Polygon gave Persona 5 a 7.5. Metacritic average is 93. The math ain't mathing.

Let's open the spreadsheet. Persona 5 sits at a 93 Metacritic average across 90+ critics. The standard deviation across major outlets is 3.4 points. Then we have Polygon's Phil Kollar at 7.5/10. That's not a difference of opinion. That's a 1.8 standard deviation outlier in a distribution where 'outlier' usually means 'the reviewer didn't finish the game.'

For context: 73 of 90 critics scored Persona 5 at 90 or above. Polygon's score lands closer to where critics put Sonic Boom: Rise of Lyric than where the consensus put one of the highest-rated JRPGs of the decade. The math ain't mathing. I ran the regression. The regression filed a complaint.

Now let's audit the review itself. Word count: 1,847. Words spent on combat, dungeon design, palace mechanics, social link system, music, art direction — the actual systems comprising 100+ hours of gameplay: roughly 600. Words spent identifying problematic depictions in a Japanese game from 2016: roughly 700. The remaining 547 are connective tissue. That's a 38% single-topic concentration ratio, which I have plotted on a chart, and the chart looks like a hockey stick.

I'm not here to argue the criticisms are wrong. I'm here to argue arithmetic. If 38% of your review is one topic, the score should reflect that the topic was 38% of your experience. Instead Polygon delivered a 7.5 — a score that suggests 'mediocre with issues' rather than 'great game with one significant problem the reviewer wanted to discuss.' Pick a lane. Either the game is a 7.5 across the board, in which case justify the other 62%, or it's a 9 with caveats, in which case score it like one.

The numbers don't add up. They never do at Polygon. I keep a chart. The chart is in red ink now. The chart has feelings.