PLAY'S · Explainer · 8 Sep 2026

Where the numbers come from

Takara Tomy publishes no usage statistics for Duel Masters PLAY'S. Not deck win rates, not play rates, nothing. Every number anyone quotes about this game's metagame was counted by a person.

Which means the numbers have properties worth knowing before you rely on them: sample sizes that vary by a factor of three, brackets that disagree with each other, and at least one widely-cited source that states plainly it is not measuring anything.

Three sources, three methods

1 · Ranked-match surveys

A volunteer group records the decks they face while playing ranked, and publishes the distribution. Several hundred matches per day, every day, split into New Division and All Division. Published at 18:30 JST.

This is the largest ongoing dataset in existence for the game. It is also the least controlled: it measures what a particular set of players ran into at whatever ranks they were playing, which is not the same as a random sample of the ladder.

A second group publishes a smaller ranked survey — around a hundred games at a time — split into all players and the top 500. The split is the useful part. Those two brackets routinely disagree.

2 · Tournament decklists

Sanctioned events collect decklists at entry, and the organiser passes the breakdown to the coverage sites. This is the cleanest data available — every entrant is counted exactly once, and nobody is estimating.

Recent field sizes, which is where it gets interesting:

30 JulREIGN DMPS VERTEX CUP319 19 AugREIGN DMPS VERTEX CUP293 22 AugREIGN DMPS VERTEX CUP251

The field shrank by 21% in three weeks. Read that carefully before drawing conclusions from it — tournament attendance moves with the calendar, the day of the week, and what else is running. A smaller field is not necessarily a smaller player base. But it does mean that a percentage from the 22 August event rests on a fifth fewer decks than one from July, and a deck at 4% is eleven decks in one and ten in the other.

3 · Tier lists

The most-read source, and the one that is not data at all. The site running the best-known tier list says so on the page itself:

This is purely our own team's assessment. Please treat it as a reference. Kamigame, on its own New Division ranking page — translated

Tier placements are editorial judgement informed by the data above. That is a legitimate thing to publish and a useful thing to read. It is not a measurement, and the site is not pretending otherwise — the misunderstanding, where it happens, is downstream.

How to read them without getting it wrong

Whole field and top bracket are different questions What people brought and what performed are separate facts, and the sources publish them separately for a reason. In the 30 July tournament, the deck with the largest share of the field was not the one doing best at the top tables. That breakdown is here.

If a number is quoted at you without saying which bracket it came from, it is close to meaningless.

Ranked surveys and tournament data measure different populations Entering a bracketed event with a submitted decklist selects for players who prepare. The ladder on a Tuesday evening does not. Expect the two to disagree, and treat the disagreement as information rather than error.

Check the date on everything Tier lists move two or three times a week. Ranked surveys are daily. A tournament breakdown is a snapshot of one afternoon. A month-old percentage in this game is not slightly stale, it is a different format — a set released in that window, a collaboration brought in new players, and the field size moved 21%.

What nobody has

Exists	Does not exist
Deck play rates, counted by hand	Official play rates
Tournament decklists	Win rates of any kind
Editorial tier assessments	Matchup matrices from real data
Legend rank cutoff, tracked daily	Player counts, retention, spend

The absence of win rates is the big one. Play rate tells you what people chose; it says nothing about whether the choice was correct. Every claim you will read about a deck being "strong" is inference from placement and popularity, not from a measured win percentage — because nobody outside the company can measure that.

Why this is worth knowing

Players in other digital card games are used to third-party trackers reporting win rates from millions of games. Nothing like that exists here. There is no client-side tracker, no public API, and no official dump.

What there is instead is people volunteering to count — one group has been doing it daily since 2020, having started from a single player logging fifteen matches a day. More on who they are.

That is a remarkable thing to sustain for six years, and it is also a reason to hold every number loosely. Hand-counted data is real data. It is just not the kind that supports three decimal places.

Sources

Tournament field sizes — 319, 293 and 251 decks for the 30 July, 19 August and 22 August events — from Kamigame's metagame data page, credited to REIGN as the organiser supplying them. Page last updated 4 September 2026.
Ranked survey cadence and sample size: BEANS, whose reports are published daily, and Kamigame's own ranked data with its all-players and top-500 split.
The editorial disclaimer is Kamigame's own wording, on its New Division ranking page.
Absence of official statistics: no such data is published by Takara Tomy. This is an observation from the absence of a source rather than a stated policy.

Percentages are not reproduced here; the underlying breakdowns are published as images by their authors and are better read at source. What is quoted are the sample sizes and the methods.