Human Limits for memo tracing in 3BLD event

[abunickabhi]

What do you guys think is the limit of tracing a 3x3 cube? (memoing the cube is not necessary in this process,and you can consider it just one passing over the entire cube, corners+edges)
Also for making it easier to reach a conclusion, lets assume the algs we get everytime is a 10 algs.
Some things you can outright do to improve, is to identify corners using 2 stickers, or avoid moving the cube that much, or getting the orientation quickly when you remove the cube cover.
(I have been doing tracing only practice since the last few months in order to make it more brain dead, so that my mental resources are more on the thinkahead process. On 10 algs scramble, I generally get about 8-9.2 seconds to one-pass/trace, which is well above the human limit.)

 

[abunickabhi]

Interesting responses to this topic taken from the FB group 'BLD support'.

Kavin says, "(0.20 reaction time)X(20 pieces) - (a lot of talent) = sub 4."

Graham says, "i know gianfranco can comfortably visually memo corners (averaging like 1.5 or something for visual corners only) and i dont think its too crazy to say its possible to do quick visual of the entire cube. Maybe sub-4, maybe 4-5 globally, idk really. That being said, probably not a very feasible way to do 3BLD anyway, because I can't imagine one would be able to do that but not suffer greatly in accuracy and also not lose a bunch of time to thinkahead during execution given you'd have to come up with the entire solve in your head on the go. "

Jeff Park says, "For the scramble, if DB is solved (or last) and the pieces to trace are next to each other, then sub4 tracing should be doable. At my peak, the main bottleneck I felt was not being able to physically rotate the cube fast enough."

20 people voted on sub-4 being the limit.
9 people voted on sub-5 being the limit.
4 people voted on sub-6 being the limit.
2 people voted on sub-7 being the limit.

 

[abunickabhi]

One of the fastest tracer, in the BLD community is Gianfranco Huanqui from Peru.

Amazing technique and optimizations from Gianfranco in his solves.
His amazing WCA track record: https://www.worldcubeassociation.org/persons/2013HUAN29
His YouTube channel: https://www.youtube.com/user/GHROCKON
Learning from Gianfranco solve is a big process, he has contributed a lot to many algsheets, and given good algs. In this video I could only cover a bit about the memorisation process, it will take tons of more videos to learn from the execution phase. Memorisation in 5 seconds is always fascinating to me, and I always am in awe when ~20 letters can be memorised in low 5 seconds in a competition setting. All the solves were successes except for the last solve taken from WC2019. There are a lot of more solves where he has gotten high 5, and low 6 second memo as well. I hope more BLDers appreciate these solves as they involve very less cube tilt, no finger tracking/tapping, fast noddons and minimal eye movements. Also, I misspoke a bit about visual memorisation, in visual memo, you just see the colours, and there are no letters to be memorised. Also visual memo is theoretically faster, as we do not have to process letter pairs and stickers, but just the color, or the comm that we have to execute.

 

[abunickabhi]

For the past few days, I am investigating, which form of tracing will be fast, ECCE or CEEC. Turns out that I am about a second faster in tracing ECCE, since I started out in this way in 2013.

I do know that the parity algs for CEEC are of better quality and a bit faster than the ones for ECCE, but I am willing to sacrifice that for faster memo times.

 

[Skewbed]

For the past few days, I am investigating, which form of tracing will be fast, ECCE or CEEC. Turns out that I am about a second faster in tracing ECCE, since I started out in this way in 2013.

I do know that the parity algs for CEEC are of better quality and a bit faster than the ones for ECCE, but I am willing to sacrifice that for faster memo times.

I think that CEEC is better also because you can detect parity more easily. It is easier to know when you are done tracing corners, so knowing if you have parity then makes tracing edges easier.

 

[abunickabhi]

I think that CEEC is better also because you can detect parity more easily. It is easier to know when you are done tracing corners, so knowing if you have parity then makes tracing edges easier.

Yes we get the parity information a bit earlier in the memo, as there are 8 corners as opposed to 12 edges. Granted the number of targets for a given scramble might vary a bit, and there might even exist scrambles where the number of corner targets exceed the number of edge targets (4+ solved edges, super rare).

What do you think are the limits though @Skewbed ?

 

[Skewbed]

Yes we get the parity information a bit earlier in the memo, as there are 8 corners as opposed to 12 edges. Granted the number of targets for a given scramble might vary a bit, and there might even exist scrambles where the number of corner targets exceed the number of edge targets (4+ solved edges, super rare).

What do you think are the limits though @Skewbed ?

Even though I think CEEC is better for most cubers, ECCE might end up being optimal because you can use extremely short-term corners methods.

In your video, you mention visual memo being human-possible for corners. I think that there might be an even faster way to do visual memo for corners that doesn't involve tracing. Instead of tracing, you could try to visually remember each corner in a predetermined order and figure out the comms as you solve it. This makes the hand and eye movements the same every time, which speeds up memo, but the amount of thinking required during execution will slow it down.

Optimal ECCE Method:
1. Trace edges and memorize using audio (nothing different)
2. Visually remember corners without tracing (speeds up memo)
3. Solve corners while figuring out what to do (slows down execution)
4. Solve edges (nothing different)

I think that if it is possible to do, sub-4 memo is possible, but you sacrifice execution time during corner execution.

 

[abunickabhi]

Even though I think CEEC is better for most cubers, ECCE might end up being optimal because you can use extremely short-term corners methods.

In your video, you mention visual memo being human-possible for corners. I think that there might be an even faster way to do visual memo for corners that doesn't involve tracing. Instead of tracing, you could try to visually remember each corner in a predetermined order and figure out the comms as you solve it. This makes the hand and eye movements the same every time, which speeds up memo, but the amount of thinking required during execution will slow it down.

Optimal ECCE Method:
1. Trace edges and memorize using audio (nothing different)
2. Visually remember corners without tracing (speeds up memo)
3. Solve corners while figuring out what to do (slows down execution)
4. Solve edges (nothing different)

I think that if it is possible to do, sub-4 memo is possible, but you sacrifice execution time during corner execution.

Yes visual memo has a lot of potential.

Back in the days when letter pairs were not invented, people used to just remember the target stickers of the 3OP alg that they were going to execute, and it was very unstructured, and a minimal method. Nowadays with very fast exec times, it is possible to do full visual memo of 3x3 in 4 seconds as well with enough practice.

 

[abunickabhi]

I do not think letter quads improve tracing as compared to letter pairs. The tracing speed overall remain the same.