Left Tracing: An OPA Idea for 4x4 (that probably doesn't work)

Kyoobies

Regular
Joined
May 20, 2022
Messages
36
Location
Southern California
This is gonna be a quick ramble because it's late, but I just learned about OPA tonight and that people are trying to implement it in speedsolving. And that's awesome, especially since 4x4 is my favorite.

But OPA recognition methods seemed tough and super rough to fit into inspection, so I was curious if I could make anything better. Inspired by some random thing I did in highschool to prevent 5x5 edge parity because the alg was annoying to learn, I present Left Tracing.

I also have no idea if it works btw since I haven't tested it and only thought of it like an hour ago; but I really wanna share it haha. I'll make much better vids soon if it does work, when I have the time for it.

Main idea

How to read the wings
 
Cool thanks. Just to be clear, you have seen Ari speedcubing's OPA tutorial right?
I just came from that video yes. That requires looking at every wing on the cube, as well as tracing the cycles of where the matching wing is at that moment.

My idea looks at the same 12 wings in the same spot and the same order every time, and just uses that info alone. So you don't need to search around the cube at all or remember if you've completed a cycle and where.

Anyway the video is uploading now, and it's a bit of an 8 minute ramble fumbling through it all with 1 hand xP
Will take a minute to upload.

Hopefully it's clear though, it goes through an entire cube/OPA inspection with a reference next to it. 😁
 
Cool thanks. Just to be clear, you have seen Ari speedcubing's OPA tutorial right?

I also thought of a way to potentially disprove its ability to work but when I try it it messes a bunch up so idk?

Even if this idea as is doesn't work at all there might be a way to use information tracing in this way to be systematic; which I feel can be potentially faster and easier still even if you needed to trace every wing.

No idea haha, and it's too late to just brute force try a bunch of scrambles and see if it holds up, and I can't think about the cube's logic enough to prove or disprove the idea. I'll have to experiment more later. We shall see!
 

I also thought of a way to potentially disprove its ability to work but when I try it it messes a bunch up so idk?

Even if this idea as is doesn't work at all there might be a way to use information tracing in this way to be systematic; which I feel can be potentially faster and easier still even if you needed to trace every wing.

No idea haha, and it's too late to just brute force try a bunch of scrambles and see if it holds up, and I can't think about the cube's logic enough to prove or disprove the idea. I'll have to experiment more later. We shall see!
Also it that just zz eo because watching through it might be. :V

Every second going by I don't become more or less confident that it works, but I do become more confused. I should sleep lol
 
I've also gotten some pushback from the name, and thinking about it I feel Half Tracing also works because it's tracing/checking the eo but for a specific half of every edge. Gets rid of the meaningless relative direction, but also gets rid of the slightly funny dumbness that I enjoy.

Feel free to share any opinions on that, Left Tracing or Half Tracing.

While I sleep and die and check if it even works tomorrow 🙃
 
I think I figured out a way that it doesn't work, under the conditions I set. But I see what it affects so when I get time I'm gonna mess around more and see if I can isolate the variables and what's affected/how the affect each other.

Looks like I may indeed need to look at every wing on the cube, but I still like this approach of gathering information so I'm gonna see if I can work with it in different ways.
 
You really do have to do some kind of cycle tracing to correctly determine the parity. Any counting method where two wing pieces can be swapped without affecting the count is necessarily incorrect. This thing makes use of EO to reduce the amount of tracing needed (compared to 4BLD wings), but of course doesn't eliminate tracing entirely.
 
You really do have to do some kind of cycle tracing to correctly determine the parity. Any counting method where two wing pieces can be swapped without affecting the count is necessarily incorrect. This thing makes use of EO to reduce the amount of tracing needed (compared to 4BLD wings), but of course doesn't eliminate tracing entirely.
Dang
 
This is gonna be a quick ramble because it's late, but I just learned about OPA tonight and that people are trying to implement it in speedsolving. And that's awesome, especially since 4x4 is my favorite.

But OPA recognition methods seemed tough and super rough to fit into inspection, so I was curious if I could make anything better. Inspired by some random thing I did in highschool to prevent 5x5 edge parity because the alg was annoying to learn, I present Left Tracing.

I also have no idea if it works btw since I haven't tested it and only thought of it like an hour ago; but I really wanna share it haha. I'll make much better vids soon if it does work, when I have the time for it.

Main idea

How to read the wings
I can't see the video. It says that the video is not available, but I'm very interested in the topic.
 
Back
Top