PLAY'S · META · 10 OCT 2026

Week two: one deck led on six days out of seven

In the format's first week four different decks took the daily lead, and none held it more than three days. In the second week one deck took it six times, five of them running. On all seven days the leading named deck was also the largest slice of the chart — in week one the survey's own unnamed bucket beat the leader twice. The week's clearest riser is a deck that was not in the six this site had been tracking.

The lead stopped moving

The one day the lead changed hands is marked. "Named" matters: the survey's unnamed bucket is excluded from this column, as it was on the nine-day page.

Week one, 24 to 30 September, went Magic Tools, Magic Tools, Magic Tools, Raid, Cotton Zenith, Jokers, Raid — four decks, longest run three days. Week two is Raid on 1 October, Jokers on the 2nd, then Raid on every day from the 3rd to the 7th. Five days running is the longest any deck has held the lead since the format opened.

The two weeks are close in size: 1,751 games in week one against 1,668 in week two. So the difference is not that there is more data now.

The one day it did not, and it was one game

On 2 October Jokers read 17.7% and Raid 16.9%. That survey had 124 games, the smallest day of the week, where one game is worth about 0.8 of a percentage point. At that size the two figures are twenty-two games and twenty-one. The lead changing hands on 2 October and not changing hands on any other day of the week are not findings of the same weight, and this page does not treat them as such.

What the tracked decks did

Percentages of games encountered, as the daily chart gives them. Baraghiara and Magic Tools are family totals — see the two sections below for why. The nine-day page called the Baraghiara column "Loop"; it is the same family, now with a second entry added to it. Rows do not sum to 100 because the smaller named decks are not columns here.

The riser was not one of the six

HADES Dragerion HADESドラゲリオン was not a column on the nine-day page. Over this week it went from 4.9% to 9.9%, and its two highest readings of the week are its last two days, 9.8% on 6 October and 9.9% on the 7th. Its low was 2.4% on 2 October, which on 124 games is three games, so the shape of the rise is firmer than its starting point.

By 7 October it is the fourth-largest named entry on the chart, behind Raid, the Baraghiara family and Jokers. That is a change in what this site should be tracking, and the column has been added here rather than left out because it was not there before.

Cotton Zenith rose for four days and then gave it back

4.9, 5.6, 9.5, 11.2, 13.1 — four consecutive rises from 1 to 5 October. 13.1% is its highest reading since 28 September, when it read 13.9% and led the format; it is not a figure the deck had never reached. Then 7.7% and 8.1%, which leaves it ending the week below its 3 October reading of 9.5%. Five days of it looking like the story of the format, and then not.

Magic Tools will not sit still, and some of that is counting

9.1, 12.1, 4.6, 6.7, 3.8, 13.3, 6.2. That is the widest spread of any deck column here — a 9.5-point gap between its best and worst days, a 9.5-point jump on 6 October and a 7.1-point fall the day after. Only the unnamed bucket moves more, and that column is not a deck.

Baraghiara: read the family, not the line

From 4 October the chart carries two Baraghiara entries, the loop and キャベッジバラギアラ. The second one appears for the first time on that day. On the loop line alone, 3 to 4 October is 17.5% down to 11.2%, a fall of 6.3. Taken as a family it is 17.5 to 15.4, a fall of 2.1. Most of the apparent collapse is the new entry being carved out of the old one.

Over the full week the family's spread is 8.4 points, wider than Raid's, so it is not a steady deck by that measure. But the width is almost entirely its first day: from 2 October on it reads 15.3, 17.5, 15.4, 14.3, 14.1, 16.9, a band of 3.4 points. Its 9.1% on 1 October is the outlier, not its later readings.

Jokers barely moved at all

13.2% to 17.7%, a spread of 4.5 points, the narrowest of any column here. Counting Baraghiara as one family, it took the lead once, on 2 October; across the other six days it was second twice, third three times and fourth once. It was never below 13.2% and never above 17.7%. A deck that neither runs away nor collapses is a different kind of fact from a deck that spikes, and the daily figures do not say which is the better deck.

The unnamed bucket, and what does not explain it

"Other" reads 17.7, 4.3, 15.8, 15.1, 11.0, 14.2, 13.7. The 4.3% on 2 October is the odd one. The nine-day page looked at it and could only account for part of it: two decks appeared on the chart that day worth 7.2 points of the 13.4-point fall, and it said plainly that the remaining 6.2 points were not explained by the bucket count. They still are not.

The tempting rule is that the bucket shrinks when the survey names more decks. It does not hold. Counting the unnamed bucket itself, the chart had 11, 13, 10, 11, 12, 12 and 13 slices across the week. 2 October and 7 October both had thirteen, and the bucket read 4.3% on one and 13.7% on the other. The bucket count is worth knowing before reading a day-to-day move, but it does not predict the bucket's size.

What else was going on

The Legend Battle ran through the whole of this week, from 1 October at 17:00, and it is solo play rather than ranked. A twenty-four-issue bug fix went live at about 18:30 on 1 October, and a further ten-issue fix at about 16:20 on the 7th. The Minogami Cup, a separate in-game tournament mode, started on 8 October and so falls outside this window.

What this page is not

It is not a tier list and it does not rank decks by strength. An encounter tally counts what people queued into; how good a deck is, is a different question and a different kind of claim. The daily samples here run from 124 to 325 games, so the small figures should be read as approximate, and a one-day move of two or three points on a small day may be a handful of games. No decklists, rankings or tier tables from other sites are reproduced here.

Sources

Daily New Division encounter surveys for 1–7 October 2026, published each following morning at beans-duelplays.com as image charts; read from the charts on 10 October 2026. Game counts are the survey's own.

Event dates and the bug-fix timings: the publisher's announcement of the Legend Battle, 1 October 2026, and its running known-issues notice as updated 1 and 7 October 2026. The event calendar on this site carries both: what is running and when.

Deck names are given in Japanese where this site has not confirmed an English name.