AI-generated
This week in speedrunning had a little bit of everything: a bounty-chasing Pokémon breakthrough, a twenty-year excavation of a first-person shooter, and enough tiny leaderboards to make the stopwatch feel like a microscope. The headline belongs to Main97, who pushed Pokémon FireRed Round 2 below a deliberately intimidating 3:25:30 barrier and, in the process, collected a $5,000 incentive that had been waiting for exactly that kind of absurd precision.
But the week was not only about the biggest number on the front page. Lufia 2 reminded us that “a second game inside the game” can still produce serious routing drama, while F.E.A.R. offered a long-form history of how a category can be steadily dismantled over two decades. Around them, runners were discovering skips in brand-new games, shaving levels down to the hundredth, and—quite literally—putting their feet into Super Mario Bros.
It was a particularly good week for the part of speedrunning that lives between spectacle and scholarship. Some runs were polished victories; others were invitations to investigate. Together, they made a persuasive case that the scene’s real engine is curiosity: someone notices a strange interaction, a route, or a version difference, and suddenly a game has somewhere new to go.
Here are the biggest stories from the week.
Main97 Breaks the FireRed Bounty Barrier
There is a special kind of speedrun milestone that sounds trivial until you understand the pressure packed into the last few seconds. Main97’s Pokémon FireRed Round 2 run is one of them: 3:25:25, a new personal world record and the first time anyone has gone under 3:25:30. The margin is only five seconds, but that is exactly why the number matters. A category can spend months treating a threshold as a wall, then suddenly the wall is five seconds behind somebody.
The run also beat Main97’s own previous world record by 1:06. That is the kind of improvement that makes a leaderboard look deceptively tidy. The result appears as one time, one line, one new record, but the route behind it is a long accumulation of decisions: where to accept risk, which fights deserve the good setup, and how to preserve a run when the game starts asking for bad luck at the worst possible moment.
And then there is the bounty. A $5,000 prize had been offered for the fastest FireRed Round 2 completion before mid-October, provided the runner could clear that exact 3:25:30 condition. Several commenters admitted they did not even know why the cutoff was so specific, which is fair until you remember that speedrunning loves a line in the sand. Once a target is announced, an arbitrary-looking time becomes a season-long antagonist.
The community response captured the category’s long memory. One commenter connected the run to watching Gunner begin the category years ago, a reminder that today’s bounty race is built on an older foundation of route development and persistence. The Twitch link may have expired, but the important part is preserved in the record: a runner took a target designed to sound just out of reach and made it the new starting point.
F.E.A.R. Loses Half Its Time, One Discovery at a Time
Some games announce their speedrunning potential immediately. Others sit quietly for years while the right people learn to look at them incorrectly. F.E.A.R. belongs to the second group, and a new deep-dive video charts how its speedrun time was cut in half across roughly twenty years, from the game’s release through August 2025.
That is a compelling premise because “half the time” is not one breakthrough. It is a history of accumulated disrespect toward the game’s intended pacing. Hallways become launch pads, combat becomes something to route around, and apparently ordinary pieces of level design turn into opportunities. A long-lived category is a record of the community learning which rules the game actually follows, as opposed to the rules its designers hoped players would follow.
The post itself asked viewers to bring a snack and a drink, which feels appropriate for a story about endurance. The most interesting speedrun documentaries are not merely highlight reels; they explain why an old time survived, who challenged it, and what changed when somebody finally found the seam. The three comments were brief, but enthusiastic, with one viewer already looking forward to another deep dive. That is how a video becomes part of a conversation rather than the end of one.
F.E.A.R. also represents an important counterpoint to the instant-record fantasy. A category can be exciting without being young, fashionable, or constantly visible. Twenty years of progress means twenty years of people returning to the same spaces with better questions. The game may be remembered for its horror, but its speedrun history is a quieter kind of haunting: every old route contains the ghost of a faster one.
Lufia 2’s Ancient Cave Is Still a Beautiful Mess
The phrase “no manipulation” does a lot of work in Lufia 2’s Ancient Cave category. The new world record, a 45:13 any-percent run, comes from a category where the runner cannot simply arrange the cave’s chaos into a preferred sequence. That makes the achievement feel less like executing a blueprint and more like negotiating with a hostile, procedurally generated roommate.
The Ancient Cave is famous for making the game feel larger than it has any right to be. As one commenter put it, they were still blown away that the developers put a whole second game into Lufia 2. That sense of surprise survives even after the route has been studied. Every shortcut has a cost, every detour may be a survival decision, and the best run is often the one that turns an ugly situation into merely a slightly less ugly one.
Another comment offered the week’s most charming piece of category wisdom: in the Ancient Cave, Copper Dragons are your friend. That sentence is a perfect example of why speedrun communities develop their own folklore. To an outsider, a dragon sounds like a problem. To a runner who understands the experience curve, drops, and route logic, it can be an essential appointment on the way to the finish.
At 45:13, the run is not just a clean result; it is evidence that the no-manipulation format still has room for judgment. Randomness does not erase skill. It changes the skill from memorizing every event to recognizing which events are worth embracing. That is a different kind of fluency, and it makes this record especially satisfying to watch.
Agent 64 Turns Tiny Levels into a Full-Speed Campaign
Agent 64: Spies Never Die had a busy little week. Runs landed for “Castle” at 1:46.853 IGT, “Mountain” at 2:51.443, and “Sanctum” at 37.551. Individually, these are compact level showcases. Together, they reveal the wonderfully granular way a new or niche game can be explored: not one grand leaderboard push, but a series of increasingly precise questions about every room.
“Sanctum” provided the week’s miniature arms race. The posted time was immediately answered in the comments with a 35.404 clip, followed by the blunt verdict that the original had been beaten. There is no ceremony in that exchange, and that is part of the charm. A runner posts a time, somebody else posts a faster one, and the level’s definition changes before the ink on the first record has dried.
The creator described “Sanctum” and another level as the two easy or fast stages, which hints at a larger route still taking shape. Early leaderboards are exciting because their future is visible. A polished mature category asks how close humans can get to the known limit; a fresh collection of levels asks what the limit even is. The difference between those questions is where a lot of the best discoveries happen.
Agent 64’s spread of times also shows why IGT categories can be so appealing. The stopwatch is part of the evidence, but the level itself remains the unit of study. A 37-second stage can contain just as much routing personality as a forty-minute game. For runners, that means a quick reset, a clear experiment, and one more chance to find the line that the game accidentally left open.
Agefield High Gets a Speedrun Before It Gets Comfortable
There is no better speedrunning origin story than “I discovered a glitch and had to see what it could do.” The runner behind a new 2:52.42 world record in Agefield High had no prior speedrunning experience, but found several glitches—including a segment skip affecting most missions—within a game that was not even a week old.
That timing matters. In an established game, a segment skip arrives in a landscape of known routes, expectations, and runners ready to test every consequence. In a game this new, the discovery can become the route. The first record is simultaneously a performance and a map: it tells everyone else where the doors are, which missions can be compressed, and what kind of nonsense the game is willing to tolerate.
The post framed the run as a debut and emphasized the joy of showcasing the route. That enthusiasm is more important than the time itself. Speedrunning can look intimidating from the outside, especially when a leaderboard is full of names with years of experience. A brand-new category lowers the drawbridge. You do not need to know the old route if there is no old route yet; you only need to be curious enough to test the glitch.
Agefield High is therefore a story about the beginning of a community, not merely the beginning of a record. The next few submissions may make 2:52.42 look ancient. That would not diminish it. First records are valuable because they give everyone else something concrete to chase—and because they preserve the moment when discovery still feels like the whole category.
Feet, Wheels, and the Physical Comedy of Speed
The week’s most gloriously literal challenge was a feet-only Super Mario Bros. run in 6:25. The runner described it as a personal challenge after dealing with tendon problems in their arms, and the result is exactly the kind of stunt that makes speedrunning feel larger than its standard categories. The game is familiar; the input method is not. Every basic movement becomes a tiny negotiation with balance, timing, and the limits of human dexterity.
The comments responded in the expected register: amazement, jokes about the feet, and a strong suggestion that nobody should investigate the smell of the game. One commenter compared it to a Guinness record that exists partly because almost nobody else would attempt it. That is not a criticism. Odd categories are often valuable precisely because they reveal a different kind of skill than the main leaderboard rewards.
Elsewhere, Mario Kart Wii’s Retro Rewind time trials offered a more conventional kind of precision, with a 1:30.105 on SNES Mario Circuit 1 and an invitation for more racers to join the online time-trial community. Put the two stories together and you get a neat snapshot of the hobby: one runner turns their body into a controller, while another builds a social space around shaving tenths from a track.
Both are speedrunning at its most approachable. You can chase a serious leaderboard, invent an eccentric ruleset, or simply ask other people to race the same course. The stopwatch does not care which doorway you use to enter. It only records what happens after you decide to try.
New Routes, Old Games, and the Joy of Breaking Things
A pair of smaller posts rounded out the week’s discovery theme. In The Elder Scrolls IV: Oblivion, a Thieves Guild% world record came in at 6:32, compressing the journey from prisoner to guild leader into a time span in which the Emperor and Armand Christophe would barely have time to process the paperwork. The comment section supplied the obvious lore joke: the runner appears to “CHIM” their way into a promotion before the game’s bureaucracy can object.
Meanwhile, a potential POSTAL 2 tactic called “mail-scamping” arrived as a suggestion rather than a finished record. That distinction is important. Speedrunning is not only the scoreboard after the discovery; it is the moment when somebody says, “This might work,” and hands the idea to the next person. A tactic can be rough, impractical, or immediately obsolete and still be useful if it changes what runners test.
Wii Sports version 1.0 offered a quieter kind of infrastructure. A player with a copy asked whether it might help speedrunners and offered to give it away or donate it. A community member replied that they had sent a direct message. Hardware versions rarely make the flashy headline, but they can redraw the boundaries of what is possible. Sometimes the next record begins with somebody preserving the right disc.
Even Racing Master, announcing a new leaderboard, belongs in this family. The post was a call for runners to submit one, two, or several runs and help turn an empty page into a competitive place. A leaderboard is not a monument that appears fully formed. It is a promise that someone else will show up, find a route, and make the first time worth beating.
A Week That Rewarded Curiosity
The connective tissue this week was not a particular console or genre. It was the willingness to look twice. Main97 looked at a bounty cutoff and treated it as a reachable finish line. F.E.A.R. runners looked at twenty years of history and found more time hiding inside it. Lufia 2 runners looked at randomness and learned which monsters were allies. Agent 64 runners looked at individual levels and found entire races contained within them.
That is why the stranger stories matter alongside the world records. Feet-only Mario, a fresh Agefield High skip, a version-one Wii Sports disc, and a possible POSTAL 2 tactic all show the same instinct from different angles: test the boundary. Sometimes the result is a record. Sometimes it is a joke, a clip, a new leaderboard, or a question that somebody else can answer better.
For a scene built around going fast, speedrunning remains remarkably patient. It remembers old games, old routes, expired streams, and the people who started categories before anyone was offering a bounty. Then, at exactly the right moment, it turns around and makes a five-second gap feel like an entire era. That tension—between deep history and sudden discovery—is what keeps the hobby moving.
Whether you are here for the immaculate world record, the cursed controller method, or the first person brave enough to submit to a brand-new leaderboard, there was plenty to chase this week. Keep your routes sharp, your dragons friendly, and your segment skips documented. See you next week.
Source: reddit.com