SI-generated
Speedrunning had one of those weeks where the past refused to stay finished. A nearly three-decade-old game crossed a barrier its community had treated as mythical, a classic Zelda category found a route-changing discovery, and the people maintaining the scene’s infrastructure had to explain some very consequential fine print. Elsewhere, a fangame marathon reached a level of endurance that makes the word “100%” feel like a dare.
That combination is the real charm of the hobby: one minute the conversation is about the perfect randomizer item, and the next it is about ownership, licensing, or a technique that makes a familiar level look newly invented. This week’s biggest stories all had the same underlying pulse. The games may be old, but the questions around them—what is possible, who maintains the record, and how a community preserves its knowledge—keep moving.
There was also plenty of the usual speedrun joy in the margins: a co-op Teenage Mutant Ninja Turtles record, a new personal best in I Wanna Be The Boshy, a pair of tiny farming-game races, and a RollerCoaster Tycoon TAS that completes an entire campaign before most players have decided what to build. But the week belonged to breakthroughs, boundaries, and the people willing to keep pushing at both.
Here are the biggest stories from the week.
Majora’s Mask Finds a New Clock to Break
“We Solved Song of Time Skip in Majora’s Mask” is the kind of headline that instantly makes a speedrunner sit up. The Song of Time is not just a memorable tune in Majora’s Mask; it is the game’s central reset button, the mechanism that turns a three-day apocalypse into a repeatable laboratory. Finding a new way to manipulate or bypass that system is therefore less like shaving a corner and more like discovering a new door in the room where the whole route is planned.
The post itself kept the details tantalizingly concise, promising that the next few weeks and months would be interesting. That restraint is appropriate. Discoveries of this size need testing before they become route gospel: runners have to establish the setup, determine its consistency, work out whether it is useful in the main category, and then figure out what the rest of the game does when the old assumptions are removed. A video can prove that something works; a leaderboard-level strategy has to prove that it works when the clock is already on fire.
The reaction from outside the Zelda specialist bubble was especially telling. One commenter admitted not following Zelda runs closely but still called the find “sick,” while another joked that Twilight Princess deserves a discovery of its own. That is the contagious part of major tech: it makes specialists reach for their notebooks, but it also reminds everyone else why speedrunning is more than a collection of old records. A solved game can still surprise you.
Expect this one to be a process rather than a single headline moment. The first breakthrough is the spark; the route changes, practice sessions, failed setups, and eventual optimized runs are the fire. Majora’s Mask may have just received another reason to be watched very closely.
Half-Life Finally Dives Under 24 Minutes
KaNangaTV’s 23:58 Half-Life run is a record with a satisfying shape: a famous game, a round-number barrier, and a community that had every reason to believe the route was already mature. Half-Life launched in 1998, which means the run arrived more than a quarter-century after the game first taught players to fear headcrabs. Yet the first sub-24-minute completion still had to be earned the hard way.
The discussion immediately focused on the possibility of a new Blast Pit strategy, a sign that this was not merely a flawless execution of a settled script. One commenter pointed out how wild it is that major skips are still appearing 28 years after release, even after landmark discoveries such as the underwater coolant-pipe clip in Lambda Core. That feeling—surely the game must be understood by now, followed by the game proving otherwise—is one of the most durable engines in classic-game speedrunning.
Sub-24 is also a useful reminder that “old” and “solved” are not synonyms. A route can be studied for years while still hiding an interaction that nobody had the right reason to test. New hardware, new tools, better video analysis, and simply a fresh runner’s willingness to question inherited wisdom can all reopen a category. Half-Life’s age gives the accomplishment its flavor, but the real story is the refusal to treat history as a finished route.
There is a particular elegance to beating this barrier by two seconds. It sounds tiny until you remember that those two seconds are the final visible edge of thousands of invisible decisions: movement lines, enemy behavior, menu timing, damage boosts, and the nerve to keep the run alive when a reset would be easier. 23:58 is a number, but it is also a very clear message that Half-Life still has room to move.
The Speedrun Terms of Service Enter the Boss Fight
Not every important speedrunning story comes with a timer on screen. Speedrun.com published a clarification of its new terms of service, and the community response was immediate, skeptical, and occasionally very funny. The top comment’s “pinky promise” joke about not trying to claw users’ IP back again captured the mood: people were not looking for a soothing summary so much as clarity about what the platform actually claims and what it intends to do with runner data.
The most serious discussion centered on licensing and AI training. The clarification explained the difficulty of finding a Creative Commons license that permits commercial use while also excluding AI training, noting that the previously used BY-NC license prevented commercial use altogether. That is a genuine legal and technical tension, but explaining the tension is not the same as resolving the trust problem. Communities want their leaderboards to be useful, their data to be portable, and their contributions not to become an unexamined asset extracted by someone else.
Several commenters zeroed in on the language itself, saying it read like a summary generated without enough human review. That criticism matters because terms are infrastructure. A runner should not need to reverse-engineer a policy document to understand whether a submitted video, run history, or username is being licensed in ways they never intended. The less precise the explanation, the more every ordinary user is pushed into the role of a lawyer.
Speedrunning has always been unusually dependent on shared platforms: leaderboards, verification queues, guides, splits, videos, and APIs make a scattered hobby feel like a connected sport. This week’s argument is therefore bigger than one website’s wording. It is about who gets to steward the record and what obligations come with being the place where thousands of communities store their history. The stopwatch may be the star, but the database is the stage.
Zelda Changes Addresses, and Everyone Checks the Moving Boxes
The Zelda speedrunning community also spent the week dealing with a different kind of infrastructure shock: the zeldaspeedruns.com domain was sold and is no longer associated with the community. The recently redesigned home is now zsr.gg, but the obvious question is what happens to everything attached to the old address—especially guides, links, and years of accumulated trust.
The top discussion was practical rather than sentimental. Can the old guides be ported? Will ownership rights make that complicated? Is the old site coming down, or will the buyer leave it online and effectively compete with the community’s new home? These are not abstract concerns for a runner trying to learn a route. A guide that loads today may redirect tomorrow, and a search result can remain confidently wrong long after a community has moved on.
That is why domain names carry so much weight in speedrunning. They are not merely signs above a storefront. They are bookmarks in practice schedules, references in videos, destinations in Discord messages, and shorthand for “this is where the vetted knowledge lives.” When the address changes, a community has to preserve not just files but context: which guide is current, which categories are active, and which old explanations should no longer be trusted.
The move to zsr.gg is an opportunity as well as an annoyance. A fresh site can make navigation cleaner, consolidate scattered resources, and give newer runners a less intimidating way into Zelda’s enormous catalog. But the transition only works if the community treats migration as part of the run: route every link, verify every resource, and leave no useful knowledge trapped behind an expired address.
The Randomizer Gives You a Shield and Calls It a Victory
Among the week’s heavier stories, one image-based post offered the perfect speedrunner joke: “Is it the best item for a speedrunner?” The accompanying discussion imagined a randomizer handing over a shield but no bombs, which is an excellent summary of the randomizer experience. The item is good. The item is absolutely not the item you need. The seed has a sense of humor.
The debate quickly expanded beyond the immediate punchline. Is the best randomly obtained item really the best item overall, or should the crown go to a scripted story item that breaks a game wide open? That distinction gets at a core truth of speedrunning: value is contextual. An object can be useless in a normal playthrough, essential in a particular route, or completely transformational when a glitch turns its intended purpose inside out.
Randomizers make that logic visible because they strip away the comfort of a known sequence. Runners are forced to evaluate movement, risk, and routing in real time, often while pretending that the seed has not just personally insulted them. The best item is not necessarily the flashiest one; it is the one that converts uncertainty into options. Sometimes that is a shield. Sometimes it is a bomb. Sometimes it is the thing that lets you skip needing either.
It is a small story compared with a world record, but it represents the week’s mood beautifully. Speedrunning is part laboratory, part library, and part comedy sketch written by hostile item tables. Even when the post is just a joke about loot, the replies reveal runners thinking in categories, routes, dependencies, and breakpoints. The stopwatch is never far away.
I Wanna Kill the Kamilia 3 Turns “100%” Into an Expedition
PlutoTheThing’s 6:09:32 I Wanna Kill the Kamilia 3 100% run is the week’s great endurance achievement. The time is already imposing, but the community context makes it land harder: even completing the vanilla game is described as a feat achieved by only a few thousand people worldwide over more than a decade. This is not a casual detour through a difficult fangame. It is a sustained campaign against precision platforming, repetition, and the tiny errors that become very expensive several hours in.
The post’s comments supplied the history lesson. One viewer remembered watching Stonk run the game ten years ago, while another noted that Pluto has taken it to a level many once considered impossible. That is how fangame communities build their own mythology: old videos become landmarks, former walls become ordinary practice, and a new runner’s time changes the imagined ceiling for everyone who comes next.
There is even a wonderfully specific wrinkle in the definition of 100%. Pluto completes all but three achievements, prompting the natural question of what those achievements are and why they are not included in the category. That is not pedantry; it is the kind of detail that makes speedrun categories meaningful. “Everything” has to be defined, especially in a game where optional challenges can be more punishing than the nominal ending.
Six hours is long enough for the run to become a relationship. Early confidence, mid-run fatigue, a late-game rescue, and the peculiar calm that arrives when the finish is finally close: endurance categories create drama differently from a two-minute sprint. Pluto’s result is a record, but it is also a map of what persistence looks like when the game keeps asking the same impossible question and the runner keeps answering.
The Tiny Records Are Still Doing Big Work
The lower-scoring posts this week offered a useful reminder not to confuse Reddit heat with speedrunning importance. A new personal best in I Wanna Be The Boshy came after a year-long gap, with the runner describing the grind as a brutal subjective bottleneck and acknowledging that the current world record is still more than two minutes ahead. That is the emotional truth of PBs: a run can be nowhere near the top of a leaderboard and still represent a wall finally giving way.
There was also a pair of Unharvest runs pushing around the nine-minute mark, a new Knights of the Old Republic unrestricted run, a new Atlantis: The Lost Empire PS1 technique promising more than a minute of savings, and a two-player Teenage Mutant Ninja Turtles: Shredder’s Revenge arcade record. They did not dominate the comment counts, but they show the scene doing what it always does: expanding sideways into every game where someone can find a timer and a reason to try.
The Atlantis post in particular has the classic speedrun-news shape: “new zombie tech,” a named time save, and exact timestamps for anyone who wants to see the trick in action. The explanation can wait; the important part is that a route everyone thought they knew has acquired a new option. In a mature community, even a niche category can be one discovery away from a different game.
And then there is the social side: a planned four-runner TimeSplitters 2 race, where the event itself is the story. Not every week needs a record. Sometimes the best contribution is giving runners a date, a bracket, or a reason to put the splits on screen together. Competition is content, but it is also community maintenance.
A Theme for the Week: Nothing Is as Finished as It Looks
Put the week’s stories together and a clear theme emerges. A 1998 shooter breaks a barrier. A Zelda category discovers a route-changing trick. Majora’s Mask finds another way to argue with time. A website changes domains, while a platform explains terms that may shape how community data is used. Even the randomizer joke turns into a discussion of what an item means when the route changes around it.
That is the strange comfort of speedrunning. The hobby is relentlessly competitive, but its best stories are rarely just “someone was faster.” They are stories about interpretation: a runner sees a level differently, a community defines a category more carefully, a researcher tests an interaction everyone else stopped noticing, or a group decides that its knowledge deserves a safer home. The record is the visible tip. Underneath it is a lot of patient collective work.
This week also showed why speedrunning needs both spectacle and stewardship. A sub-24 Half-Life run brings people in; clear rules and durable guides help them stay. A six-hour fangame completion inspires awe; a thoughtful category definition tells the next runner what mountain they are actually climbing. The scene grows when those pieces reinforce one another instead of competing for attention.
So keep an eye on the next few months. Majora’s Mask may be entering a new research cycle, Half-Life’s barrier may already be under threat, and the discussions around platform policy and community ownership are not going away just because the wording has been clarified. The games are old. The work is not. See you next week.
Source: reddit.com