I’ve been making my way through Marvel’s Spider-Man 2. This game was released in 2023 for the PS5, a headliner title that served as a showcase for both the console’s graphics and its loading time capabilities1—so much so that when I fire it up in the living room, my kids watch it like an animated movie. It’s also the direct sequel to the surprisingly good Spider-Man on the PS4 as well as its Miles Morales expansion, and draws its story elements from a mixture of the Marvel Cinematic Universe and the long-running Spiderman comic books. On the flip side, the game doesn’t deviate too much from its predecessors, opting rather to add more of everything: city size, movesets, (super-)villains, side quests, and—set pieces.
A bit of background first. Video games have long-employed cutscenes, which are breaks in player control to play pre-programmed and pre-animated segments, most often to advance the game’s storyline. One of the earliest and simplest examples is in the original Super Mario Bros, whereupon defeating Browser at the end of every level trilogy, you reach Toad in the castle’s innermost sanctum, who solemnly declares that “Our Princess is in Another Castle!” Of note here is how the player loses control of Mario and the camera during each of these sequences, as well as the custom text display and character that’s not reused anywhere else in the game2.
By the time games moved to be rendered in 3D in the PlayStation 1 era, there was an added dimension of movement and freedom camera, with in-game objects built from sharp and blocky polygons. The hardware was still limited in computing and memory capacity, so custom behaviors that were not a part of the core gameplay had to be separately loaded or streamed from the CD, instead of kept in memory to be readily accessed. At the same time, pre-rendered 3D graphics—Toy Story was released around this time—looked much better, with a level of realism that helped tell games’ stories. This confluence of factors created a dynamic, where a player would take an action to trigger a story sequence, and then the game would fade to a prerendered cutscene. The quality and amount of detail contrasting the 2 parts were noticeably different, as if the higher fidelity graphics were there as a point of reference, to help conjure the missing details during the gameplay portions when you took control again of your character. Final Fantasy VII was my first recollection of this technique being used for a major title, but it’s continued to be used and refined through the years and generations of hardware3.
By around the PS3/PS4 timeframe, gaming hardware got fast enough to render most cutscenes in the game’s native gameplay engine, intact with all the visible player customizations and the surrounding environment; the developers no longer needed to fade or cut into videos as much. To reintroduce autonomy, game designers tapped into something called Quick Time Events (QTEs): gameplay sequences that require the player to press buttons on the controller to perform character actions, but they’re custom-animated and are built specifically for this scene. Players who were used to years and years of sitting back when a cutscene shows up are now reminded to pay attention, as failing a button press or button mash could cause a premature “game over,” or at least some loss of health and status.
The problem with QTEs was that they were intended to inject interactivity during more passive parts of a game’s storytelling, but they were overused in too many games, and with too many scenarios a button prompt—or 10 of them—still felt decoupled from the character’s actions. Modern games allow the player to run around, jump, climb, attack, defend, pan the camera, manage their inventories, and a dozen combinations of these discrete activities that grant player autonomy; QTEs reduce all those possible actions to 1 or 2 dimensions by pressing on a few buttons, and since the sequences are so tightly scripted, the outcomes are either pass/fail or just pass. The “Press F to Pay Respects” meme from Call of Duty was so bad because it was such a ham-fisted attempt to layer on an interactive element to a scene that didn’t need it.
Alright, back to set pieces. With modern PS5-era games, cutscenes and QTE-like actions are much better integrated and generally feel more natural. Game designers have learned to restrict axes of control to tell their stories. For instance, a normally 360°, free-moving camera would be locked behind Kratos in God of War, with only 90° of movement so the player is forced to focus on what’s in front of them. At the same time, the edges of the environment are blocked off, nudging the player away so they can trigger the character position to progress. While normal gameplay would have the player hit the “attack” button to swing their axe at an enemy, in this cutscene they’ll instead swing it to chop wood. It results in a tightly directed sequence with limited control for the player—what I’m calling here, a set piece.
Spider-Man 2 takes this type of hybrid scripted gameplay to a new level that makes them feel even more epic. In the game, Spiderman has a wide variety of moves—including all the punching, kicking, web-slinging, and spidey powers that define the comic book hero. All these actions are available to the player during regular gameplay4, and the tutorial sections introduce each in turn. The game designers explicitly wanted the gameplay to flow and for the characters to always preserve momentum in their actions.
But during some of these set pieces, Spiderman would be swinging through falling debris, webbing a safety net for innocent bystanders, or zipping into a concrete pipe, with actions that are similar to what you’d do in the open world but with higher stakes and more cinematic results. The game cleverly uses the same prompts as the core gameplay, so the player may not know that it’s going to cut into a directed sequence until after they perform the button presses and Spiderman dutifully zips to the next ledge. Sometimes the set piece goes into a prolonged cutscene, other times it returns you to the action a few seconds later. The dynamic nature of how much control the player has, and when they have it, strikes a much better balance than the QTEs of prior generations.
I’m fascinated by how video games are evolving their storytelling. As mentioned above, in the early days it was good enough to write a sentence or two to lay out the rationale for your journey5; eventually, these plot points would be interwoven throughout the game as it tells its story interactively. It’s taken many iterations to get these mechanisms to feel unintrusive—that it’s not a movie with button prompts6, or a bunch of mini-game skill checks that end up feeling generic and uninspired. AI is the next frontier here; there are already some demos that showcase conversations with NPCs that are unscripted, though, at this early juncture, the interactions are aggressively bland and don’t measure up to crafted sequences. Give it a few more generations.
Seriously, the game’s fast travel times are near-instant.↩
In fact, these are the only parts of the game that write out full sentences.↩
One tell, which still applies today, to say the Final Fantasy VII Rebirth remake rendered on the PS5, is whether your prerendered cutscene has its characters wearing their customized outfits, or are still sporting their generic looks.↩
So much so that it’s a bit overwhelming; Spider-Man already had a bunch of moves, and the sequel here assumes familiarity with the previous game and adds on even more.↩
Credit goes to Double Dragon for taking it to the next level in the NES era, showing your girlfriend’s kidnapping at the start so you’re justified in beating up their army of goons.↩
Although there are now subgenres that cater to that sort of gameplay if you’re so inclined.↩