Do we need more than one run per game/category?

2018-03-20 09:12:52 am

To be clear, this isn't related to SDA in any way.

The question is... what is the purpose of having an evolution in routes and run times? I was watching some of SummoningSalt's world record progression series (very well-done btw) and here's my main thoughts:

1) How many such record runs there are is a function of how much effort the each runners/trickers etc. happened to put into it, albeit long-standing (milestone) records were more likely the result of more effort.
2) Multiple WRs are not functionally different from the same person's successive PBs, except if there's an implication of route changes each time instead of grinding the same route.
3) A speedrun is a demonstration of more fundamental (and interesting) rules and mechanisms that arise from the game's code within the specific paradigm of completing a playthrough as fast as possible. It's not the primary item of interest. This can easily be seen when you think about the way people commentate their runs: "Here you're supposed to... but you can also just...". That is, it is some non-obvious time-saving alternative in such and such a gameplay situation that is being brought up and demonstrated. It's not "here I'm expressing my creative liberty by holding down a button that makes the character spin around for no particular reason". Thus an individual speedrun's worth is measured in its efficiency in channeling the game's inherent "speedrun-geist", and so an obsoleted run has to be demoted.

That having been said, old runs may still have some intrinsic value as route/technique demonstrations. They just typically aren't as efficient in that function as a dedicated database with more focused demonstration videos would be.

So what does that leave us with? I guess some people are interested in the competitive aspect of it, in which case I suppose WRs by different runners ARE different from PBs by the same one. I don't think most casual viewers, the middle of the bell curve, care much who the runner is though. There are other points to make:

- The amount of effort required to produce a speedrun (SS, segmented, ILs or TAS) can grow to a very substantial amount. This adds a side effect similar to why some Guinness record for something like "Longest distance someone's drawn a cart for" might pique anyone's interest: nobody cares about the activity itself, just how difficult or ridiculous it sounds, because it rises above our everyday aspirations and what we could picture ourselves doing. This has a particular name I've forgotten. Thus a speedrun will share some of that with any arbitrary record. While it's something to recognize – for sure – it's just not what we're fundamentally interested in in speedruns most of the time.
- You could argue that venerating record holders and their runs is a necessary driving force. While it's obvious it's not necessary, and serves in fact as a distraction to those with "pure speedrunning" (for its own sake) in mind, it may at least serve as a vent that allows the record holder, no doubt coming off a long soul-eating grind, to cool down for the arbitrary duration of their reign.

There's probably more to say but I'll leave that for the rest of you...

Thread title:

To be clear, this isn't related to SDA in any way.

The question is... what is the purpose of having an evolution in routes and run times? I was watching some of SummoningSalt's [url=https://www.youtube.com/watch?v=qFCF4I09Avc]world record progression[/url] series (very well-done btw) and here's my main thoughts:

1) How many such record runs there are is a function of how much effort the each runners/trickers etc. happened to put into it, albeit long-standing (milestone) records were more likely the result of more effort.
2) Multiple WRs are not functionally different from the same person's successive PBs, except if there's an implication of route changes each time instead of grinding the same route.
3) A speedrun is a demonstration of more fundamental (and interesting) rules and mechanisms that arise from the game's code within the specific paradigm of completing a playthrough as fast as possible. It's not the primary item of interest. This can easily be seen when you think about the way people commentate their runs: "Here you're supposed to... but you can also just...". That is, it is some non-obvious time-saving alternative in such and such a gameplay situation that is being brought up and demonstrated. It's not "here I'm expressing my creative liberty by holding down a button that makes the character spin around for no particular reason". Thus an individual speedrun's worth is measured in its efficiency in channeling the game's inherent "speedrun-geist", and so an obsoleted run has to be demoted.

That having been said, old runs may still have some intrinsic value as route/technique demonstrations. They just typically aren't as efficient in that function as a dedicated database with more focused demonstration videos would be.

So what does that leave us with? I guess some people are interested in the competitive aspect of it, in which case I suppose WRs by different runners ARE different from PBs by the same one. I don't think most casual viewers, the middle of the bell curve, care much who the runner is though. There are other points to make:

- The amount of effort required to produce a speedrun (SS, segmented, ILs or TAS) can grow to a very substantial amount. This adds a side effect similar to why some Guinness record for something like "Longest distance someone's drawn a cart for" might pique anyone's interest: nobody cares about the activity itself, just how difficult or ridiculous it sounds, because it rises above our everyday aspirations and what we could picture ourselves doing. This has a particular name I've forgotten. Thus a speedrun will share some of that with any arbitrary record. While it's something to recognize – for sure – it's just not what we're fundamentally interested in in speedruns most of the time.
- You could argue that venerating record holders and their runs is a necessary driving force. While it's obvious it's not [i]necessary[/i], and serves in fact as a distraction to those with "pure speedrunning" (for its own sake) in mind, it may at least serve as a vent that allows the record holder, no doubt coming off a long soul-eating grind, to cool down for the arbitrary duration of their reign.

There's probably more to say but I'll leave that for the rest of you...

reality

2018-03-23 03:28:39 pm

Nostalgia Kills

I'm really torn in two directions by this topic...

I hate to give a response that implies I think that games I don't run should have multiple categories and the noes that I do run shouldn't but I can't really work my thoughts out right now.

Pro - multiple categories

The mindset of doing something like a 70 star Super Mario 64 run, or Banjo Tooie with or without Cutscene Warp 14 VS 70 is completely different. Both are of course no longer near their old status of being the "any% category" but the combination of having multiple runners who've practicted the Old Any% route, and the greater volume of mechanics in the classic route, means that there is some reason to preserve it.

As a viewer I really enjoy having multiple categories, especially for games that can evolve into credit-warps (Skyward Sword / Ocarina of Time) or for (mostly PC) games that transform from having out of bounds in one or two places, to runs (2-3 years later) that find ways to play almost the entire game out of bounds (portal/Halo 1). I really enjoy being able to watch something that I can partially connect to the way I played the game casually.

Generally for games like this (or human total control in Super Mario World).. I think the full brokenness of the current WR route is interesting to the "casual viewer" once in something like a marathon to show off the extent that the game can be pushed etc... But people hoping to watch a runner's stream will in many cases find that a more normalized (old-route speedrun) can last longer in an environment where they watch the runner play through the game many times over. There are a lot of games where watching people compete in "outdated" categories is much more viewer friendly than the current run.

IL games like Super Monkey Ball or F-Zero can also be like this... the current WR route is fine to watch in its completed form, but since it will lead to nearly constant pause-resets in a streaming enviornment, no reset categories (which almost always use older, safer routing) become much more watchable in the long term and to watch them in the making instead of only when fully finished.

Con - multiple categories

However, I think that there is a problem with "category bloat" despite this.

From my personal experience with Monaco (an IL speedrunning environment), for most given levels. 30 or so people grinded the level using the A route, and the record passed between multiple players multible times using the same route (with barely tighter cornering, etc). If we were archiving it, anyone along that range would have been an equal representative of the record. Eventually, the record is beaten a different way, eg coming up with a new route B. The B route becomes practicesby again all 30 or so runners and no one run the A route anymore. And so on to Route C.

However, because runs last a maximum of 4 minutes, no one in the Monaco community really cares to document the A or B routes compared to the current one. Differences between full routes can be as drastic as using a differnt item in a 1-invenotry slot game, or as simple as forcing the RNG to make parts of the level skippable. Attempting to get a record with an old route is physically impossible, and because we use in-game timing there is not even a PB incentive to do so - someone can have a lower IGT with a badly executing new route than brilliantly executed old route.

It is important that the difference between the B and C routes can take a huge amount of time to develop. On some levels we competed on the B route for 9-15 months before ever developing a new strategy. These are the levels where there is perhaps the most incentive to record who is the champion of the B route strategy despite the glut of C route and D route players with higher times. These LONG STANDING - old routes feel more worth chronicling due to the sheer man-hours put into them, and for many levels, also as a kind of testament for the older "pioneer" runners who eventually left the game. Some of these levels also have an aesthetic quality to them due to guard manipulation and health manipulatuion when contrasted to modern IL strategies that just bet on chance and take huge amounts of intentional damage.

I think that feeling is magnified a lot for people doing longer speedruns (especially of N64-era games) Even though most of the communnity accepts and competes in the new routes, there is a kind of great pride in being able to be remember who was able to run the level the best using the old strategy. However, I feel like for IL games and for very short (NES era games) it is more reasonable to simply never chronicle the old strategies, since it is easier to get the entire community to adapt to start prodicing runs using the new strategy.

It's also important that for more free-form games (Metroid Prime/Super Mario 64)... that the old category of 70 star is itself made of multiple different 70 star routes, eg... Where those 70 stars are gotten from can vary between two 70 star runs.. I think further representing things like 70 stars toxic maze, 70 stars tiny huge island red coins, 70 star shifting sand land secret (or all the combinations of these) would be completely ridiculous.

I'm also not sure what an archtype website's role in demonstrating the differnce between old and current route's should really be... In some games, oldish routes are only interesting when contrasted against multiple other old routes, while in other games (Super Mario 64) their are actually active runners in the old route category alongside the modern main categories...

I'm really torn in two directions by this topic...

I hate to give a response that implies I think that games I don't run should have multiple categories and the noes that I do run shouldn't but I can't really work my thoughts out right now.

Pro - multiple categories

The mindset of doing something like a 70 star Super Mario 64 run, or Banjo Tooie with or without Cutscene Warp 14 VS 70 is completely different. Both are of course no longer near their old status of being the "any% category" but the combination of having multiple runners who've practicted the Old Any% route, and the greater volume of mechanics in the classic route, means that there is some reason to preserve it.

As a viewer I really enjoy having multiple categories, especially for games that can evolve into credit-warps (Skyward Sword / Ocarina of Time) or for (mostly PC) games that transform from having out of bounds in one or two places, to runs (2-3 years later) that find ways to play almost the entire game out of bounds  (portal/Halo 1). I really enjoy being able to watch something that I can partially connect to the way I played the game casually.

Generally for games like this (or human total control in Super Mario World).. I think the full brokenness of the current WR route is interesting to the "casual viewer" once in something like a marathon to show off the extent that the game can be pushed etc... But people hoping to watch a runner's stream will in many cases find that a more normalized (old-route speedrun) can last longer in an environment where they watch the runner play through the game many times over. There are a lot of games where watching people compete in "outdated" categories is much more viewer friendly than the current run.

IL games like Super Monkey Ball or F-Zero can also be like this... the current WR route is fine to watch in its completed  form, but since it will lead to nearly constant pause-resets in a streaming enviornment, no reset categories (which almost always use older, safer routing) become much more watchable in the long term and to watch them in the making instead of only when fully finished.

Con - multiple categories

However, I think that there is a problem with "category bloat" despite this.

From my personal experience with Monaco (an IL speedrunning environment), for most given levels. 30 or so people grinded the level using the A route, and the record passed between multiple players multible times using the same route (with barely tighter cornering, etc). If we were archiving it, anyone along that range would have been an equal representative of the record. Eventually, the record is beaten a different way, eg coming up with a new route B. The B route becomes practicesby again all 30 or so runners and no one run the A route anymore. And so on to Route C.

However, because runs last a maximum of 4 minutes, no one in the Monaco community really cares to document the A or B routes compared to the current one. Differences between full routes can be as drastic as using a differnt item in a 1-invenotry slot game, or as simple as forcing the RNG to make parts of the level skippable. Attempting to get a record with an old route is physically impossible, and because we use in-game timing there is not even a PB incentive to do so - someone can have a lower IGT with a badly executing new route than brilliantly executed old route.

It is important that the difference between the B and C routes can take a huge amount of time to develop. On some levels we competed on the B route for 9-15 months before ever developing a new strategy. These are the levels where there is perhaps the most incentive to record who is the champion of the B route strategy despite the glut of C route and D route players with higher times. These LONG STANDING - old routes feel more worth chronicling due to the sheer man-hours put into them, and for many levels, also as a kind of testament for the older "pioneer" runners who eventually left the game. Some of these levels also have an aesthetic quality to them due to guard manipulation and health manipulatuion when contrasted to modern IL strategies that just bet on chance and take huge amounts of intentional damage.

I think that feeling is magnified a lot for people doing longer speedruns (especially of N64-era games) Even though most  of the communnity accepts and competes in the new routes, there is a kind of great pride in being able to be remember who was able to run the level the best using the old strategy. However, I feel like for IL games and for very short (NES era games) it is more reasonable to simply never chronicle the old strategies, since it is easier to get the entire community to adapt to start prodicing runs using the new strategy.

It's also important that for more free-form games (Metroid Prime/Super Mario 64)... that the old category of 70 star is itself made of multiple different 70 star routes, eg... Where those 70 stars are gotten from can vary between two 70 star runs..  I think further representing things like 70 stars toxic maze, 70 stars tiny huge island red coins, 70 star shifting sand land secret (or all the combinations of these) would be completely ridiculous.

I'm also not sure what an archtype website's role in demonstrating the differnce between old and current route's should really be... In some games, oldish routes are only interesting when contrasted against multiple other old routes, while in other games (Super Mario 64) their are actually active runners in the old route category alongside the modern main categories...

LotBlind

2018-03-25 09:53:23 am

You should probably read the title again and see if your reply is still relevant.

reality

2018-03-30 09:00:43 pm

Nostalgia Kills

My last post was all over the place - I adressed both games having multiple categories and gamess having multiple runs within the same category.

I think if I keep things in line with the OP -

For highly grinded / competetive games, the recorded videos of different runners PB's are largely the same. At it's most extreme this means that there are games
where the "WR" run's aesthetic qualities outweighs other runs throughout its entire duration (Super Castlevania 4).

In less competetive games, a WR recording can be slightly sloppy or unaesthetic in certain areas despite being the fastest overall per the category (sometimes only in ways that fellow runners will notice aesthetically). This is easier to see if runner's use "splits" (or for games that have definite seperate levels over their total duration)

In most cases a "bad" WR can still be used as a tutorial video.

There are outlier cases where 1st, 2nd, 3rd, get comparable times using different tricks. Or else the WR uses tricks on level 1-3, but doesn't make use of tricks on level 4, while the Second best time (or even an unranked time) features tricks on level 4, and/or uses multiple attempts to land the level 1-3 tricks. The Second best time possibly doesn't use level 1-3's tricks at all if the Level 4 trick was that much of a benefit.

This is especially the case for certain tricks that aren't quite TAS-only, but the community considers them to be TAS-only. Such tricks can only become part of the main run once the community has been "forced" to do so by someone getting the WR with them. This can be years after they are first known. Sometimes even years after they were succusfully shown performed by humans.

In the case of a single runner's succesive PBs being different from each other - it's the difference between a run made from the sum of all their best splits, and one that has to make do with mostly light green/red and a few gold splits.It's almost unheard of (except for the first run after a major skip has been found) for the runner's WR to be made of ALL gold splits relative to their other runs.

To use a Monaco example- On the level Pearls Before Swine - Frank and Mopo finish within 2 seconds of each other. However , the level is made up of 4 floors - and almost all of the time Mopo loses to Frank is due to bad routing on the first floor, while Mopo's "splits" of floors 2-4 are actually faster. An ideal third speedrun of the level would combine their strategies.

However, because speedrunners prefer to use the current WR as reference, most of the runners are unwilling to grind the level, because watching either video in isolation shows a near perfect execution of their own routing, which makes people think that the level has already been pushed to its limits.

I think the use of the Summoning Salt videos made me think that "more than one run per game/category" implied hosting for example a Metroid Prime any% with 2006 strats, a Metroid Prime any% with 2012 strats, and one with 2017 strats.

As far as an archiving website goes - I think it isn't the archiving websites job to host multiple runs in such a way. The speedrunners within their own communities should make sure to double check, (for the multiple routes getting similiar timing ad to properly combine them later), but usually the timeline of that happening is fairly quick, and thus needless for an archive website - Ideally a submitted run that picked and choosed strategies (unless for an extremely long game that justified some level of safety strats) would be rejected on the basis of a cleaner/more complete run being submitted in the (very near) future, even if the submitted run was currently the WR .

Using an entire second place run (or a multiple year old run) as a techinqiue/strategy reference, is probbably only practical for IL-type runs and very short NES-era games. Otherwise it's obviously better to to have knowledge base / community videos for individual parts of the run or glitches that happen hours into the run, etc. But again, this is something that is the responsbility of individual game communityes, not an archive website.

My last post was all over the place - I adressed both games having multiple categories and gamess having multiple runs within the same category.

I think if I keep things in line with the OP -

For highly grinded / competetive games, the recorded videos of different runners PB's are largely the same. At it's most extreme this means that there are games 
 where the "WR" run's aesthetic qualities outweighs other runs throughout its entire duration (Super Castlevania 4).

In less competetive games, a WR recording can be slightly sloppy or unaesthetic in certain areas despite being the fastest overall per the category (sometimes only in ways that fellow runners will notice aesthetically). This is easier to see if runner's use "splits" (or for games that have definite seperate levels over their total duration)

In most cases a "bad" WR can still be used as a tutorial video.

There are  outlier cases where 1st, 2nd, 3rd, get comparable times using different tricks. Or else the WR uses tricks on level 1-3, but doesn't make use of tricks on level 4, while the Second best time (or even an unranked time) features tricks on level 4, and/or uses multiple attempts to land the level 1-3 tricks. The Second best time possibly doesn't use level 1-3's tricks at all if the Level 4 trick was that much of a benefit.

This is especially the case for certain tricks that aren't quite TAS-only, but the community considers them to be TAS-only. Such tricks can only become part of the main run once the community has been "forced" to do so by someone getting the WR with them. This can be years after they are first known. Sometimes even years after they were succusfully shown performed by humans.

In the case of a single runner's succesive PBs being different from each other - it's the difference between a run made from the sum of all their best splits, and one that has to make do with mostly light green/red and a few gold splits.It's almost unheard of (except for the first run after a major skip has been found) for the runner's WR to be made of ALL gold splits relative to their other runs.

To use a Monaco example- On the level Pearls Before Swine - Frank and Mopo finish within 2 seconds of each other. However , the level is made up of 4 floors - and almost all of the time Mopo loses to Frank is due to bad routing on the first floor, while Mopo's "splits" of floors 2-4 are actually faster. An ideal third speedrun of the level would combine their strategies.

However, because speedrunners prefer to use the current WR as reference, most of the runners are unwilling to grind the level, because watching either video in isolation shows a near perfect execution of their own routing, which makes people think that the level has already been pushed to its limits.

I think the use of the Summoning Salt videos made me think that "more than one run per game/category" implied hosting for example a Metroid Prime any% with 2006 strats, a Metroid Prime any% with 2012 strats, and one with 2017 strats.

As far as an archiving website goes - I think it isn't the archiving websites job to host multiple runs in such a way. The speedrunners within their own communities should make sure to double check, (for the multiple routes getting similiar timing ad to properly combine them later), but usually the timeline of that happening is fairly quick, and thus needless for an archive website - Ideally a submitted run that picked and choosed strategies (unless for an extremely long game that justified some level of safety strats) would be rejected on the basis of a cleaner/more complete run being submitted in the (very near) future, even if the submitted run was currently the WR .

Using an entire second place run (or a multiple year old run) as a techinqiue/strategy reference, is probbably only practical for IL-type runs and very short NES-era games. Otherwise it's obviously better to to have knowledge base / community videos for individual parts of the run or glitches that happen hours into the run, etc. But again, this is something that is the responsbility of individual game communityes, not an archive website.

LotBlind

2018-03-31 10:25:29 am

Well, doesn't sound very different from me then. You conclude on the same point there, that most of the time demonstrations should not equal complete speedruns – it's lazy for one thing and clumsy for another. Furthermore there's tricks and strategies that don't crop up in any actual runs, at least not at the time. So at least from that point of view, multiple runs are certainly not required to exist, unless the effects of choices made early into the run splash down into the later parts in a way that makes it difficult to split it into "Trick A", "Trick B" etc. in which case a more longitudinal demonstration (which is still usually not the same as the whole run) is in order.

However, there's more to be said on this topic, and perhaps even arguments for multiple runs. I never meant whether or not SDA should host them, just whether the old records have some intrinsic value.

SDA Forum Index -> Community -> Gaming Discussion -> Do we need more than one run per game/category?