PG Soft confirms 96.95 percent on Mahjong Ways 2 and 96.92 percent on the original. Neither page states a ways count, a payline total, or a volatility label.
Published 2026-09-17Not changed since publicationNo review date set yetSources not re-read since publication
PG Soft's own page for Mahjong Ways 2 states a return of 96.95 percent. 96.95%◦ That same page names a maximum win multiplier of 100,000 times the stake. 100,000×◦ Read directly today, the page's own embedded data block still lists exactly those two figures, unchanged since this site last checked it on 2026-08-14, down to the same field names PG Soft has used since that first check.
The page also states the grid's physical shape, in one sentence: five reels, four rows, with one extra row layered onto reels two, three and four. That sentence is the entire layout statement PG Soft gives a reader. No line count follows it. No ways count follows it either, despite a title built specifically around that word.
The page's own description names one feature the original never had. During the Free Spins Feature, every symbol on reel three, other than the wild and the scatter, turns into a golden Mahjong tile, and any golden tile that lands in a winning line converts to a wild before the round pays out. The same feature carries a multiplier that climbs to ten times as the free spins continue. Both numbers come from the studio's own wording, not from a third-party summary of it.
Mahjong Ways, PG Soft's 2020 launch of the format, states a return of 96.92 percent. 96.92%◦ Its maximum win sits lower than the sequel's, at 25,000 times the stake. 25,000×◦ The grid is a plain five reels by four rows, with no irregular row anywhere on it, and no golden-tile conversion feature either. That mechanic belongs to the sequel alone, on the wording of both pages read side by side today.
The original's own free spins are stated in exact numbers PG Soft rarely repeats elsewhere on the site: three scattered symbols anywhere on the reels award twelve free spins, and every further scatter in that trigger adds two more. A climbing multiplier runs through the round, the same as the sequel's, capped at ten times.
Beyond that free-spins count, both pages name the identical three mechanics stored on this database: free spins, a wild symbol transform, and the climbing multiplier itself. Nothing further is named on either page. Both leave volatility blank. Neither states a payline count. A title built specifically around the word "ways" turns out to be the format PG Soft documents least on its own site.
Six later PG Soft titles, all built on a cascading, symbols-collapse-and-refill engine rather than the wild-transform format Mahjong Ways uses, do better, even though none of them carries "ways" in its own name. Caishen Wins and Ganesha Fortune both state a range of 2,025 to 32,400 ways on their own game-information sheets, tied to a base bet of 20. Jurassic Kingdom and Rise of Apollo both state 64 to 46,656, which happens to be the exact ceiling a fixed six-reel, six-row grid allows once every reel is filled with the smallest symbol.
Leprechaun Riches states the same fixed grid and a range of 576 to 46,656. Cocktail Nights states 400 to 15,625, matching its irregular six-reel, five-row grid with an extra row layered onto four of the reels, the same layout shape PG Soft uses on Mahjong Ways 2 but applied across more reels.
None of the six resolves its range to one number, because the true count on any given spin depends on how many oversized symbols, ones that occupy two, three or four symbol spaces at once, happen to land. PG Soft states a floor and a ceiling on each sheet and leaves the arithmetic to the reader.
A range is still a published figure, sourced and checkable against the studio's own page. Mahjong Ways and Mahjong Ways 2 do not offer even that much, on either title's page, on either date this site has checked it.
The obvious objection is timing, not branding. PG Soft may simply have started documenting newer releases more fully as its own reporting practice matured, and the six cascading titles named above could just be recent enough to inherit that habit rather than any decision tied to the word "ways" itself. Checking that objection against a release date would settle it cleanly.
It cannot be checked that way, because PG Soft does not publish a release date on most of these pages either. Five of the six comparison titles carry no wide-release date in this record at all. The one that does, Caishen Wins, carries a launch date of 2020-04-08, the same year Mahjong Ways itself launched.
That single date weakens rather than confirms a simple story where only the newer titles get full documentation. What can be said is narrower: on this site's own catalogue, the two titles built specifically around the word "ways" sit among the thinnest pages held for the studio, and a name cannot state a mechanic its own page never states, whatever the reason turns out to be.
Three things are missing here and stay missing rather than get filled in. PG Soft names BMM and GA as its accreditors across its whole catalogue, but publishes no certificate specific to either Mahjong Ways page. Neither page states a volatility label, so this record carries none rather than inferring one from the max win alone. And Mahjong Ways 2's own page carries no dated release field at all, only an internal record timestamp this site does not treat as a launch date.
Every slot record on this database follows the same rule: a figure the studio has not published does not appear here as a guess. That rule covers every slot this site has read a provider's own page for, not only these two. [the casino brands on this database](/casinos) is where to check which operators actually carry either title today, a separate question from what the studio's own page says about it.
If PG Soft adds a ways count to either page tomorrow, this record changes from that page, and not from a rewrite.
PG Soft's own page states 96.95 percent, confirmed directly from the page's embedded data as of today. The original Mahjong Ways states a slightly lower 96.92 percent on the same basis.
No. The page states the grid shape, five reels and four rows with an extra row on three of them, but names no ways total and no payline figure. The original Mahjong Ways does not publish one either.
PG Soft states 100,000 times the stake for Mahjong Ways 2. The original Mahjong Ways states a lower ceiling of 25,000 times the stake.
Both pages name the identical three mechanics: a wild symbol transform, a climbing free-spins multiplier, and nothing beyond those two. The sequel adds an irregular grid and a second theme, but no new mechanic and no ways figure the original was missing.
Caishen Wins and Ganesha Fortune both state 2,025 to 32,400. Jurassic Kingdom and Rise of Apollo both state 64 to 46,656. Leprechaun Riches states 576 to 46,656, and Cocktail Nights states 400 to 15,625.
Not a bibliography — these are the addresses behind the numbers this site stores about the 8 records this piece is about, with the date we read each one. Anything the article argues from that we have not stored as a figure is not in this list, and we would rather say that than pad it.
Nothing on this page has been corrected since it was published on 2026-09-17. When something is, it is listed here with the date and what changed — not quietly edited. How we handle a correction.
PAGCOR-licensed casinos rarely lock only the bonus a player claims. The wagering requirement usually covers the whole balance, deposit included, until the combined total clears the multiple.
Mental and Fire in the Hole xBomb are both Nolimit City titles with six-figure win caps. Only one published number explains why one pays its cap roughly seven times more often than the other.
Pragmatic Play states one RTP for Gates of Olympus and its sequel, 96.50 percent on both game pages, but the sequel's stated 15,000x maximum win lives only in a 2023 launch article, not on either spec page.
Bangko Sentral's rail caps every InstaPay transfer at ₱50,000, but the number that actually stops a Philippine casino deposit is a second ceiling each bank sets on its own, and it is rarely the same figure twice.
