact that John Wick is just mortal enough for the number of bad guy’s he’s facing to matter. Each scene needs a sense of “progress,” where the stakes are constantly changing, and sometimes the change in stakes is as simple as, “There are five bad guys, oh no!” Bang bang, pivot, bang bang. “Okay, now there are only three bad guys.” (It’s harder to do this when you’ve been injected with super soldier serum and wear a suit made of high-tech blast resistant stretch fabric: Captain America subduing five bad guys doesn’t feel meaningfully different from him subduing three bad guys, even if the way he punches them is really cool.)

Stakes matter! If threat level scales linearly with the number of bad guys on screen, then each scene will have a natural eb and flow to it as bad guys get gradually picked off (or as more of them stream into the room, or pick themselves up off the floor and reach for the gun they just dropped).

One of the MCU scenes that actually did this better than most is the famous Captain America elevator scene: first, Crossbones and two guys get onto the elevator with cap.

Seemingly nothing out of the ordinary. But why is that guy resting his hand on his hip so close to his gun? Several floors later, the elevator opens, and four more guys get on. They’re wearing suits, like you’d expect from people who just showed up for a day at the office. This is headquarters, there’s nothing to be worried about. So why is that guy sweating? 

Then, the  door opens again, and three more guys get on – and these guys are wearing tac gear. But hey, it’s Jack – I know Jack, he was in the first ensemble movie, he can’t be one of the bad guys…so why is he standing directly between me and the door?

It’s a great example of slowly amping up the tension by gradually adjusting the threat level up. The scene even amps up the tension by having the magnetic handcuff, which leaves Cap in various stages of incapacitation throughout the fight. And he has to fight his way up from the bottom. But we very quickly go from this iconic shot:

…then after literally three seconds of Cap delivering a rapid series of strikes to incapacitate most of the mooks, he’s back to fighting two or three of them at a time.

Despite there being ten bad guys in the elevator, it’s kind of hard to get a clear fix on how many of them attacking Cap at any given moment, all of the others existing in various states of injury and recovery after they get the wind knocked out of them. In fact, the only shot that allows us to get a full body count is after the fight is over:

It’s still a fun scene, among my favorite MCU moments, succeeding at being tense, fun, and memorable.

The bigger issue is the need to be “epic”: threat level can’t scale linearly with the number of bad guys on screen when you want a scene where there are literally hundreds of enemies to fight off.

There’s a certain point at which the marginal effect of another bad guy on screen is effectively zero. Obviously, this is the case when you have hundreds or dozens of enemies on screen, but you run into it sooner than you might think: there’s a real sense in which “group of seven bad guys” doesn’t feel different from “group of six bad guys.” Our brains just categorize both as “a pretty big cluster.” Research tends to come to slightly different results on this, but it seems like humans count “one, two, three, four, many.” Once you have five bad guys on screen, adding a sixth bad guy doesn’t do anything to change the stakes.  (That’s the problem with having a hero who’s so strong that you need to throw 10 bad guys at him to pose any threat.)

If you cap the total number of on-screen bad guys at five, then each enemy the hero defeats meaningfully changes the stakes, and John Wick does this a lot. In many cases, the flow just comes from watching the number of bad guys on screen decrease linearly as the hero picks them off, one by one. It gives the scene a natural scenes of progress, and it can sometimes be played for comedy, like Neal Stephenson does in Snow Crash in a scene described by a sniper’s dialog:

"It’s, like, one of them drug dealer boats,” Vic says, looking through his magic sight.  "Five guys on it.  Headed our way.“  He fires another round.  "Correction.  Four guys on it.”  Boom.  "Correction, they’re not headed our way anymore.“  Boom.  A fireball erupts from the ocean two hundred feet away.  "Correction.  No boat.”

That’s a (very short) scene with flow. You can’t rearrange the beats; every beat leads to the next. Follow the logic, and arrive at the punchline.