The Last of Us Tops the Top 100: A Ranking That Never Published Its Rubric

The day Kinda Funny Games published its list of the 100 greatest games of all time, I opened the piece, scrolled to the bottom of the page and looked for the methodology. There was nothing there. No weighting, no criteria, no voter list, no collection window. The Last of Us sat at number one, 99 names trailed behind it, and immediately below ran two opposing currents of argument: one side insisting Part I is a cinematic masterpiece, another insisting Part II is the real advance in gameplay systems, a third dragging Sekiro into the fight. Ten years of reading arguments like this have taught me an uncomfortable rule: people argue about the ranking long before they agree on what the ranking is measuring.

The structure of the article is itself a signal. The controversy comes first, the ranking second — cause and effect reversed to hold tension at maximum. By the time the list appears, the reader has already been invited into a confrontation.

The evidence behind the controversy is far thinner than the controversy itself. The piece quotes “one gamer,” “some players feel,” “many people.” No platform, no count, no date. In my trade, that is the condition under which a claim gets removed from the dataset, not promoted to the front page.

The article does record a worthwhile question and then leaves it hanging: does this list judge by gameplay quality, by popularity, or by historical influence? The defence of the top slot cites historical influence and the Naughty Dog lineage, from Uncharted to The Last of Us. The opposition cites systems depth. Two different yardsticks are applied to one list, and nobody states which yardstick is official.

At the end, the author cools things down: choosing the greatest game of all time has never been easy. Even the dissenters, the piece notes, concede that Part I is one of the most outstanding games in history. That framing saves face on both sides while quietly constructing a consensus nobody ever measured.

The tail of the article carries a block of related headlines: LCK viewership doubling after six years of franchising, GAM Esports appearing as a VCT Pacific 2027 partner, LOUD returning to VCT Champions 2026, MLBB Vietnam mentioned for global competitive potential, a K/DA appearance controversy, and a sponsor-named VISA VMC Fall 2026 event. All of it in headline form only — no figures, no sourcing, no measurement window. I filed them into a verification queue and kept them out of the piece.

Take the LCK headline. Six years of franchising, viewership doubled — if true, that is a strong argument for the fixed-slot model. But such a comparison only holds once platform migration and the inclusion of co-streaming are stripped out. Without a measurement window and a baseline, I can only log it as a signal to track, not a conclusion to cite.

I have spent most of my career doing something many editors consider pointless: reconstructing timelines to the second for events other people describe in a single sentence. That principle was born in round 12 of the 2026 PFL season, when Jordan Minta left the pitch after 28 minutes. I was seventeen, writing for a community blog, and instead of copying down “cramp,” I rewound to the fourteenth play before the injury, mapped the running line, and cross-checked it against the way the opposing back line turned. I found the hamstring mechanism in a Philippine passage of play while Europe was looking elsewhere.

Since then, every conclusion of mine needs a timestamp, a location, and a specific frame. A medical report that says “serious injury” without grade, site, or mechanism has no usable value. A ranking that publishes no weighting belongs to the same category: it states a conclusion while denying anyone the right to audit the route. Apply frame-reading to a ranking and every line needs three things — who nominated it, on which criterion, at what point in time. Missing all three, the position is just a sentence of enthusiasm set in bold.

What I found in this article is a methodological error, not a disagreement of taste. Nobody is actually arguing about whether The Last of Us is excellent; they are arguing about what “excellent” is measured by. The winner of the debate is not the side with the better case — it is the side that got to choose the yardstick first.

If the list were mine to build, the methodology would read: historical influence 35%, systems depth 25%, narrative and staging 25%, durability over time 15%. Change the weights and you change the top name. That is precisely why rankings stay silent about weights. Publish them and the argument collapses in an afternoon.

Esports suffers from exactly this disease, except nobody has even bothered to build the list. Football counts every hamstring tear; esports lives inside its own medical darkness. And esports lives inside a second darkness, about how greatness is counted. A player with three titles in two years and a player with eight straight years inside the top four are two kinds of greatness that cannot be compared directly. Both fan camps are right, because each is holding a different yardstick and calling it the shared one.

What I would do if handed the assignment: fix four axes, assign weights, then score each player season by season rather than career-wide. International titles 30%, seasons inside the domestic top four 25%, time spent leading the meta 25%, influence on how others play 20%. Change the weights and the leader changes. The whole problem sits right there.

Across ten years of tracking matches and transfer windows in Southeast Asia, I learned to separate two things everyone else merges: medical risk and transfer risk. A player can carry a low re-injury risk and still be a bad signing, and the reverse. Rankings need the same separation. Editorial opinion and evaluative data are two different goods, sold in two different shops, and only one of them is allowed into my aggregation table.

The dichotomy the article builds — Part I as safe, Part II as ambitious — is its most durable asset, because it does not die with the ranking. In esports the same pair shows up every season: the champion that wins with a proven style gets called boring, the team that shifts the meta gets called reckless until the day it wins. One choice, two labels, depending on the result.

The headline block also reveals an editorial direction: VCT Pacific 2027, VCT Champions 2026, MLBB Vietnam, GAM Esports. A media outlet widening its topical surface toward Southeast Asia and the Pacific. I mark that as a media signal, not a competitive one — the two get read interchangeably and produce bad forecasts about regional strength.

I tell young editors this often: the body does not lie — it simply speaks a language the medical room has not translated. Rankings behave the same way. They do not lie. They speak a language nobody has bothered to translate, and leave the reader to guess.

The counter-hypothesis I am obliged to put on the table: the missing methodology was removed on purpose. A ranking with clear weights generates a few hundred comments. A ranking with no weights generates a few hundred thousand. The controversy here is the product, not the by-product. The line “even critics concede” is a consensus-manufacturing device: concede a weak claim, collect a strong one.

I have to own an error of my own. On first read I filed this piece under the esports vertical, because it sat beside a block of esports headlines and because I read it with professional habit. On review, the entire body contains not a single esports entity: no competitive title, no tournament, no roster, no patch, no organiser. I pulled it out of the vertical and logged the mistake. A mislabelled item distorts entity counts, topic distributions and sentiment baselines across an entire dataset — a far higher cost than one wrong article.

Next time an esports “greatest of all time” list appears, I will ask for the weights before I ask for the positions. If whoever built it cannot answer, I will read it as editorial opinion — worth quoting, not worth counting. And if someone in Manila or Jakarta builds such a list with public weights, I will be the first to open the file and recount every line, even if the result does not support what I believe.