OLL execution and PLL recognition at the same!

[Aergyl]

This is a trick for solving LL completely without pausing, but without any advanced 1LLL algorithms or recognition techniques. I'll start by explaining how it works for OLL 23, since that is the only case that, at least for me, is actually just as fast as doing regular OLL + PLL.

Step 1: AUF so that the headlights point towards you.

Step 2: Perform R' D R D' R' D R while taking a mental note on the color of the UFL sticker (which doesn't move at all).

Step 3: Perform U'.

Step 4: Do PLL recognition and R' D R D' R' D simultaneously. This can be done since all of the stickers used for two-sided PLL recognition, except for FUR, don't move at all. As for FUR, we simply imagine the color we took a mental note of being there.

Step 5: Cancel into PLL (works especially well if you use a lot of PLL's that start with R moves from homegrip).

In other words, we use the 2-twist algorithm (R' D R D' R' D R) U' (R' D' R D R' D' R) to solve OLL 23, but we start recognizing PLL halfway through. Therefore, if we compare this to doing regular OLL, the execution is slightly slower, but we can start PLL recognition faster.

The same technique can be used for OLL 24 and 25, by simply changing the U' to U or U2 in step 3. Also, it's possible to extend this technique to other OLL's by simply doing some fast OLL that makes it into OLL 23, 24 or 25 (just like some people do for ZBLL). For most cases, I think OLL + PLL is better, so I would only recommend doing this for fun!

 

[Aergyl]

Also, if you use this trick, I would recommend doing RUD A Perms, since there's a variant for almost every angle for which you can easily cancel 3 moves (saving 6 moves in total). The algorithms are:

R' D R U R' D' R U R' D R U2 R' D' R
R' D R U2 R' D' R U' R' D R U' R' D' R
R' D R U R' D' R U2 R' D R U R' D' R
R' D R U' R' D' R U2 R' D R U' R' D' R
R' D R U' R' D' R U' R' D R U2 R' D' R
R' D R U2 R' D' R U R' D R U R' D' R
R D R' U R D' R' U R D R' U2 R D' R'
R D R' U2 R D' R' U' R D R' U' R D' R'

There's also very long E perms that are very good because of a lot of moves getting cancelled. There are a few variants, but the ones I like the most are:

R' D R U' R' D' R U R' D R U R' D' R U R' D R U' R' D' R
R' D R U R' D' R U' R' D R U' R' D' R U' R' D R U R' D' R

There's also a nice Y Perm that cancels the same amount of moves from one angle:

R' D R2 U' R' U' R U R' F' R U R' U' R' F D' R

I'll update this if I find more algorithms that works well with this technique!

 

[Deleted member 36445]

Thanks for your contribution, but being able to predict pll having to use beginners method for oll is a pretty heavy tradeoff

There is a well known technique that can be used to cut down the possibilities of pll cases and using this with 2 side recognition can lead to nearly optimal recognition times

 

[Aergyl]

A

Thanks for your contribution, but being able to predict pll having to use beginners method for oll is a pretty heavy tradeoff

There is a well known technique that can be used to cut down the possibilities of pll cases and using this with 2 side recognition can lead to nearly optimal recognition times

First of all, I totally agree that there are other ways of solving OLL 23 even faster than with my method (most notably with ZBLL). And if your PLL recognition time is very fast (possibly by using ROLL/JOLL), OLL + PLL can be faster in theory.

Also just to be clear, with the exception of OLL 23, I think OLL + PLL is always better (for OLL 24 and 25 it can be pretty close though).

However, doing this trick for OLL 23 isn't a tradeoff of at all over OLL + PLL, at least not for me. I can execute the normal algorithm (starting from homegrip) in about 0.7 and the 2-twist in about 0.9. So as long as I pause at least 0.2 seconds between OLL and PLL, it's actually faster. I'll be happy to listen to any arguments against that, but just saying that it's a pretty heavy tradeoff because it's part of the beginners method doesn't really make any sense.

Finally, the aim of this trick is first and foremost to execute LL without pausing, and with as little advanced stuff as possible, not to make it as fast as possible

 

[OreKehStrah]

At that point just use the sune combo pure U alg and not pause at all or have to think about recog.

 

[Aergyl]

At that point just use the sune combo pure U alg and not pause at all or have to think about recog.

How would you do PLL recognition at the same time as that without slowing down considerably? Also RUD 2-flip is just as fast, so I don't understand what you're trying to say.

 

[OreKehStrah]

How would you do PLL recognition at the same time as that without slowing down considerably? Also RUD 2-flip is just as fast, so I don't understand what you're trying to say.

If the RUD solution was just as fast people would be using that as the alg instead of sune combos

 

[Aergyl]

At that point just use the sune combo pure U alg and not pause at all or have to think about recog.

Ohhh I think I understand now. Do you mean doing PLL recognition before starting the 2-twist? That is certainly possible, but not the goal of the trick, which is to do it at the same time as OLL execution. Also, doing it before is obviously just worse than just doing OLL + PLL normally, and doesn't even work well for every angle.

 

[Aergyl]

How would you do PLL recognition at the same time as that without slowing down considerably? Also RUD 2-flip is just as fast, so I don't understand what you're trying to say.

I realize that I was a bit unclear about the goal of the technique, so I'll try to clarify that. The main goal is to have LL completely pauseless after identifying the OLL case. In particular, any extra recognition pause before starting OLL execution is not allowed. As far as I know, using the RUD 2-twist is the only way of being able to do this consistently.

 

[Deleted member 36445]

As far as I know, using the RUD 2-twist, zbll and 1LLL are the only way of being able to do this consistently.

 

[Aergyl]

If the RUD solution was just as fast people would be using that as the alg instead of sune combos

Not neccessarily true, but I see your point. A lot of 3BLD solvers use RUD though, and it's fast to execute. However, the reason why I chose RUD, is because you can do PLL recognition simultaneously. That's not possible with the RU 2-twist.

 

[OreKehStrah]

Not neccessarily true, but I see your point. A lot of 3BLD solvers use RUD though, and it's fast to execute. However, the reason why I chose RUD, is because you can do PLL recognition simultaneously. That's not possible with the RU 2-twist.

That only matters if you don’t recognize zblls quickly

 

[Aergyl]

Well

That only matters if you don’t recognize zblls quickly

Well, I literally said before, and I quote: First of all, I totally agree that there are other ways of solving OLL 23 even faster than with my method (most notably with ZBLL).

If you read the clarification I made, I said that any pause after OLL recognition was not allowed. Even the fastest people knowing ZBLL will pause sligthly (at least sometimes) to do ZBLL recognition after identifying the OLL. Again, I'm not saying that my technique is the best thing ever (ZBLL is obviously faster) and that everyone should use it. It simply fullfills the goal I made, and ZBLL does not.

 

[Aergyl]

Funny, but 1LLL and ZBLL requires additional recognition time, at least sometimes, which goes against the goal of the technique (see the clarification).

 

[Aergyl]

Actually, I take back what I said about OLL 23 being just as fast with this technique. With a lot of testing, I realized it's actually even faster than regular OLL + PLL for me. I pause on average around 0.5 seconds between OLL and PLL. Therefore, the sligthly slower execution (about 0.2 seconds) of the 2-twist followed by PLL without pausing is faster.

 

[Aergyl]

Thanks for your contribution, but being able to predict pll having to use beginners method for oll is a pretty heavy tradeoff

There is a well known technique that can be used to cut down the possibilities of pll cases and using this with 2 side recognition can lead to nearly optimal recognition times

By the way, what do you mean by nearly optimal recognition time? It would be interesting to know if it can be below 0.2s on average with ROLL/JOLL, because if it can't, my method has the potential to be even better for OLL 23 (of course still worse than ZBLL though).

 

[Deleted member 36445]

By the way, what do you mean by nearly optimal recognition time? It would be interesting to know if it can be below 0.2s on average with ROLL/JOLL, because if it can't, my method has the potential to be even better for OLL 23 (of course still worse than ZBLL though).

for that technique you always end with headlights on left if you don't have diagonal swap. so using 2 side recogntion you can have the best reaction possible because the number of cases are reduced.
also you can reduce the number of possibilites because you know where some edges will end

 

[Aergyl]

for that technique you always end with headlights on left if you don't have diagonal swap. so using 2 side recogntion you can have the best reaction possible because the number of cases are reduced.
also you can reduce the number of possibilites because you know where some edges will end

I apologize if the question was unclear. I already know how ROLL/JOLL works. The question is what you mean by nearly optimal recognition time. How fast is that on average? In other words, what's the time difference on average between using ROLL/JOLL vs. not pausing at all between OLL and PLL?

 

[Aergyl]

for that technique you always end with headlights on left if you don't have diagonal swap. so using 2 side recogntion you can have the best reaction possible because the number of cases are reduced.
also you can reduce the number of possibilites because you know where some edges will end

Actually, never mind that question. You seem pretty confident that ROLL/JOLL is a lot better than my method for OLL 23. Can you explain why?

 

[ruffleduck]

How much slower must you execute OLL to have enough time to predict the PLL before finishing the OLL?