Order of the Sinking Star vs The Witness and Braid
How Order of the Sinking Star compares to Jonathan Blow's earlier games, The Witness (2016) and Braid (2008), from puzzle count and structure to design method.
FAQ
Is Order of the Sinking Star a sequel to The Witness?
How many puzzles does Order of the Sinking Star have compared with The Witness?
Is Order of the Sinking Star linear like Braid?
Does Order of the Sinking Star use the same one-rule-set approach as The Witness?
How long is Order of the Sinking Star?
Jonathan Blow has now directed three games, and the differences between them say more about his method than any through-line does. Braid was a compact time-manipulation platformer. The Witness was an island built from a single visual language. Order of the Sinking Star is something else again: four worlds, several designers’ systems, and a hub that never tells you where to go.
Quick Answer
The comparison comes down to structure and method. Braid (2008) and The Witness (2016) were both tightly unified — one idea, expanded patiently. Order of the Sinking Star takes four worlds built on other puzzle designers’ foundations, lets each teach its own rules, then collides them in the Gold Rooms. It is much larger, openly nonlinear, and deliberately less unified. That is the trade.
| Braid (2008) | The Witness (2016) | Order of the Sinking Star (2026) | |
|---|---|---|---|
| Form | 2D side-scrolling platformer | 3D first-person island | 3D top-down grid puzzle adventure |
| Core idea | Time manipulation as a mechanic | One panel language, endlessly varied | Four separate worlds that later combine |
| Structure | Discrete stages | One island, gated by understanding | Free-roam hub open from the first minute |
| Puzzle count | Small, stage-based | A much smaller single-rule-set catalogue | “Over 1000” in official store copy; Blow has claimed more than 1400 |
| Design lineage | Original | Original | Built on four foundation games by other designers |
| Stated length | Short | Tens of hours | Blow’s own estimates: up to 250 hours normally, past 500 for full completion |
Official source: Order of the Sinking Star on Steam
What the earlier games actually were
Braid arrived in 2008 and became the reference point for a certain kind of indie puzzle-platformer: hand-painted, melancholy, and built around a single mechanic — control over time — that the game re-examined from a new angle in each world. Its puzzle count was modest and its stages were discrete. You finished a stage or you did not.
The Witness arrived in 2016 and went the other way. It put the player on one island and gave them one interaction: drawing lines on panels. Everything else — the mazes, the audio logs, the environmental statues, the way the island’s layout itself encodes rules — was variation on that single language. Its catalogue of puzzles was substantially smaller than what Order of the Sinking Star claims, and its ambition was depth rather than breadth.
Order of the Sinking Star keeps the rule-discovery interest and drops the unification. It is the third game Blow has directed, and it has been reported as spending roughly a decade in development — a span that shows in the sheer volume of systems.
Puzzle count: 1000+ versus a smaller, unified catalogue
The official store copy for Order of the Sinking Star still says “over 1000 interconnected, hand-crafted puzzles”. Blow has separately claimed more than 1400 puzzles. Those are two different figures from two different sources, and neither is a measured count: the store text has not been updated to match the director’s claim, and the director’s claim has not been reflected back into the store listing. The demo alone is reported by players at 105 puzzles, which gives a sense of how the full set is distributed across four worlds and the shared Overworld.
The Witness, by contrast, was built around one mechanic explored exhaustively. Its puzzle catalogue was much smaller, and that was a design choice rather than a limitation: with a single rule set, the game could teach one language and then spend its entire length making that language mean different things.
The relevant question is not which number is bigger. It is what the extra puzzles buy. In a four-world game with combining mechanics, more puzzles means more room to teach each world’s verb properly before the collision, and more surface area for the combination phase afterwards.
[!note] The two counts do not conflict so much as disagree. One is marketing copy that has sat unchanged, the other is a statement the director has made in interviews. Neither is a measured, final count, and both were the current public position as of September 24, 2026.
One island versus four worlds that collide
The Witness is one island. Its map is a single continuous place, and progress is measured by understanding rather than by unlocks in the conventional sense.
Order of the Sinking Star opens as what the developers describe as an anthology of separate worlds. The Overworld is the hub, and it is fully explorable from the beginning of the game — the four routes branch off in four directions, and the map fog recedes as you make progress rather than as you complete a required sequence.
Then the structure changes. Reaching the Gold Rooms triggers what the official material calls a violent structural shift that makes the worlds collide. Clearing one Gold Room bridges two different realms, brings their characters face to face, and opens one of six large dual-combination regions. The developers have named two of those pairings: the Explorer’s light beams crossed with the Trader’s magic mirrors, and the moving company crossed with the Sailor’s stone-skipping.
The full pairing list for all six Gold Rooms has not been published. What is public is that the collision is the point: the worlds were built to be combined, and the second half of the game is about those combinations rather than about any single world’s rules.
The supercollider method
The clearest break from the earlier games is how Order of the Sinking Star was designed. Rather than inventing one core mechanic and expanding it, Thekla built what Blow has described as a game design supercollider: they took the foundations of four existing puzzle games by other designers and collided those ideas, producing what he calls a combinatoric explosion of possibilities.
The four foundation games, as described in the official material:
| Foundation game | Designer | Official description |
|---|---|---|
| Skipping Stones to Lonely Homes | Alan Hazelden | A serene puzzle game about throwing stones into water to create stepping paths |
| Mirror Isles | Alan Hazelden | A clever top-down game focused on swapping places with reflections across islands |
| Promesst 1-3 | Sean Barrett | An abstract grid-puzzler about exploring a cave filled with mysterious beams of light |
| Heroes of Sokoban 1 / 2 / 3 | Jonah Ostroff | A fantasy-themed puzzle game where you manage distinct hero classes with unique movement and pushing mechanics to reach the exit |
Additional design contributions came from Zach Polansky (Enigmash), Patrick Traynor (Patrick’s Parabox) and Marc ten Bosch (Miegakure), along with several others. The world themes correspond naturally to those foundations, but the official material does not state a one-to-one mapping, so it is safest to describe each world as built on or inspired by one of them.
Braid and The Witness were both single-author visions in the sense that mattered: one designer’s rule set, extended. Order of the Sinking Star is explicitly a collision of several designers’ rule sets. That is a different kind of authorship, and it shows in how the puzzles feel — more mechanical variety, less insistence that everything reduce to one idea.
Nonlinear free-roam versus sequence
Braid’s stages were discrete and ordered by the player’s own choice of world. The Witness gave you an island but still funnelled understanding through a rough progression of areas.
Order of the Sinking Star is the most permissive of the three by design. The Overworld is free from the start, there is no single correct route, and a player who is stuck can leave a world, cross the hub, and work elsewhere. The design intent is that the collision phase rewards you for having sampled broadly, so wandering is not a failure of discipline.
[!tip] If you bounced off The Witness because a single stuck panel could halt the whole island, this is the structural difference that matters most. Here, a stuck board is a local problem, not a global one.
Length: the 250-hour and 500-hour figures
Both headline numbers are Jonathan Blow’s own estimates, given to IGN on December 11, 2025. He put a normal pace, without full completion, at up to 250 hours, and full completion past 500 hours. The official FAQ’s line about more than 500 hours of gameplay awaiting the most dedicated puzzle solvers restates that same estimate rather than measuring anything independently.
That distinction matters when you compare the three games. These are not measured playtimes, not averages drawn from a released game, and not review data. They are the director’s expectations for a game that had not shipped when he gave them, recorded here as of September 24, 2026.
For comparison with the earlier games, treat the scale gap as the headline. Braid and The Witness are games you finish in a sitting-spanning stretch; Order of the Sinking Star is presented as a game you live in for months. Crucially, the story does not require solving everything — you can advance without clearing the full set — so the 250-hour and 500-hour estimates describe two different kinds of player rather than two difficulty settings.
What is not a fair comparison
No aggregate review score, Metacritic figure or awards record exists for Order of the Sinking Star as of September 24, 2026, because the game is not out until October 8. Any comparison that ranks the three games by review reception is therefore premature.
It is also worth being precise about what Braid and The Witness were. Braid’s reputation rests partly on how it used a platformer’s language to talk about regret and memory, which is not a puzzle-count achievement. The Witness’s reputation rests on the discipline of refusing to add a second mechanic. Order of the Sinking Star is attempting neither of those things; it is attempting scale, variety and synthesis. Whether that succeeds is a question the launch will answer, not the pre-launch material.
Related
- Beginner tips and first hour — how to approach the free-roam hub
- Overworld map guide — the structure that replaces The Witness’s single island
- Gold Rooms explained — the collision point that has no equivalent in the earlier games
- How many puzzles and how long to beat — the 250-hour and 500-hour figures in context
- Achievements guide — what is and is not public before launch
This page is re-checked at launch, when review reception and final puzzle counts become verifiable.
Related Articles
guides
Is Order of the Sinking Star on Xbox? No, and Here Is Why
No. Order of the Sinking Star is not on Xbox or Game Pass, and there is no Switch 1, PS4 or Epic version either. It launches October 8, 2026 on PC, PS5 and Switch 2.
guides
Order of the Sinking Star Achievements Guide
Steam confirms Order of the Sinking Star has achievements, but no list is public before launch. What is confirmed, what is not, and what to expect on October 8.
guides
Order of the Sinking Star Beginner Tips: Your First Hour
What a new player needs in the first hour of Order of the Sinking Star: how the grid works, which world to start in, how to read the art, and how to avoid burnout.
Was this helpful?
Thanks for the feedback!