Speedrun Game Chronicles: Doom

From the beginning, Doom speedrunning grew out of something the original PC game already supported: recording a play session as a compact “demo” file rather than as video. Guides from the mid-1990s were already teaching players how to use the DOS executable to record and play back .lmp demos, and they noted a key reality that shaped early competition: demos were tightly tied to the exact game version and could fail to play correctly if recorded on a different build. That created an early norm where runners cared not only about the route, but also about the precise environment used to produce a run.

Because the run artifact was a demo file, the first serious record-keeping centered on sharing and collecting demos, often through simple file hosting and FTP-style distribution rather than modern streaming. The earliest long-running “public record” for classic Doom was Compet-N, a database site that began in the mid-1990s and organized tables of best-known performances from submitted demos. In that era, “leaderboards” were less about a unified web platform and more about trusted curators maintaining category tables and preserving the underlying files so that anyone could replay and verify them.

Early goals also solidified into Doom’s own category language instead of the later, more universal Any% and 100% labels. In classic Doom culture, fast-completion runs on Ultra-Violence became “UV Speed,” while full-clear performances that emphasize killing and secrets became “UV Max,” and those terms remained foundational even as the rest of speedrunning standardized different naming conventions. Just as importantly, the timing culture was built around what the engine could reproduce: Doom runs were measured through deterministic playback, where fine granularity exists at the level of “tics” (the game’s 35-per-second rhythm), even when public-facing times often appeared in more human-friendly units.

As the game’s ecosystem expanded through new releases and ports, Doom speedrunning gained a second layer of “origin story” that still matters today: compatibility. The community’s long-standing expectation is that a legitimate run should be replayable in the intended environment, which is why discussions of version behavior, port accuracy, and replay integrity became part of the scene’s foundation rather than a later technical footnote.

Classic Doom speedrunning organized itself around a shared expectation that runs should be replayable and checkable, not just watchable. The center of gravity became archives that could store demo files, preserve category tables, and keep the community’s historical record intact, with the Doom Speed Demo Archive explicitly built around hosting demos and maintaining long-term continuity. Doomworld functioned as a practical hub for discussion and submission traffic, with DSDA directing runners to post demos through Doomworld’s speed demo forum culture rather than treating the archive as a standalone upload silo.

As the scene matured, its communication channels broadened without abandoning its demo-first identity. Forum threads and long-form writeups remained important for preserving hard-won knowledge, but modern runners also leaned into real-time spaces, especially Discord, for feedback, troubleshooting, and route discussion. DSDA’s own newcomer guidance points runners toward asking questions in Discord, reflecting how “how-to” knowledge often lives in pinned messages, shared docs, and quick video examples even when the authoritative run artifact is still a demo file.

In parallel, Speedrun.com became a prominent public-facing directory for Doom boards, linking out to community infrastructure like Discord and DSDA and making it easier for newcomers to find categories, resources, and discussion threads. Submissions there follow the familiar speedrun.com rhythm of a runner posting a run and a moderator later marking it verified, while Doom-specific verification often benefits from attaching a demo because it lets moderators and other runners confirm source port behavior, stats, and input legitimacy with more confidence.

Doom’s classic category family grew out of the way the game itself reports completion and statistics on the intermission screen, and the way the community preserved runs as playable demo files. On the Doom Speed Demo Archive, the “Any%” equivalent is UV Speed, which asks you to finish a level or a full game segment as fast as possible on Ultra-Violence (skill 4), while the “100%” equivalent is UV Max, which adds the requirement of 100% secrets and all monsters dead (with a few long-standing exceptions like lost souls and endlessly spawned enemies in specific contexts).

From that core, Doom’s ruleset expands through clearly labeled challenge styles that change the run’s objective without changing the basic identity of the game. NM Speed and NM 100S move the same ideas onto Nightmare difficulty, UV Fast and UV Respawn introduce command-line behavior changes that alter monster speed or respawn behavior, and restriction categories like UV Tyson, Pacifist, and NoMo focus the skill test on limited weapons, monster interaction rules, or pure movement optimization with monsters removed.

What the community is measuring depends on the board, but Doom is unusual in that the game provides built-in timing and a deterministic replay format. Runners can rely on level time and totals shown at the end of each map, and episode-style runs have long-standing conventions for how to interpret that in-game timing across the structure of an episode. On Speedrun.com boards for Ultimate Doom, categories are typically grouped by episode bundles, and submissions often record the platform and source port used, reflecting the reality that timing, compatibility, and verification all intersect in this game.

The “big rule decisions” that define Doom’s leaderboard identity are mostly about legitimacy and compatibility rather than about banning one specific trick. Traditional competition expects a single-segment run and treats tool assistance as its own category, with explicit guidance that changing game speed, recording in segments, or using automation and informational overlays beyond what is allowed pushes a run into TAS territory. Cheats and content edits are also policed in a practical way: many cheats do not survive the demo format, and the ones that do are not accepted for traditional categories, while sprite edits are forbidden outright.

Version and platform policy follows that same principle. Source ports are allowed and often recommended, but a run is only considered legitimate if its demo can be played back in the intended environment, which is why runners commonly use ports designed around faithful compatibility and settings that mimic specific executables. That is also why Doom communities frequently treat PC demo competition as its own ecosystem, with other releases and ports living on separate boards rather than being forced into a single “one size fits all” leaderboard.

Doom speedrunning’s most enduring technical arc is the gradual conversion of raw movement into repeatable, route-level time saves. As runners pushed optimization, “going fast” stopped meaning simply running clean lines and started meaning maintaining high-speed movement through corners, doorways, and cramped geometry with disciplined inputs. The classic baseline is straferunning, including SR50-style movement that exploits the way forward motion and strafing combine to exceed normal running speed, turning even simple hallway travel into a skill test of angle control and consistency.

From there, Doom’s biggest paradigm shifts came from engine quirks that enabled sequence breaks without changing the objective of reaching exits quickly. Tricks like wallrunning and glides made it possible to bypass intended chokepoints by exploiting collision and positioning rules, either by building abnormal speed along certain wall alignments or by slipping through spaces the map was never meant to allow. Those techniques did not just shorten individual rooms. They reshaped how runners think about levels, because a route could be rebuilt around a single “make-or-break” execution that skips a key, a trigger, or an entire fight.

Strategy evolved alongside those mechanics into a more deliberate form of risk management. Doom’s enemy pressure is not simply avoided; it is steered. Runners learned to treat monster placement and projectile traffic as part of the route, using positioning to reduce blocking, invite infighting, and protect momentum. The guiding idea is consistency under stress: a fast route is only valuable if it can be reproduced across attempts, and Doom’s deep movement toolkit made it possible to trade small, safe gains for occasional high-variance “breaks” where the run lives or dies on one glide line or one wallrun setup.

Tooling then accelerated that refinement by making practice and verification easier without abandoning demo-first standards. Compatibility-focused source ports became central to modern training because they preserve replay integrity while adding quality-of-life features for recording, playback, timing diagnostics, and video dumping. DSDA maintains an explicit allowed-ports culture tied to intended environments, and DSDA-Doom in particular is designed around speedrunning workflows, including command-line options for recording demos, replaying them, and timing them in gametics. PrBoom+ and related ports also contributed long-standing practice conveniences while maintaining broad demo compatibility, which matters because the community’s historical record depends on reproducible playback rather than “trust me” video alone.

Finally, Doom’s tech culture has a research loop that keeps pushing the ceiling upward: tool-assisted experiments and deep documentation often clarify what is theoretically possible, even when a technique remains brutally difficult for real-time execution. That kind of analysis has been especially influential for understanding precision movement and wall-interaction optimization in Doom-engine play, and it feeds back into human routing as runners translate “in principle” ideas into reliable setups.

One of Doom’s defining milestones is that the scene matured with verification built in. Because runs could be recorded as demo files and replayed deterministically, early competition quickly favored results that other people could test rather than simply watch. That expectation helped create a “best times” culture long before modern streaming, with archives and tables acting as the community’s scoreboard and memory.

A second milestone was the rise of dedicated institutions that treated preservation as part of the sport. Compet-n established a model for curated tables and demo collection, and later the Doom Speed Demo Archive expanded the idea into a broader public record for classic Doom and related WAD communities. What changed here was not only visibility, but standards: submitting, categorizing, and storing the underlying demo became the norm, and the “history of the game” stayed readable because the evidence was preserved alongside the claims.

Category identity is another landmark, because Doom’s leaderboards developed a vocabulary that matched the game’s own structure. Labels like UV Speed and UV Max became more than shorthand. They turned into a shared language for what the scene values, how completeness is defined, and how difficulty settings function as part of the rules rather than a mere toggle. That common category core helped Doom remain cohesive even as new ports, new mapping ecosystems, and new platforms arrived around it.

On the technical side, Doom’s “route eras” are often marked by movements from intuition to documentation. Guides that formalized the physics and the strange edge cases of Doom’s movement created a bridge between experimentation and repeatable execution, which made high-end techniques teachable instead of purely mythical. In practice, this is where speedrunning shifted from simply knowing where to go into knowing how the engine can be pushed, through repeatable movement standards, glides, and wall interactions that reframe what is possible inside familiar maps.

Modern tooling and showcase culture then widened the tent without abandoning the demo-first roots. Source ports designed around speedrunning workflows, like dsda-doom, made practice, recording, and timing more convenient while still respecting the community’s obsession with replay integrity. At the same time, marathon showcases introduced Doom to audiences who did not grow up inside its archive culture, and those runs often served as “watershed” explanations of why Doom’s verification norms are different from many other speedgames. Doom (1993) appearing at Games Done Quick events is one example of that visibility shift, and listings like Speedrun.com helped make the modern category map easier for newcomers to find, even when the deeper record still lives in demo archives.

Robin Woehler — E1M1 — UV Max — 0:54.71 — 1995-08-08 — DooM v1.9 — an early UV Max-era demo that shows the foundational “kill, grab, exit” shape before later route compression — DSDA leaderboard entry

Yonatan Donner — E1M1 — UV Max — 0:32.49 — 1998-08-08 — DooM v1.9 — representative late-1990s refinement, when quicker max clears began to look “modern” in pacing — DSDA leaderboard entry

Thomas Pilger — E1M1 — UV Speed — 0:09.91 — 1999-02-11 — DooM v1.9 — a classic-era fast exit time that helps mark how close early runners were getting to the map’s hard movement constraints — DSDA leaderboard entry

Jonathan Rimmer — E1M1 — UV Max — 0:29.83 — 2001-04-13 — DooM v1.9 — early-2000s polish that kept pushing max times down while staying inside old-school demo norms — DSDA leaderboard entry

Grazza — E1M1 — Stroller — 0:20.74 — 2005-01-22 — DooM v1.9 — a good “culture” entry: Doom’s scene has long supported side-categories alongside the headline speed/max boards — DSDA leaderboard entry

4shockblast — E1M1 — UV Speed — 0:08.97 — 2019-02-23 — PRBoom v2.5.1.4cl3 — emblematic of the sub-9 era that established the modern standard for Hangar’s opening sprint — DSDA leaderboard entry

Meowgi — E1M1 — NM 100S — 0:17.69 — 2023-02-12 — DSDA-Doom v0.25.2cl3 — a standout “hard ruleset” performance that shows how execution and survival routing diverge from UV boards — DSDA leaderboard entry

aconfusedhuman — E1M1 — UV Speed — 0:08.74 — 2024-08-02 — DSDA-Doom v0.27.5cl3 — a modern peak example where the run is basically about perfecting tiny movement and line choices rather than discovering new structure — DSDA leaderboard entry

Zero-Master — Episode 1 — UV Max — 17:49 — 2023-12-21 — DSDA-Doom v0.27.5cl3 — a major episode-level benchmark that illustrates how “max” routing becomes a sustained resource-management puzzle across maps — DSDA history entry

MatrixCL — Episode 1 — UV Max (TAS) — 13:06 — 2025-03-25 — DSDA-Doom v0.28.0cl3 — useful as a TAS ceiling marker: it frames what’s theoretically possible when execution limits are removed — DSDA history entry

ZELLLOOO — Episode 4 — UV Speed — 4:08 — 2023-10-09 — DSDA-Doom v0.26.2cl3 — a strong “episode speed” showcase where route flow and exit discipline matter as much as raw movement — DSDA history entry

kuckkuck — D1All — UV Max — 1:16:53 — 2025-04-22 — DSDA-Doom v0.28.3cl3 — a full-game UV Max landmark that anchors the “complete Doom” story and gives the page a big, evergreen capstone entry — DSDA leaderboard entry

id Software. Doom Instruction Manual (1993). Hosted by Internet Archive. https://archive.org/details/doom-manual/page/n1/mode/1up

DoomWiki.org. “Versions of Doom and Doom II.” https://doomwiki.org/wiki/Versions_of_Doom_and_Doom_II

Doomworld. “Patches.” https://www.doomworld.com/classicdoom/info/patches.php

Doom Speed Demo Archive. “Intro, Rules, and Guidelines.” https://dsdarchive.com/intro

Doom Speed Demo Archive. “Doom / The Ultimate Doom Leaderboard.” https://dsdarchive.com/wads/doom/leaderboard

Doom Speed Demo Archive. “Doom / The Ultimate Doom History.” https://www.dsdarchive.com/wads/doom/history

Speedrun.com. “Ultimate Doom (1993).” https://www.speedrun.com/doom1

Speedrun.com. “Ultimate Doom Guides.” https://www.speedrun.com/doom1/guides

Games. “How To Make Doom .lmp Demo’s…” (Doom Help Service / DHS). https://www.gamers.org/dhs/helpdocs/demomake.html

YouTube. “The Ultimate DOOM: E1M1: Hangar – UV-Pacifist in 0:08.74.” Channel: aconfusedhuman. https://www.youtube.com/watch?v=ND5jA5xzFB0

YouTube. “22 Year Old Speedrun For The VERY FIRST Map of DOOM Beaten!” Channel: 4shockblast. https://www.youtube.com/watch?v=1UkeFwJ-yHI

YouTube. “Ultimate Doom: Episode 1 – UV-Max Speedrun in 17:49.” Channel: Zero Master. https://www.youtube.com/watch?v=ykIhK_HjVjg

Doomworld. “Welcome to the Speed Demos Forum! (Getting Started, Port Usage + Info, Co-op, etc.).” https://www.doomworld.com/forum/topic/130670-welcome-to-the-speed-demos-forum-getting-started-port-usage-info-co-op-etc/

Andrew Stine (Linguica). “The Doom Movement Bible.” Hosted by DoomWiki.org. https://doomwiki.org/wiki/The_Doom_Movement_Bible

DoomWiki.org. “Straferunning.” https://doomwiki.org/wiki/Straferunning

DoomWiki.org. “Wallrunning.” https://doomwiki.org/wiki/Wallrunning

kraflab. “dsda-doom.” GitHub repository. https://github.com/kraflab/dsda-doom

Games Done Quick. “Doom.” GDQ VODs. https://gdqvods.com/game/doom

YouTube. “Ultimate Doom by KingDime in 27:55 – AGDQ2019.” Channel: Games Done Quick. https://www.youtube.com/watch?v=5yfXcsJTJfI

Shacknews. “22-year-old Doom E1M1: Hangar speedrun record finally broken.” April 10, 2019. https://www.shacknews.com/article/111159/22-year-old-doom-e1m1-hangar-speedrun-record-finally-broken

PCGamesN. “Doom’s first mission has been beaten in eight seconds.” April 12, 2019. https://www.pcgamesn.com/doom/doom-e1m1-speedrunning-record