Edit history:
VorpalEdge: 2010-10-05 10:12:06 pm
VorpalEdge: 2010-10-05 10:03:55 pm
VorpalEdge: 2010-09-23 09:43:20 pm
VorpalEdge: 2010-09-23 09:33:34 pm
VorpalEdge: 2010-09-23 09:15:28 pm
VorpalEdge: 2010-09-23 09:06:26 pm
VorpalEdge: 2010-09-23 09:02:41 pm
mikwuyma: 2010-09-23 08:53:03 pm
mikwuyma: 2010-09-23 06:14:04 pm
mikwuyma: 2010-09-22 04:46:47 pm
welcome to the machine
I have control of the first post now! The most recent revision of the rules is below.
Quote:
At Speed Demos Archive (SDA), we publish high-quality speedruns, which are fast playthroughs of video games. As with any competitive undertaking, there has to be a standard set of rules to ensure fair play. This page establishes rules and expectations for runs submitted to SDA.
We only publish the fastest runs submitted to us. Players are expected to use every method at their disposal, including glitches, to minimize time; side issues such as entertainment are secondary. To be published, a run must be fast enough, show no evidence of cheating, and have good audio and video quality. We also highly encourage attempts to improve upon currently published speedruns. If a player submits a run faster than a published run, the faster run will replace the slower run provided it meets the above requirements for acceptance.
Not every game is fit for speedrunning. A game has to have a definite ending, so a game like <i>The Sims</i> or <i>World of Warcraft</i> doesn't qualify. Additionally, it's impossible to speed up some games. Fixed autoscrollers or rhythm games, for instance, are generally not allowed, because speedruns would not be any faster than normal playthroughs.
Submissions must have direct-feed footage of the entire run from start to finish. Some players prefer to just post a final run time, but we publish speedruns, not records. For more information about recording your run and our standards for audio and video quality, see the documents at our <a href="http://speeddemosarchive.com/kb/Main_Page">knowledge base</a>. For information about submitting a run, see our <a href="http://speeddemosarchive.com/lang/submit_en.html">submission page</a>.
When a run is submitted, it is distributed to verifiers for peer review. Verifiers check runs for things such as cheating, mistakes that cost time, missing footage, and bad audio/video quality. A more detailed explanation of verifiers’ duties and the exact criteria for judging submissions is available at the <a href="http://speeddemosarchive.com/kb/Verification_Guidelines">verification guidelines</a> page. When verifiers finish reviewing a run, their notes are posted at the <a href="http://forum.speeddemosarchive.com/board/speedrun_verification.html">verification board</a>. If a run is accepted, it is timed and eventually published on the main site.
MAIN CATEGORIES
There are two main sets of categories: the single-segment and segmented categories, and the completion percentage categories. Most runs will be either single-segment or segmented and have a completion percentage.
Speedruns that beat a game in one sitting without loading save files or quitting are single-segment runs. Loading a save file is only permitted in single-segment runs for games or categories (such as New Game +) that require doing so to progress. If the player saves and then retries multiple parts of the game individually, the run is considered segmented. Segmented runs are expected to be faster than single-segment runs in return for the ability to retry parts of the run multiple times; if a single-segment run is faster than a segmented run, the single-segment run will obsolete the segmented run (but not vice versa). For some level or track-based games, such as racers, segmented runs are replaced by an Individual Level table where the time for each track or level is displayed separately. If an individual level table does not exist for a game, the first submission of individual level runs must complete every race or level in that table's category.
There are also three completion percentage categories. The first is the any% category, in which players do whatever it takes to get to the end of the game as fast as possible. This can be considered the default category.
Next is the low% category, wherein players collect the minimum number of items and upgrades necessary to complete the game. What counts or doesn't count can be subjective, so check the [link]low%/100% definition thread[/link] for more information. Note that the low% category does not restrict usage of items players are forced to obtain. For instance, Mega Man always collects a weapon after defeating each robot master or maverick in a Mega Man game. Since players are forced to collect these weapons, runs that do not use them when they would save time will be rejected. Lastly, since low% runs are defined by the number of items and upgrades they pick up, a run that picks up fewer collectibles will obsolete a run that collects more, even if the new run is slower as a result.
Finally, there is the 100% category, where players collect "everything" in the game. As with low%, how to define 100% for any given game is subjective unless the game tracks percentage, and many games do not have a 100% category. Therefore, it is recommended to check the [link]low%/100% definition thread[/url] and ask questions if necessary. We generally prefer concise, reasonable definitions (instead of "comprehensive" laundry lists) that allow for enough differentiation from the any% category. A good example is "all stages". A bad example is "all items" in an RPG. Do you need max of each? Mutually exclusive items? If we're getting all items, why not all skills? Max levels? And so on.
Some games are designed in such a way that multiple completion percentage categories may apply to the same run. For instance, skipping all optional items may be the fastest way through some games, while other games might require you to fulfill 100% conditions to beat them. In these cases, the overlapping categories are identical and such runs are treated as any% runs for simplicity.
OTHER CATEGORIES
Besides the basic categories, there are several other categories runs could fit into, depending on how they were performed. These include:
Major skips: Some games have particularly egregious glitches or skips that deserve a category of their own. Examples of these are select warping in <i>The Legend of Zelda: Link's Awakening</i>, or summon warping in <i>Castlevania: Circle of the Moon</i>. Unfortunately, it's impossible to create a solid definition of major skips that works for every game, so major skips categories are tailored to each game that warrants it.
Game-dependent: These categories include things like different difficulty settings or game modes, or even multi-player runs. As a side note, multi-player runs are not distinguished by the number of players; only single-player and multi-player are categories, and single-player runs can obsolete multi-player runs.
Version differences: Many games are available on multiple systems. Players may produce runs on any of these systems, but many viewers would rather watch runs on their original systems. If the differences between game versions are severe, such as the differences between the NTSC and PAL versions of many older games, they will be classified as separate categories. If a game has been patched and there are multiple versions, we recommend using whichever version is fastest for running.
Deaths: For games in which dying significantly lowers the final time and changes the strategy throughout the run, runs with deaths can be categorized separately.
TIMING
We prefer to use consistent in-game timers wherever possible. If an in-game timer isn't available, a simple real-time measure is used. Timing begins when the player first gains control of the game's character, and stops when control is lost after the final event in a game. For segmented runs, timing for a segment stops at the first system-dependent activity, usually the actual saving. When loading, timing resumes at the point when the game was saving or displaying the password. For runs over three hours, seconds are rounded down to the nearest minute in the final time because seconds are insignificant.
For games that let players save and reload anywhere, a half-second penalty is added for each manual save. This penalty counters a common problem where games drop frames immediately after a load and discourages players from potentially using thousands of segments in a run. The half second penalty applies to both quick-saves and saves made through a menu, but menu time is not counted when saving in a menu. We do not apply a penalty for segmentation in games where saves can only be performed at save points or in games where players warp to a predefined location upon reloading, since only being able to save at save points or reload at warp points is itself a restriction on segmentation.
Some games have a timer that can't be seen at the end of the game. In most of these cases the timer will be ignored. For some games, however, such timers are displayed on a menu screen that would be visible before the final event in the game; if this is the case, that time will be noted and real time will be appended from that point.
For manually timed games with an existing run, improvements must save at least one additional second. If a run is over three hours long, the minimum improvement allowed is one minute, because seconds are less significant past the three-hour mark. Improvements smaller than a second are only acceptable if a game has a visible timer that measures fractions of a second.
BANNED CHEATS/HACKS
First, and most obviously, NO CHEATING. Only fairly produced runs are accepted. Ask on the forum if there are any questions about what we permit. If there is no answer on the forum, then try asking <a href="http://forum.speeddemosarchive.com/profile/mikwuyma.html">Mike</a>. We also reserve the right to reject any run for any reason (including, but not limited to, suspected cheating or objectionable content, such as pornography).
Things that are banned include:
-Cheat devices: Game Genies, Action Replays, and similar devices are banned.
-Hardware modification: Players may not modify their systems in any way (e.g. overclocking it to reduce lag). The only extra hardware allowed are mod chips and boot disks used for playing imports, video mods to allow old systems to output higher quality audio/video, and official add-ons, such as the Nintendo 64 expansion pack. For example, the PlayStation 2 HDD is allowed, whereas the HD Loader is not. Additionally, to answer a commonly-asked question: since glitches that involve pressing left and right or up and down at the same time require pressing buttons hard enough to break or circumvent the crossbar underneath d-pads, glitches that involve pressing these buttons at the same time are banned. This section only applies to consoles; PC players can feel free to bump their graphics settings up as much as they can.
-Game modification: Removing or altering a game's internal files or hardware is forbidden. This rule includes things like mods or using the crooked cartridge trick in <i>The Legend of Zelda: Ocarina of Time</i>. Mods may be accepted as runs of a separate game at staff discretion.
-Third party controllers: Third-party controllers are allowed, but players must only use features present on controllers that were officially bundled with the system. Thus, turbo-fire is not allowed, except on systems such as the TurboGrafx-16 that come with official turbo-fire controllers. If a game provides a turbo-fire option, it may be used. Players that display very rapid firing in runs may be expected to show video camera proof of their button mashing ability.
-Virtualization: We will not accept speedruns recorded on emulators (ZSNES, VBA, DOSBox, etc.) or other virtualization software (Wine, VMware, etc.). Emulators commonly allow for recording games frame-by-frame and then playing back the input at normal speed. Also, most emulators and virtualization programs have minor inaccuracies in timing and slowdown that prevent accurate comparisons between runs. Finally, it is illegal in most countries to obtain ROMs. Note that an exception is made for officially sanctioned emulators such as the Game Boy Player for GameCube, Virtual Console, or GameTap. Due to problems with some official emulators we prefer that players use original hardware whenever possible.
-Codes: Using a beneficial cheat code is not allowed. This includes codes that give more lives, reduce damage, and so on. Codes that are only cosmetic, like suitless Samus in <i>Metroid</i>, are allowed. A code that increases the game's difficulty, such as <i>Donkey Kong Country 3</i> 105%, may qualify as a separate category.
-Scripts: Many PC games allow players to use scripts or macros to automate certain actions. These are banned. In the past, scripts were not explicitly banned from SDA, so there are currently a few runs that use scripts. However, no further submissions using scripts will be accepted.
FREQUENTLY ASKED QUESTIONS
Q: I want to do a single-segment run of a game, but I’m forced to quit at a certain part. Does that make my run segmented?
A: Exceptions to the no-quitting rule for single segment runs are made for games where quitting is required, provided there is video camera proof to show that the run isn’t segmented.
Q: I'm trying to record a single-segment run, but the run is too long for one DVD. Is it okay if I switch DVDs?
A: Yes, but be sure you switch DVDs while the game idles at a pause screen or another screen without much action.
Q: I’m trying to manually time a run for comparison purposes, but SDA timing ends earlier than I expected. Where did the discrepancy come from?
A: Timing ends when a player loses control after the final event in a game. This means that time isn’t counted if a player regains control later on (perhaps for an interactive ending or credits sequence).
We only publish the fastest runs submitted to us. Players are expected to use every method at their disposal, including glitches, to minimize time; side issues such as entertainment are secondary. To be published, a run must be fast enough, show no evidence of cheating, and have good audio and video quality. We also highly encourage attempts to improve upon currently published speedruns. If a player submits a run faster than a published run, the faster run will replace the slower run provided it meets the above requirements for acceptance.
Not every game is fit for speedrunning. A game has to have a definite ending, so a game like <i>The Sims</i> or <i>World of Warcraft</i> doesn't qualify. Additionally, it's impossible to speed up some games. Fixed autoscrollers or rhythm games, for instance, are generally not allowed, because speedruns would not be any faster than normal playthroughs.
Submissions must have direct-feed footage of the entire run from start to finish. Some players prefer to just post a final run time, but we publish speedruns, not records. For more information about recording your run and our standards for audio and video quality, see the documents at our <a href="http://speeddemosarchive.com/kb/Main_Page">knowledge base</a>. For information about submitting a run, see our <a href="http://speeddemosarchive.com/lang/submit_en.html">submission page</a>.
When a run is submitted, it is distributed to verifiers for peer review. Verifiers check runs for things such as cheating, mistakes that cost time, missing footage, and bad audio/video quality. A more detailed explanation of verifiers’ duties and the exact criteria for judging submissions is available at the <a href="http://speeddemosarchive.com/kb/Verification_Guidelines">verification guidelines</a> page. When verifiers finish reviewing a run, their notes are posted at the <a href="http://forum.speeddemosarchive.com/board/speedrun_verification.html">verification board</a>. If a run is accepted, it is timed and eventually published on the main site.
MAIN CATEGORIES
There are two main sets of categories: the single-segment and segmented categories, and the completion percentage categories. Most runs will be either single-segment or segmented and have a completion percentage.
Speedruns that beat a game in one sitting without loading save files or quitting are single-segment runs. Loading a save file is only permitted in single-segment runs for games or categories (such as New Game +) that require doing so to progress. If the player saves and then retries multiple parts of the game individually, the run is considered segmented. Segmented runs are expected to be faster than single-segment runs in return for the ability to retry parts of the run multiple times; if a single-segment run is faster than a segmented run, the single-segment run will obsolete the segmented run (but not vice versa). For some level or track-based games, such as racers, segmented runs are replaced by an Individual Level table where the time for each track or level is displayed separately. If an individual level table does not exist for a game, the first submission of individual level runs must complete every race or level in that table's category.
There are also three completion percentage categories. The first is the any% category, in which players do whatever it takes to get to the end of the game as fast as possible. This can be considered the default category.
Next is the low% category, wherein players collect the minimum number of items and upgrades necessary to complete the game. What counts or doesn't count can be subjective, so check the [link]low%/100% definition thread[/link] for more information. Note that the low% category does not restrict usage of items players are forced to obtain. For instance, Mega Man always collects a weapon after defeating each robot master or maverick in a Mega Man game. Since players are forced to collect these weapons, runs that do not use them when they would save time will be rejected. Lastly, since low% runs are defined by the number of items and upgrades they pick up, a run that picks up fewer collectibles will obsolete a run that collects more, even if the new run is slower as a result.
Finally, there is the 100% category, where players collect "everything" in the game. As with low%, how to define 100% for any given game is subjective unless the game tracks percentage, and many games do not have a 100% category. Therefore, it is recommended to check the [link]low%/100% definition thread[/url] and ask questions if necessary. We generally prefer concise, reasonable definitions (instead of "comprehensive" laundry lists) that allow for enough differentiation from the any% category. A good example is "all stages". A bad example is "all items" in an RPG. Do you need max of each? Mutually exclusive items? If we're getting all items, why not all skills? Max levels? And so on.
Some games are designed in such a way that multiple completion percentage categories may apply to the same run. For instance, skipping all optional items may be the fastest way through some games, while other games might require you to fulfill 100% conditions to beat them. In these cases, the overlapping categories are identical and such runs are treated as any% runs for simplicity.
OTHER CATEGORIES
Besides the basic categories, there are several other categories runs could fit into, depending on how they were performed. These include:
Major skips: Some games have particularly egregious glitches or skips that deserve a category of their own. Examples of these are select warping in <i>The Legend of Zelda: Link's Awakening</i>, or summon warping in <i>Castlevania: Circle of the Moon</i>. Unfortunately, it's impossible to create a solid definition of major skips that works for every game, so major skips categories are tailored to each game that warrants it.
Game-dependent: These categories include things like different difficulty settings or game modes, or even multi-player runs. As a side note, multi-player runs are not distinguished by the number of players; only single-player and multi-player are categories, and single-player runs can obsolete multi-player runs.
Version differences: Many games are available on multiple systems. Players may produce runs on any of these systems, but many viewers would rather watch runs on their original systems. If the differences between game versions are severe, such as the differences between the NTSC and PAL versions of many older games, they will be classified as separate categories. If a game has been patched and there are multiple versions, we recommend using whichever version is fastest for running.
Deaths: For games in which dying significantly lowers the final time and changes the strategy throughout the run, runs with deaths can be categorized separately.
TIMING
We prefer to use consistent in-game timers wherever possible. If an in-game timer isn't available, a simple real-time measure is used. Timing begins when the player first gains control of the game's character, and stops when control is lost after the final event in a game. For segmented runs, timing for a segment stops at the first system-dependent activity, usually the actual saving. When loading, timing resumes at the point when the game was saving or displaying the password. For runs over three hours, seconds are rounded down to the nearest minute in the final time because seconds are insignificant.
For games that let players save and reload anywhere, a half-second penalty is added for each manual save. This penalty counters a common problem where games drop frames immediately after a load and discourages players from potentially using thousands of segments in a run. The half second penalty applies to both quick-saves and saves made through a menu, but menu time is not counted when saving in a menu. We do not apply a penalty for segmentation in games where saves can only be performed at save points or in games where players warp to a predefined location upon reloading, since only being able to save at save points or reload at warp points is itself a restriction on segmentation.
Some games have a timer that can't be seen at the end of the game. In most of these cases the timer will be ignored. For some games, however, such timers are displayed on a menu screen that would be visible before the final event in the game; if this is the case, that time will be noted and real time will be appended from that point.
For manually timed games with an existing run, improvements must save at least one additional second. If a run is over three hours long, the minimum improvement allowed is one minute, because seconds are less significant past the three-hour mark. Improvements smaller than a second are only acceptable if a game has a visible timer that measures fractions of a second.
BANNED CHEATS/HACKS
First, and most obviously, NO CHEATING. Only fairly produced runs are accepted. Ask on the forum if there are any questions about what we permit. If there is no answer on the forum, then try asking <a href="http://forum.speeddemosarchive.com/profile/mikwuyma.html">Mike</a>. We also reserve the right to reject any run for any reason (including, but not limited to, suspected cheating or objectionable content, such as pornography).
Things that are banned include:
-Cheat devices: Game Genies, Action Replays, and similar devices are banned.
-Hardware modification: Players may not modify their systems in any way (e.g. overclocking it to reduce lag). The only extra hardware allowed are mod chips and boot disks used for playing imports, video mods to allow old systems to output higher quality audio/video, and official add-ons, such as the Nintendo 64 expansion pack. For example, the PlayStation 2 HDD is allowed, whereas the HD Loader is not. Additionally, to answer a commonly-asked question: since glitches that involve pressing left and right or up and down at the same time require pressing buttons hard enough to break or circumvent the crossbar underneath d-pads, glitches that involve pressing these buttons at the same time are banned. This section only applies to consoles; PC players can feel free to bump their graphics settings up as much as they can.
-Game modification: Removing or altering a game's internal files or hardware is forbidden. This rule includes things like mods or using the crooked cartridge trick in <i>The Legend of Zelda: Ocarina of Time</i>. Mods may be accepted as runs of a separate game at staff discretion.
-Third party controllers: Third-party controllers are allowed, but players must only use features present on controllers that were officially bundled with the system. Thus, turbo-fire is not allowed, except on systems such as the TurboGrafx-16 that come with official turbo-fire controllers. If a game provides a turbo-fire option, it may be used. Players that display very rapid firing in runs may be expected to show video camera proof of their button mashing ability.
-Virtualization: We will not accept speedruns recorded on emulators (ZSNES, VBA, DOSBox, etc.) or other virtualization software (Wine, VMware, etc.). Emulators commonly allow for recording games frame-by-frame and then playing back the input at normal speed. Also, most emulators and virtualization programs have minor inaccuracies in timing and slowdown that prevent accurate comparisons between runs. Finally, it is illegal in most countries to obtain ROMs. Note that an exception is made for officially sanctioned emulators such as the Game Boy Player for GameCube, Virtual Console, or GameTap. Due to problems with some official emulators we prefer that players use original hardware whenever possible.
-Codes: Using a beneficial cheat code is not allowed. This includes codes that give more lives, reduce damage, and so on. Codes that are only cosmetic, like suitless Samus in <i>Metroid</i>, are allowed. A code that increases the game's difficulty, such as <i>Donkey Kong Country 3</i> 105%, may qualify as a separate category.
-Scripts: Many PC games allow players to use scripts or macros to automate certain actions. These are banned. In the past, scripts were not explicitly banned from SDA, so there are currently a few runs that use scripts. However, no further submissions using scripts will be accepted.
FREQUENTLY ASKED QUESTIONS
Q: I want to do a single-segment run of a game, but I’m forced to quit at a certain part. Does that make my run segmented?
A: Exceptions to the no-quitting rule for single segment runs are made for games where quitting is required, provided there is video camera proof to show that the run isn’t segmented.
Q: I'm trying to record a single-segment run, but the run is too long for one DVD. Is it okay if I switch DVDs?
A: Yes, but be sure you switch DVDs while the game idles at a pause screen or another screen without much action.
Q: I’m trying to manually time a run for comparison purposes, but SDA timing ends earlier than I expected. Where did the discrepancy come from?
A: Timing ends when a player loses control after the final event in a game. This means that time isn’t counted if a player regains control later on (perhaps for an interactive ending or credits sequence).
Thread title: