← Duel Masters English

PLAY'S · Rules · 21 Oct 2026

Two cards change on 22 October, and one has led the chart on eleven of twenty days

The publisher posted the notice at 17:00 on 20 October, in the same hour as the three notices for the seventh collaboration, and it applies at the maintenance on the 22nd. One of the two gives its name to the deck that has led this site's daily readings more often than any other. The other gains power and an option. Only the first is being compensated — which, by the publisher's own stated policy, is the publisher saying which of the two is a downgrade.

First reading, 24 October — and this page still is not calling it a result The 23 October chart published this morning: 227 games, the first full day under the changed card. Two things stand in it.

is not a named slice on it. That deck had appeared on every one of the twenty-nine days this site has recorded, 24 September to 22 October, without missing one. Below the survey's cutoff a deck sits in the unnamed bucket — 13.4 that day — rather than being absent from the format, so this is a disappearance from the chart and not from the game.

, the build the chart names without that card, is second at 11.9. Across nine earlier days its best was 4.3.

That is one day. The standing example is in week four: Jokers read 11.4, then 26.5, then 14.3, then 22.7, then 20.8 across five days with nothing altered in the game at all. The test set here is whether is still unnamed on the 24th and the 25th and the Raid-less build still sits above 8. If that fails, this note stays as the record that it failed.

25 October — day one of the two holds. The 24 October chart: 146 games, thirteen named slices, and is not among them for a second day. reads 11.6, against 11.9 the day before and a best of 4.3 across the nine days before that. Both conditions pass. The second day decides it, and the 25th's chart publishes on the morning of the 26th.

26 October — the test is met. The 25 October chart is the largest sample since the change at 290 games, and is not a named slice on it either. That is three days running. reads 16.6 and is tied for first with ; across the three days it has gone 11.9, 11.6, 16.6, against a best of 3.6 on the eight days it appeared before the change.

Corrected, 26 October. This paragraph first said “a best of 4.3 on the nine days it appeared before the change”. The 4.3 is from 22 October, which is after the change — play resumed at 16:00 that day with the new card text, as the note below this one says. Eight of those nine days fall before the change and their best is 3.6. The 4.3 is the weaker figure in any case: on a 46-game day it is two games, where 3.6 on 17 October is six games out of 166. The three days are read against the whole series on the data page of 26 October.

What that is and is not. The deck was a named slice on 29 of the 32 days this site has recorded, every one from 24 September to 22 October, and the three it is missing are the three full days since its card changed. A streak ending exactly there, with the build that omits the card tripling in the same window across samples of 227, 146 and 290, is about as clean as this survey gets. It is still not a verdict on whether the change was right, which would need a reason the publisher has never given — its notice says balance adjustment and names no deck.
Resolved, 24 October The maintenance ended at 16:00 on 22 October, having run from 23:15 the night before — the emergency maintenance ran on into the scheduled one, and finished an hour late. These changes are live.

The reasoning below about which day is the first clean one holds, and the overrun does not change it: play on the 22nd began at 16:00, so that day is a part-day under the new text and 23 October is the first full one. Its chart published on the morning of the 24th and is in the note above. The publisher is giving five card pack tickets as an apology for the outage, claimable until 23:59 on 24 October.
Update, 02:00 on 22 October — when these apply has moved An emergency maintenance began at 23:15 on 21 October with no end time given, for what the publisher states is system security measures. Its notice says the 10:00–15:00 maintenance on the 22nd may be run on continuously from it, depending on the situation.

The changes below are applied at that maintenance, so the time they take effect is no longer fixed, and the reasoning further down about which day is the first clean one moves with it. The publisher's website was also unreachable at 02:00, so none of this is documented on a citable page yet.
What this page covers, and when Written on 21 October from the card-ability notice of 20 October and the maintenance notice of the 19th. The change has not happened yet and nothing here has been seen in the running game. Abilities are described in this site's own words rather than reproduced.

The two changes

“B-Reishu” Raid · DMPP-36 · Super Rare · PLAY'S original Before

Arrives with Speed Attacker outright. While your hand is one card or fewer, its keeps every Fire creature you control of cost 7 or less from being destroyed, so long as its power is above zero.

After

Speed Attacker only while three or more Fire creatures sit in your mana zone. The protection keeps only this creature from being destroyed, on the same hand-size and power conditions.

Unchanged

2, Double Breaker, and the shield burn when it arrives or when the opponent casts a spell without paying for it.

Kaya no Flam Arte · DMPP-35, March 2026 · Super Rare · PLAY'S original Before

On arrival or on attack it puts the top card of your deck into your mana zone, and if that card is a Twinpact it must go straight into play. It gives your Twinpact creatures Mach Fighter.

After

The same mana placement, but putting the Twinpact into play is now your choice. It gives your Twinpact creatures Mach Fighter and +2000 power.

What the Raid change removes

Two things, and they are not the same size.

Speed Attacker stops being a property of the card and becomes a condition on your mana zone. Before, the Raid attacked the turn it landed, always. After, it does that when three or more Fire creatures are sitting in your mana — a state of the board rather than a line on the card.

The protection stops covering the board and covers only the card. This is the quieter of the two and the easier to misread, so it is worth being exact: the clause that named your other Fire creatures is gone from the new text, and nothing replaces it.

The correction this site had to make in September still stands The page written on this card in September said in its first version that the cheap copy would destroy itself at end of turn and could simply be waited out. That was wrong, and was corrected on 12 September: the protection stopped it, because the Raid is itself a Fire creature of cost 7.

The new wording does not obviously change that. It drops the clause naming other creatures and leaves the protection with no stated subject, which in this game's card text means the creature carrying it. On that reading the Raid still protects itself, and the cheap copy still does not die on its own.

That reading is this site's, taken from the Japanese, and the change has not happened yet. Check it in the running game after the 22nd before relying on it. The September error came from reading one ability at a time; this one would come from reading one clause at a time.

The compensation says which change is which

The notice restates the publisher's standing policy: it collects and analyses match data, changes cards when it judges it needs to, and where a change is a downward revision it compensates — by altering the DM points the card disassembles for, or by handing out items.

Compensation here is attached to one card only. Disassembling the Raid pays 2400 DM points instead of 600 for a normal copy and 1700 for a premium one, from the 22 October maintenance to the 26 November maintenance, reverting afterwards. The Sec version can be disassembled in both its normal and premium forms but cannot be generated, so a Sec copy cashed in cannot be got back.

Nothing is paid for Kaya no Flam Arte. By the publisher's own rule, that is it saying the Flam Arte change is not a downgrade — which matches what the change does: an option where there was an obligation, and power on top.

Players were asking for this in September, and the notice says nothing about that

At Battle Arena 2026 Summer in July, the Raid deck was submitted by over 30% of entrants in both preliminary rounds — nearly one in three at the game's flagship tournament. That figure is the publisher's own, from the deck distribution it published on 18 July.

This site recorded a post from 11 September putting the two together: that the card should have been nerfed or restricted on the strength of that submission rate. That is what one player wrote, paraphrased and anonymised at the time, and it is reported here as that and nothing more.

The notice gives no reason. It says the change is for balance adjustment and restates that the publisher collects and analyses match data. It names no deck, no tournament and no figure. Six weeks sit between the two, and nothing in the notice connects them — so this page does not connect them either.

Where this lands

This site reads the daily ranked survey and records it. Against that series, as of the last reading published:

DayGamesChart leader
16 Oct13018.5
17 Oct16624.1
18 Oct14715.0 26.5
19 Oct848.3 15.5
20 Oct11018.2 22.7

The five most recent readings. 20 October is the day the notice went out, so its play falls on both sides of the 17:00 post; the chart for it published on the morning of the 21st. Sample sizes are small — at 84 games a single percentage point is under one game — and all five fall inside the ranked season that opened at 00:00 on the 16th, so none of them is comparable to a day before that.

What this page is not saying

That the change worked, or did not. There is now exactly one reading from a game that has it, and it is in the note at the top of this page; one day is not a result, whichever way it points.

And the first one will not be clean. The maintenance runs from 10:00 to 15:00 on the 22nd, so play that day happens on both sides of the change and the survey for it cannot separate them. The first full day under the new text is 23 October, and its chart publishes on the morning of the 24th.

Even then, one day is one day. read 11.4, then 26.5, then 14.3, then 22.7 across 17 to 20 October — four readings, no two alike, with nothing changed in the game at all. A deck is not a card, and a week is the shortest thing this site has been willing to call a move.

It went out in the collaboration's shadow

All four of the day's notices carry the same timestamp, 17:00 on 20 October. Three of them are the collaboration's — the set, the deck builders, the Legend Battle — and this is the fourth. On X the two card posts went up at 17:00 and 17:01.

September had the same collision and the opposite result: a change shared its maintenance with a new set, an alternate win condition and a rotation, and the changes that mattered in it were All Division ones — the format that was not the story. This one is not that. is in New Division, which is where every number above comes from.

Sources The change list, the before and after text for both cards, the compensation figures and window, and the statement of the publisher's compensation policy: the news post of 20 October 2026, . Abilities are described here rather than reproduced. The maintenance window of 10:00 to 15:00 on 22 October: the maintenance notice of 19 October.

That is a PLAY'S original with no paper printing, from DMPP-35 of March 2026 at Super Rare with a Sec version: the Duel Masters Wiki's pages for the card and for the set, which agree on both. The profile of is this site's own, from September, as are the Battle Arena submission rate — originally from the official deck distribution of 18 July 2026 — and the player post of 11 September, which was paraphrased and anonymised there and is not quoted here.

Deck readings are from this site's record of the daily ranked survey, which is read off published charts by hand. Kaya no Flam Arte is this site's romanisation, not a name taken from the community translation patch, which cannot be consulted from here; the Japanese name is the one to search on. A name rendered this way has been wrong once before and was corrected the next day.