Multiple reviews are good, because they give us different perspectives from other people on the same game, which can in turn allow us to discern whether or not a game is worth going for, or simply to look back on said game differently than what we saw it as previously.

Also I heard Ghostwire Tokyo suffered on Xbox than it did on PS, so in this case, yes, multiple reviews are absolutely welcome and much needed for this sort of topic (be it performance or just a different opinion a year later).

"VGChartz's Lee Mehr hated this game, but man oh man did VGChartz's Craig Snow pull through for us and gave it a 9/10!" - potential accolades trailer

The aggregate site question is definitely another good argument for the professionalism strand in OP.  Though it seems places like MetaCritic would be more accommodating, since they split up votes between different versions.  So long as the synopsis on Meta outlining the special conditions for something like No Man's Sky, perhaps most of the audience wouldn't feel like either is more "official" than the other.  For example:

No Man's Sky 2016 on PS4:  "Buggy, mismanaged, and unwelcoming, NMS represents the pitfalls of a near-infinite universe."  

No Man's Sky years later on Xbox One: "The long journey since NMS' contentious launch years ago fulfills more of its initial concept's promise, though yada yada yada."  

In my mind, as long as it's carefully worded, I think both can be considered "official" for their respective time.  But like you said: maybe all supplementary "x years later reviews" could be slotted as score-less or articles instead.

TBH, I don't read reviews because my tastes don't seem to align with main stream media.

Reviews feel stale to me, and not sure multiples would be worth me reading.

Reviews can also become predictable if say the same person writes it on their own. After 100 reviews you naturally fall into a pattern of the things you talk about.

However here is an idea (and it isn't exactly a new one but rarely is it done anymore), rather then have 3 articles, why not try say three reviewers where they all discuss the game provide their final verdict on the game.

But instead of it being written in one tone like a traditional review, have say a few set questions and treat it like a panel debate and you can have all three reviewers agree, disagree and provide their input to each question or topic of discussion. In essence write it in a way that when you read it, it is like seeing three mates having a natural discussion on a topic and provide their views on it. I think this way it will cut out having redundant essays to read, when say 2 of the 3 people agree on an aspect of the game and if they write it on their own it is repeated. It also makes it less predictable what will be said in each review as the people discussing the game bounce of each other and things come to mind that you may have missed if you wrote a review on your own.

As for final verdict,

Person 1 = 8/10

Person 2 = 9/10

Person 3 = 9/10

Get the average of the three and round up to nearest 0.5 or 0.  This way at least you also cut out wild ranges for multiple reviews as you are providing a combined avg consensus. 

I feel reviewing again should only be done if there are significant changes to the game. I feel Ghostwire Tokyo hasnt seen much difference between both versions as the changes only consist of a new mode or few collectables/side missions.

I don’t see the point really. I wouldn’t read more than one review. If I wanted to, I’d go to metacritic because there there’d be a much wider range anyway.

I see.  Perhaps I put too much stock in the advertised changes to Ghostwire's recent update.

I feel like I'm split between your threshold and Machina's issues with multiple reviews overall.  If a hypothetical TLOU Pt. II Director's Cut came along that made gameplay updates & notable alterations to story pacing, I don't think I'd reflexively shoot down someone else reviewing it here.  I guess the tension comes back to how Machina said would be the "real review."  I'd hate to see responses that acted like a more positive review would be like "fixing past errors," if you know what I mean.

If you had like a director's cut then that would almost be it's own database entry I would think, or at least identified as a variant to the original game. If that reviewer was smart, they would then make reference to the original review and maybe touch base about the thing they agreed on in terms of the negatives, and if those negatives have been addressed and whether they believed it was a superior experience to the original and worth the purchase again. Conversely they could also conclude that the changes made, actual deter from the original game experience and they recommend to those who haven't bought it yet to find the original copy.