A review published on launch day was written under a schedule most readers never see. The amount of the game the writer covered is often the most important unstated variable.
The window is shorter than the game is long
Review copies typically arrive a week or two before release, and a large role-playing game or open world can take considerably longer than that to finish thoroughly.
The writer is also usually working on other coverage in the same period, so the available hours are a fraction of the calendar time.
Finishing at all can require playing far more intensively than any ordinary player would, which changes how pacing and repetition are perceived.
Intensive play distorts specific judgements
A game designed to be played across two months feels different when compressed into ten days, and repetition that would be spaced out becomes immediately obvious.
Conversely, systems intended to become tedious over time may not have time to become so, which can flatter a game's later hours.
Careful reviewers flag this, and the better ones describe how they played rather than presenting the assessment as a neutral observation.
Multiplayer cannot be assessed in advance at all
Competitive and cooperative games depend on populated servers, and pre-release sessions run with a small organised group bear little resemblance to public matchmaking.
Server stability, queue times, matchmaking quality and the behaviour of the community only exist after launch, and those are frequently the deciding factors.
This is why many outlets now publish impressions first and a scored review later, treating the launch period as part of the assessment rather than an inconvenience.
Some content is structurally unreachable
Endgame activities, late progression and content gated behind long play are often out of reach within a review period, and a reviewer may be told about them rather than reaching them.
Optional content presents the same problem, since a reviewer covering the main path thoroughly cannot also cover a large volume of side material.
What gets assessed is therefore weighted towards the opening and the critical path, which is also the part most players see.
Disclosure is what makes the review usable
A stated play time, a note on which version was tested and an explanation of what was not reached let a reader calibrate the assessment themselves.
Outlets increasingly publish these details, and their absence has become a reasonable signal that the review should be read more cautiously.
The alternative, waiting until everything can be covered, would place the review long after the moment readers need it, which is the constraint the whole format works within.