Released in August 2008, Braid is a puzzle platform game designed by Jonathan Blow and developed by Number None. It debuted on Xbox Live Arcade for Xbox 360 and later expanded to PC and other platforms, which helped it become a common reference point for the late-2000s indie wave. A modern rerelease, Braid, Anniversary Edition, brought the game to a wide set of current systems while keeping the identity of the original intact.
Braid speedruns work because the game’s signature mechanic, time manipulation, turns mistakes into information rather than failure. Most platforming risks can be tested aggressively, then rewound cleanly, which encourages runners to push movement lines and puzzle solutions without the usual “death tax” found in many platformers. The run itself is a chain of short stages inside themed worlds, and the central optimization problem is routing: choosing a fast path through levels while gathering the required puzzle progress to unlock the ending sequence.
What makes Braid distinct as a speedgame is that the route is not only about raw execution, but also about controlling causality. Time powers change from world to world, so the fastest solutions often come from understanding how objects, player position, and time states interact, then turning that understanding into repeatable patterns. The game also includes built-in time trial style “speed runs” after completion, which has historically reinforced the idea of replaying levels for cleaner solutions and quicker clears.
From the start, Braid’s earliest “speedrun culture” grew out of the fact that the game itself treats speed as a formal challenge. After finishing the story once, the pause menu opens a dedicated Speed Runs mode built around six preset challenges, including a Full Game target time of 45:00, and those challenges are framed around clean execution and puzzle familiarity rather than pure movement tech. The game also bakes in key timing behaviors that shaped early routing habits, like pausing the clock during certain non-platforming moments and letting runners breathe during the jigsaw screens between worlds.
Because those rules live inside the game, some of the earliest route knowledge spread the old fashioned way: players talking it out on message boards, comparing where the clock does and does not tick, and building practical “safe but fast” solutions for rooms that could punish even small mistakes. Discussions from the release era already show runners thinking in terms of timer management, noting that the clock stops while assembling the between-world puzzle and continues all the way through the Epilogue exit, which effectively defines what “counts” in a full game attempt.
As soon as full runs started getting recorded and shared more widely, speedrunning Braid began to look like a modern scene: videos, forum uploads, and archived runs that preserved both the route ideas and the quirks of the game’s timer. A notable early example hosted by Speed Demos Archive explicitly calls out a core timing norm that would stick around: the in game timer can read noticeably shorter than the actual recording because it does not count time spent in certain between-world puzzle segments, and runners were already comparing results across different release ecosystems like Xbox Live Arcade and PSN listings.
Over time, as communities centralized on Speedrun.com, the “standard goals” became clearer and more leaderboard-friendly: full game categories such as Any% and 100%, with additional variants that focus on specific completion landmarks like stars. That consolidation also reinforced the game’s long running preference for using the in game timer as the common language for comparing runs.
Later, re-releases created natural version splits without rewriting the core identity of the run. Braid, Anniversary Edition is tracked separately, with its own leaderboard structure and category set, which lets the community preserve the older routes while giving the updated release its own standards.
As Braid speedrunning matured, the community’s public footprint consolidated around Speedrun.com, which functions as the scene’s shared record. The game page gathers the core pieces in one place: full-game leaderboards, level boards, a forum for rules and community discussion, and a library of guides and resources, with a prominent link out to the game’s community chat. A parallel page for the rerelease keeps its own leaderboards and discussion spaces, which helps the broader community keep standards clear across versions without forcing everything into one combined board.
Day to day conversation tends to live off the leaderboard itself. Older scenes often began on general gaming forums and early video uploads, but the modern shape is more centralized: most coordination happens in Discord, with wider visibility and casual questions spilling over into places like Reddit, Twitch chats during practice streams, and tutorial or showcase uploads on YouTube. Speedrun.com explicitly treats game-linked Discord servers as part of the same community ecosystem, with expectations tied to the site’s code of conduct.
The culture that forms around that structure is practical and documentation-heavy. Knowledge gets preserved through routing writeups, autosplitter setup help, and technique tutorials that live in the game’s guide index, so newer runners can learn the shared “standard” approaches without having to rediscover them from scratch. Submissions follow the same general rhythm as most Speedrun.com boards: runners submit a time with video, and moderators verify by checking that the run meets the category rules and that the timing is supported by what’s on-screen. The site also sets expectations for patience and process, including guidance to wait before escalating questions and a general window for moderators to handle submissions, which keeps verification from turning into a constant conflict point in smaller communities.
On Speedrun.com, Braid’s full game leaderboard is built around completion thresholds rather than a long list of niche variants. The core set is Any%, 6 Stars, and 100%, with an additional miscellaneous board for 30 puzzle pieces that captures a more collection focused style of run. A separate leaderboard exists for Braid, Anniversary Edition, and that board keeps the same overall shape while making the star-count goals more explicit by listing 6 Stars and 8 Stars alongside Any% and 100%.
In terms of what is actually being measured, the public boards surface in-game time as the key number, and the leaderboard layout shows an IGT column and a separate Time column for context. Individual run pages reinforce that standard by labeling the recorded value as In-Game Time, and submissions are moderator-verified with evidence, which can include specific IGT proof. In practice, that means most “rule” debates that matter to readers boil down to what is allowed under the category’s ruleset and what the moderators will accept as clear proof, rather than disputes over manual timing or load removal.
Platform policy is permissive in the sense that the boards track multiple platforms and label them directly in the leaderboard and run details. The original release’s page lists PS3, X360, and PC, while the Anniversary Edition page lists PS4, PC, Switch, plus additional platforms. Beyond full-game runs, the community also maintains an individual level structure, and a published guide points runners to an external IL leaderboard and asks submitters to coordinate through the community hub. For runners using a timer, another guide documents an autosplitter workflow for the most common timing setup, which helps keep splits and resets consistent across attempts.
Braid’s speedrun strategy gradually shifted from “solve the puzzles quickly” into “solve them in a way that preserves momentum and minimizes downtime.” The most stable routes lean on repeatable room patterns, fast exits, and clean transitions between puzzle pieces, because each world’s time rules reward runners who can commit to a plan and use rewinds as controlled setup rather than as recovery. That evolution is visible in how the community archives knowledge now, with tutorials, advanced technique notes, segmented demonstrations, and tool assisted references all treated as legitimate ways to study the route and tighten execution.
Movement tech became a second layer of optimization once runners started treating the game’s physics as something to exploit, not just something to tolerate. Ladder jumps and corner interactions are a common example, because they turn ordinary platforming moments into chances to gain extra height or forward speed when timed precisely. At the higher end, tool assisted research has documented how Braid’s collision geometry can create invisible “booster” points at surface intersections that fling Tim left or right, and mapping those boosters became its own kind of community craft because they are not obvious from normal play.
The game’s most distinctive advances come from treating time mechanics as programmable logic. Runners developed setups that deliberately build multiple timelines, then rewind and fast forward to make earlier actions happen while Tim’s current position stays advantageous. A well known example in community documentation is “Key Magic” in the level Elevator Action, where parallel timeline behavior is used to stage lever pulls, key movement, and door access across separate rewinds so the room collapses into a compact sequence instead of a slow puzzle.
Glitches and edge case mechanics added another layer, especially when they interact with rewinds and pauses. TASVideos documentation describes “Springload,” a world 4 only behavior where a specific bounce and rewind timing can accumulate unusual horizontal velocity, turning an enemy bounce into a burst of speed that changes how certain rooms can be approached. The important point for the human scene is not the exact frame story, but the pattern: Braid’s rewinds can store and release motion in surprising ways, so new discoveries tend to look like new ways to convert puzzle states into movement states.
As the scene became more standardized, tooling helped make practice and comparison easier. The LiveSplit autosplitter guide is part of that, because it gives runners a consistent way to track attempts and splits without reinventing setup each time. The community also maintains an individual level ecosystem through a shared spreadsheet leaderboard that feeds back into full game routing, since strong IL solutions often become the blueprint for future full run optimizations.
One of the earliest milestones for Braid as a speedgame was that it arrived with an internal framework for playing fast, not just finishing. The built-in Speed Runs mode, unlocked after a first completion, gave runners a shared starting point for what a “real attempt” looked like and normalized the idea that the game’s own clock mattered. That internal timer also reinforced a distinctive standard for the scene, because it pauses during specific between-world and menu-like moments that don’t reflect execution inside levels.
A second milestone was the shift from scattered early records and forum knowledge into a stable public archive. Older communities preserved routes through individual uploads and legacy sites, but the modern identity of the scene is anchored by a central record on Speedrun.com, where categories, moderation, guides, and resources sit side by side. When Braid’s rerelease was given its own board, it created a clean version split that protected older standards while letting the updated release develop rules and routing without constant ambiguity.
Another major milestone type was the way tool-assisted research fed back into human play. Braid’s physics and time rules are full of edge cases that are hard to “feel” in a normal run, and TASVideos documentation helped formalize what was possible by turning weird interactions into named, testable concepts. Even when a runner never uses a TAS directly, that kind of research influences how the community thinks about momentum, collision behavior, and rewind setups, which in turn shapes how routes get rewritten to be both faster and more repeatable.
Finally, marathon showcases served as watershed visibility moments that brought the game to broader speedrunning audiences and helped legitimize the run’s standards. Braid’s presence at Games Done Quick events showed how the game reads on a big stage, where explaining time mechanics and route logic is part of the performance, and those showcases tend to accelerate documentation and onboarding afterward because new viewers become new runners.
CtrlAltDelicious – Any% – IGT 22:42.480 – Jan 2026 (listed as 1 month ago) – PC (version not listed) – The current top-listed Any% run on the main leaderboard, representing the modern benchmark pace for the category.
GhERmann – Any% – IGT 22:56.570; RTA 24:41.540 – circa 2021 (listed as 5 years ago) – PC (version not listed) – A defining top-tier time from the post-early era that still anchors the front of the board.
KanBan85 – Any% – IGT 23:30.720; RTA 25:36.235 – circa 2021 (listed as 5 years ago) – PC (version not listed) – A high-water-mark run from one of the scene’s central organizers and competitors.
P-why – Any% – IGT 24:14.430; RTA 26:26 – circa 2018 (listed as 8 years ago) – PC (version not listed) – A rivalry-era run whose description captures the back-and-forth chase for tiny improvements and the feeling of “getting WR back.”
AngryMagikarp – Any% – IGT 24:24.780 – circa 2015 (listed as 11 years ago) – PC (version not listed) – An older front-page time that reflects how established the classic Any% route became as the community matured.
BakaRecca – Any% – IGT 26:08 – circa 2016 (listed as 10 years ago) – PS3 (version not listed) – A notable console entry that also shows the archival fragility of older runs when original video links disappear.
AlexGanonSm64 – Any% – IGT 26:10.080 – circa 2021 (listed as 5 years ago) – X360 (version not listed) – A strong Xbox 360 time that helps illustrate the category’s multi-platform presence beyond PC.
GhERmann – 6 Stars – IGT 25:38.020; RTA 27:54.703 – circa 2022 (listed as 4 years ago) – PC (version not listed) – The current top-listed 6 Stars run, representing the fastest execution of the game’s “minimal ending collection” route family.
KanBan85 – 6 Stars – IGT 26:16.900 – circa 2021 (listed as 5 years ago) – PC (version not listed) – A marquee run explicitly tied to Awesome Games Done Quick practice, showing how marathon prep and routing discipline feed into leaderboard performance.
GhERmann – 100% – IGT 2:12:39.370; RTA 2:15:04.670 – circa 2022 (listed as 4 years ago) – PC (version not listed) – The current top-listed full completion run, reflecting the optimized “everything” line through the worlds.
Bum – 100% – IGT 2:20:15 – circa 2014 (listed as 12 years ago) – PC (version not listed) – A historically valuable older full-run VOD that still functions as part of the public record for how 100% used to be routed and executed.
Thekla, Inc. “Braid, Anniversary Edition.” Accessed February 2, 2026. https://braid-game.com/
Valve Corporation. “Braid, Anniversary Edition” (Steam store page). Accessed February 2, 2026. https://store.steampowered.com/app/499180/Braid_Anniversary_Edition/
Valve Corporation. “Update notes for May 17th.” Steam Community. May 17, 2024. Accessed February 2, 2026. https://steamcommunity.com/ogg/499180/announcements/detail/6215608421523930333
Valve Corporation. “Update Notes for June 3rd.” Steam Community. June 3, 2024. Accessed February 2, 2026. https://steamcommunity.com/games/499180/announcements/detail/4192367467634414898
Nintendo. “Braid, Anniversary Edition for Nintendo Switch” (store listing). Accessed February 2, 2026. https://www.nintendo.com/us/store/products/braid-anniversary-edition-switch/
Wikipedia contributors. “Braid (video game).” Wikipedia. Accessed February 2, 2026. https://en.wikipedia.org/wiki/Braid_(video_game)
Speedrun.com. “Braid” (leaderboards and game hub). Accessed February 2, 2026. https://www.speedrun.com/braid
Speedrun.com. “Braid” (guides index). Accessed February 2, 2026. https://www.speedrun.com/braid/guides
Speedrun.com. “Individual Levels Leaderboard” (Braid guide). Accessed February 2, 2026. https://www.speedrun.com/braid/guides/7efba
Speedrun.com. “LiveSplit Autosplitter” (Braid guide). Accessed February 2, 2026. https://www.speedrun.com/braid/guides/7ecvy
Speedrun.com. “Key Magic” (Braid guide). Accessed February 2, 2026. https://www.speedrun.com/braid/guides/h11sn
Speedrun.com. “Braid” (resource link). Accessed February 2, 2026. https://www.speedrun.com/braid/resources/hrviv
Speedrun.com. “Braid, Anniversary Edition” (leaderboards and game hub). Accessed February 2, 2026. https://www.speedrun.com/BraidAE
Speedrun.com. “Braid, Anniversary Edition” (resources). Accessed February 2, 2026. https://www.speedrun.com/BraidAE/resources
Speedrun.com. “Site Rules” (verification and conduct guidance). Accessed February 2, 2026. https://www.speedrun.com/support/learn/site-rules
Speedrun.com. “Any% in 22m 42s 480ms by CtrlAltDelicious” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/y2rjkjwm
Speedrun.com. “Any% in 22m 56s 570ms by GhERmann” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/zprjedvm
Speedrun.com. “Any% in 23m 30s 720ms by KanBan85” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/mk36r0lz
Speedrun.com. “Any% in 24m 14s 430ms by P-why” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/yl9rlnny
Speedrun.com. “Any% in 24m 24s 780ms by AngryMagikarp” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/kz578dgm
Speedrun.com. “Any% in 26m 08s by BakaRecca” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/kz5vlvnm
Speedrun.com. “Any% in 26m 10s 080ms by AlexGanonSm64” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/z052qr9m
Speedrun.com. “6 Stars in 25m 38s 020ms by GhERmann” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/y9n8edez
Speedrun.com. “6 Stars in 26m 16s 900ms by KanBan85” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/m3g2vpqz
Speedrun.com. “100% in 2h 12m 39s 370ms by GhERmann” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/y2vjk4jz
Speedrun.com. “100% in 2h 20m 15s by Bum” (Braid run page). Accessed February 2, 2026. https://www.speedrun.com/braid/runs/7ylj8nmp
TASVideos. “Braid” (game page). Accessed February 2, 2026. https://tasvideos.org/1452G
TASVideos. “Submission #9074: keylie & Del’s Linux Braid ‘Speedrun mode’ in 19:24.98.” Submitted May 17, 2024. Accessed February 2, 2026. https://tasvideos.org/9074S
TASVideos. “Submission #3968: keylie’s Windows Braid ‘speedrun mode’ in 22:15.28.” Accessed February 2, 2026. https://tasvideos.org/3968S
TASVideos. “Submission #6954: keylie’s Linux Braid in 19:39.58.” Accessed February 2, 2026. https://tasvideos.org/6954S
TASVideos. “Topic 22307: keylie’s Linux Braid in 19:39.58 (Boosters discussion).” Accessed February 2, 2026. https://tasvideos.org/Forum/Topics/22307
TASVideos. “Topic 14094: #3968 keylie’s Windows Braid in 22:15.28.” Accessed February 2, 2026. https://tasvideos.org/Forum/Topics/14094
StrategyWiki. “Braid/Speed Runs.” Accessed February 2, 2026. https://strategywiki.org/wiki/Braid/Speed_Runs
GameFAQs. Anathema09. “Braid – Speedrun Mode FAQ.” October 14, 2008. Accessed February 2, 2026. https://gamefaqs.gamespot.com/xbox360/943284-braid/faqs/54421
GameFAQs. “Speed Run Achievement” (message board thread). October 25, 2008. Accessed February 2, 2026. https://gamefaqs.gamespot.com/boards/943284-braid/46174872
Speed Demos Archive. “Braid” (game page). Accessed February 2, 2026. https://speeddemosarchive.com/Braid.html
Games Done Quick. “Braid by cwolf20 in 27:12 – SGDQ 2016 – Part 70.” YouTube video. Accessed February 2, 2026. https://www.youtube.com/watch?v=jeQJNyHDW2g