/vr/ - Retro Games

>>4591481
Fuck speedruns. Go play arcade games.

>>4591486
or speed run an arcade game.

>>4591481
The hardest speed games are Super Mario 64 and Super Metroid. Expect to pour hundreds of hours into them to get even a semi-respectable time. Want to challenge for a world record? See you in 4-5 years.

>>4591481
The thing is only the most popular or simplest games have people that near the skill ceiling. For most games, we can only speculate how high the skill ceiling goes.

>>4591521
>Super Metroid
It's a good speedgame, but it's not on the level of SM64.

2D games by their very nature are easier to optimize in a speedrun than 3D games due to the significantly lower possibility window (less variability in movement, etc.)

Half-Life 1, Spyro 1, F-Zero X/GX, Banjo Kazooie, Mirror's Edge, etc, all have higher skill ceilings than Super Metroid.

and players with respectable times have thousands of hours, not hundreds. this goes for Super Metroid too.

>>4593784
>The thing is only the most popular or simplest games have people that near the skill ceiling.
This isn't true at all.

The reality is that games with higher skill ceilings attract more players and have greater longevity. Games that reach their skill ceiling early die early.

>>4593784
>The thing is only the most popular or simplest games have people that near the skill ceiling.
This isn't true at all.

The reality is that games with higher skill ceilings attract more players and have greater longevity. Games that reach their skill ceiling early die early.

There will be exceptions to the rule, of course.

>>4594265
One weird thing about SM64 is that due to the fixed camera angle increments, speedrunning it gets as close to digital input as 3D games can get.

>>4594595
>he fixed camera angle increments
this actually isn't the case most of the time.

>>4594595
What are you talking about? SM64's camera is insanely dynamic and requires a greater level of input detail than most 3d games.

>>4594265
>>2D games by their very nature are easier to optimize in a speedrun than 3D games due to the significantly lower possibility window (less variability in movement, etc.)

the possibility window? sure, the state space of next possible movements I could make includes hundreds or thousands of mario tiptoeing animations in a thousand different directions, at about 3 or 4 different thresholds of speed. so? why do you think those extra controller states being possible makes the skill ceiling higher? almost none of them are useful. aren't you just holding a direction 99% of the time anyway? also, the game is only 30fps. people like to pretend 1-frame tricks are difficult to get but then fail to recall that detail.

I can see half life, but mirror's edge? I can only take that seriously in the context of this post >>4593784 and assume you meant that the ceiling is high but nobody's near it. SM has had a ton more effort poured into it.

>>4594680
The SM64 speedrun uses thousands of different directional inputs, the Super Metroid speedrun can only use up to 8. This is quantifiable. There are also hundreds of thousands of more positions your character can exist inside the game due to an additional axis. This is even the case while following general established paths (which 2D games also do).

The human theory TASes for the highest skill ceiling 3D speedgames are always significantly further away from the RTA record as the highest skill ceiling 2D speedgames. This is because 3D games have much higher variability and are harder to play near-perfectly.

3D speedgames, in general, are harder to optimize than 2D speedgames. There is some grey area and overlap, but the extremes of 3D will always be harder than the extremes of 2D.

>also, the game is only 30fps. people like to pretend 1-frame tricks are difficult to get but then fail to recall that detail.

This isn't relevant. 1 frame "tricks" have almost nothing to do with what make speedrunning hard. Movement and high execution core mechanics is what makes speedrunning hard.

>I can see half life, but mirror's edge

Yes, Mirror's Edge.

>>4594714
You obviously didn't get it, so let me try it. If you only eat chocolate ice cream, and stand a has three flavors with one chocolate and stand b has 200 flavors with one chocolate, which stand is harder to make a decision at?

>>4594721
That's a bad analogy. Here's a better one:

There are two equally sized spinning wheels at the county fair. One wheel is divided into 8 sections. The other wheel is divided into thousands of sections.

You have to hit one specific target on each wheel with a dart. Which wheel offers the easier target?

>>4594731
>You have to hit one specific target on each wheel with a dart.
Each section is a target. Sorry if that wasn't clear.

>>4594714
>The SM64 speedrun uses thousands of different directional inputs, the Super Metroid speedrun can only use up to 8. This is quantifiable.
and like I said before, that doesn't map onto difficulty. why would it? there are huge ranges of joystick angles that work for every required action, and there are even larger ranges of joystick angles that are completely irrelevant because they aren't useful. you can spin in circles for days, making a huge state space of inputs, and it doesn't make the game harder, because it's only asking you to hold angle X because it's the only one required.
>There are also hundreds of thousands of more positions your character can exist inside the game due to an additional axis.
so? you can be in more places, once again, this doesn't map onto difficulty. the game has half as many input frames over time, meaning there can only be at most half as many required inputs over any amount of time to define any sequence of gameplay. why are you dismissing framerate entirely?
>The human theory TASes for the highest skill ceiling 3D speedgames are always significantly further away from the RTA record as the highest skill ceiling 2D speedgames.
examples? the mario 64 120 tas isn't human theory and does impossible things like exact 30hz mashing of multiple buttons (30hz instead of 60 because it's only 30fps)
>Yes, Mirror's Edge.
you literally move the camera toward the goal and hold the run button

>>4594783
>there are huge ranges of joystick angles that work for every required action
This isn't true, and the fact that you would say this demonstrates a fundamental lack of understanding for how 3d games are optimized.
>it's only asking you to hold angle X because it's the only one required.
The optimal angle is harder to achieve with a joystick than the optimal angle on a D-Pad. You need considerably more accuracy.
>so? you can be in more places
Yes, which means there are more sub-optimal places you can be, offering a much smaller margin of error for optimal play.
>examples?
The SM64 human theory TAS.
>the mario 64 120 tas isn't human theory
120 star has a human theory TAS, yes.
>you literally move the camera toward the goal and hold the run button
Laughable ignorance.

>>4594783
The brute force input Super Metroid TAS isn't even 6 minutes faster than the RTA record. The human theory TAS is likely even slower.

The 120 star human theory TAS is 8 minutes faster than the RTA record, and the brute force input TAS is nearly 20 minutes faster.

>>4594797
https://www.youtube.com/watch?v=TNTv68Ngbs4 - Super Metroid 100% TAS

twitch tv/videos/128933680?t= - Super Metroid 100% RTA

fucking lol.

>>4594783
>why are you dismissing framerate entirely?
not him, but i've never found framerate to actually play a role in speedgame difficulty, whereas i frequently find 2d games easier to get good times in than 3d games.

>>4594797
>>4594804
I mean, 2D games are even considerably less complex to TAS than 3D games. You can ask any TASer this and they will tell you the same. The time investment and complexity of TASing a 3D game is much larger.

>>4594734
In that case yes, wheel 2's target is harder to hit - if your aim is bad. -
If you know what you're doing, it doesn't matter how small the target is, or how many potential moves there are. The most basic strategy will eliminate the vast majority of worthless moves before you even begin, which is why I went with flavors of ice cream. When it comes to squeezing out the last few frames of perfection, then your many targets come back into play, and that's for stuff like tapping L/R every other frame to move one pixel faster in Super Metroid. The major breakthroughs all come from route selection, and glitch and skip discoveries.

In other words, you're both right, but you're actually arguing about two separate things.

>>4595205
>If you know what you're doing, it doesn't matter how small the target is, or how many potential moves there are.
Except it will take more practice and understanding to get to the point where you know what you're doing, and even then you'll miss sometimes because of the intrinsic lower probability involved.
>When it comes to squeezing out the last few frames of perfection, then your many targets come back into play
Yes, I'm talking about skill ceilings and high level play. How difficult it is to learn how to do a bad speedrun of the game is not relevant to this conversation.
>The major breakthroughs all come from route selection, and glitch and skip discoveries.
Major breakthroughs don't matter because virtually everyone does them as soon as they're discovered, including mediocre runners. What sets the good players apart from the mediocre runners are how proficient they are in the nuances of movement.

>>4595250
I can't see the goalposts from here!

>>4595328
>thread is about skill ceilings in speedgames
>first comment in reply chain is about how 3d games are harder to optimize than 2d games
>"hurr don't shift the goalposts"

>>4595358
>first comment in reply chain is about how 3d games are harder to optimize than 2d games
>they're not
>keeps changing argument so he can win

>>4595432
>they're not
They are.
>keeps changing argument so he can win
My argument hasn't changed. I described in very specific detail why they're harder to optimize. You're the one who changed it to "muh major breakthroughs", which have very little to do with the skill ceiling in a game.

>>4595205
dude, anyone can learn the route of a game. that's not what makes speedrunning hard. you have a very fundamental misunderstanding of how optimization in 3d games works. there is always more time to save in a 3d game than a 2d game, because there are so many more actual ways to save time. you don't even speedrun a 3d game, so you obviously have no fucking idea what you're saying, but many of us have experience in both 2d and 3d games.