The MCU Phase 4 and Beyond (2021-present) Review
4/10
Infinity War and Endgame resolved eleven years of setup. What followed was a slate built to fill a streaming service, with thirty-one projects across two phases in roughly four years.
The 4 is arithmetic. I’ve seen twenty-four of the thirty-one and scored them, and the mean comes out at 3.75 with a median of 5. Rounding to the phase as a whole gives a 4, and the number hides a pattern that tells you more than the average does. That pattern is a split. Thirteen of the twenty-four sit at 5 or above and nine sit at 2 or below, with almost nothing in between. This is a machine producing good work and near-worthless work with no middle ground.
Watching this run was a letdown I never expected from the studio that pulled off Endgame. The good entries reminded me what this franchise could do, the bad ones felt like homework set by a marketing department, and the whiplash between them wore me down faster than a steady decline would have.
The Numbers
| Project | Phase | Score |
|---|---|---|
| Spider-Man: No Way Home | 4 | 8 |
| Loki, season 1 | 4 | 7 |
| What If…?, season 1 | 4 | 7 |
| Guardians of the Galaxy Vol. 3 | 5 | 7 |
| Loki, season 2 | 5 | 7 |
| WandaVision | 4 | 6 |
| Hawkeye | 4 | 6 |
| Black Widow | 4 | 5 |
| Shang-Chi | 4 | 5 |
| Doctor Strange in the Multiverse of Madness | 4 | 5 |
| The Falcon and the Winter Soldier | 4 | 5 |
| Moon Knight | 4 | 5 |
| Deadpool & Wolverine | 5 | 5 |
| Eternals | 4 | 3 |
| Black Panther: Wakanda Forever | 4 | 3 |
| Thunderbolts* | 5 | 2 |
| Thor: Love and Thunder | 4 | 1 |
| Ms. Marvel | 4 | 1 |
| Ant-Man and the Wasp: Quantumania | 5 | 1 |
| Captain America: Brave New World | 5 | 1 |
| Agatha All Along | 5 | 1 |
| Ironheart | 5 | 1 |
| She-Hulk: Attorney at Law | 4 | -1 |
| The Marvels | 5 | -1 |
Phase 4 averages 4.4 across fifteen scored projects with a median of 5. Phase 5 averages 2.67 across nine with a median of 1.
The median is the number that indicts it. A mean can be dragged down by two disasters. A median of 1 means the typical Phase 5 project, and not the worst one, sits near the floor.
That median of 1 is the damning number. One disaster is bad luck. A typical project sitting near the floor means the work was being waved through on summaries, and the people paying for tickets were the ones who found out.
The gap between the two phases is 1.73 points on a ten point scale. That’s the distance between a body of work you’d recommend with reservations and one you’d tell somebody to avoid unless they already know in advance which specific entries are worth the time.
Films and series decline together. That rules out the easy explanation. Phase 4 films average 4.29 and Phase 5 films 2.5. Phase 4 series average 4.5 and Phase 5 series 3.0. The television was never the problem people said it was. Seven projects are unscored because I haven’t seen them, and their absence probably flatters the result. Nothing in the unseen group carries the reputation of a No Way Home or a Guardians Vol. 3.
The unseen list is itself a finding. Werewolf by Night, the Holiday Special, Secret Invasion, two seasons of What If, Echo and Daredevil: Born Again went past without ever becoming appointments, which for somebody who saw every film of the first three phases in a cinema is the clearest measurement available of what the volume did.
Scoring your own field produces information that impressions don’t. I would have said this period was uneven. The distribution says something more specific, and I couldn’t have arrived at it by remembering.
Look at the shape and not the average. Two bodies of work with the same mean can be entirely different objects, one clustered and one split, and only one of those has a fixable problem.
Apply it to your own output. Rank everything you’ve written, and the pattern will tell you whether your weak work is weak in a consistent way or whether you have two separate modes.
What the Good Ones Share
Every project scored at 6 or above is either self-contained or a continuation by the people who started it. Loki in both seasons has a governing idea about determinism, a bureaucracy with rules it establishes and obeys, and a lead who wants something. Guardians Vol. 3 is James Gunn finishing a decade-long story about a cruelty done to one character. WandaVision spends most of its run inside a formal idea about grief.
No Way Home is the outlier at 8 and appears to break the rule, since it depends entirely on external material. What it depends on is twenty years of films the audience already saw. That’s a completely different proposition from requiring a series that aired last month. The projects at 1 or below share the opposite trait. New leads, spin-offs and handovers, made to introduce somebody instead of to be about something, with the introduction as the entire purpose. I liked Loki a great deal, and Guardians Vol. 3 got to me the way the old films did, because Gunn cared about one damaged character and finished his story. The handover projects are the worst of the lot. Nobody making them seemed to care who these people were, only that they were now on the board.
Both of the Phase 5 sevens confirm it from the other direction. Guardians Vol. 3 and Loki season 2 are continuations by returning creators of things established in Phase 4. That means every good thing in Phase 5 was planted before the volume decision took hold.
Curation Does Not Scale
Phases 1 to 3 released two or three films a year, and Kevin Feige personally shepherded each. Phase 4 onward released around forty-seven projects in five years across film and streaming.
That process has a throughput limit set by one person’s attention. At three films a year a producer reads every draft and watches every cut. At ten projects a year across two formats, they’re approving summaries.
What disappears first isn’t craft. That can be hired. It’s selection, because judgment can’t be delegated without becoming something else, and the split distribution is what selection failure looks like from outside.
The selection failure frustrates me most, because it was a choice. Feige had shown across three phases what his attention could do, and the studio spread it over so much product that it stopped meaning anything. I’d have cut the slate in half and been happy to wait. The visual effects situation is the same problem in another department, with vendors reportedly working impossible schedules across overlapping projects. That’s a supply line where the delivery dates are fixed and the work is the variable.
Quality control is a bottleneck by design. Remove it to raise throughput and the quality control goes with it, and there’s no version where you keep both. A contract specifying a book a year is a decision about quality disguised as a decision about scheduling. The rate is the constraint that determines everything else.
The uncomfortable part is that median output usually stays acceptable, so nothing forces a correction until the audience stops turning up.
Homework Is Not a Story
Plot began moving between formats. An arc starts in a series, resolves in a film, and references a special that aired between them. That assumes total consumption. Somebody watching only the films meets characters behaving as though a relationship exists, and the film can’t pause to explain without insulting everybody who did the reading.
The first three phases avoided it. Every film worked alone and the connective material was a reward for attention. Optional depth became mandatory prerequisite, and the difference is whether your audience can start anywhere.
The multiverse compounded it. If a death can be undone by locating another version of the person, the audience has been instructed not to invest, and no amount of well-staged spectacle recovers from that instruction.
The multiverse is where I stopped caring about deaths in these films. If anybody can come back, nothing in the fight scenes costs anything, and I don’t see how the studio thought bigger spectacle would make up for that.
The Audience Learned to Wait
Opening weekends held up longer than quality did, and the decline shows in second weekends and in the gap between openings and totals.
Ten years of consistent quality created a habit of turning up without checking reviews. That’s the most valuable thing a franchise can own. Roughly three years of split output spent it, and habits of that kind are far easier to lose than to build. Streaming accelerated it. A film reaching a subscription service within months trains its own audience to weigh a ticket against a short wait. The split distribution makes that worse than a flat decline would. If everything were mediocre, an audience would lower its expectations and keep attending out of habit. When half the output is good and half is close to worthless, the rational response is to wait for the verdict, and waiting is the behavior that removes an opening weekend.
The correction has begun, with a reduced slate and several projects canceled or delayed. The casting disruption of recent years forced rewrites at exactly the moment the plan needed stability. That’s outside anybody’s control and made the volume problem worse.
When output volume is set by a business requirement instead of by what the work needs, the work is the variable that absorbs the difference. That’s why this belongs on a page about films that needed somebody to say no. The person who should have said it was looking at a slide about annual content hours, and no scene in any of these projects could have fixed that decision. The audience habit is the asset being spent. It takes a decade to build, it’s invisible while it holds, and the spending only becomes measurable once it’s gone.
Should you watch MCU Phase 4 and Beyond?
Watch the ones that are about something and skip the rest. The 4 is what’s left after the failures drag the good entries down.
I’m more irritated than sad about it, because the good ones prove the studio still knows how to do this and chose volume anyway. The average is 3.75, and the failures and the successes are further apart than they’ve ever been.
