• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Proving UF5

abunickabhi

Super Moderator
Staff member
Joined
Jan 9, 2014
Messages
8,983
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
Hello everyone,

This is a serious attempt to make this algset work or atleast use a subset of it. The method is used in blindsolving to solve 4 edges at once using a ~10.4 move commutator.

Disclaimer:

5-style is an experimental algset, it does not work at the elite level currently. I love 5-style algset. It solves 5 edges on the 3x3 and will be mainly used in MBLD event. A lot of top BLDers are not convinced that it is useful and would make anyone faster. I do not want to feel good/superior or boost my ego by putting out these updates. I do not get enough free time to drill the algs or revise them on Anki. I have propagated misinformation about 5-style and letter quads in the past. I hope to not do it again. A lot of years have been spent developing it and there is no inertia currently to learn it. I have one passed 51k algs, but that does not mean I can reproduce them in actual solves, that factor is variable and depends on the practice and sharpness of the cuber.

My goal for 100 points in MBLD is wishful thinking and I have made no progress towards it. As a BLDer, I have made no progress from 2017 because of life and a lot of work on 5-style. I am not the best person to do this set, this is just an attempt. Please do not troll, I will forgive and not block, but I dislike trolls. It is a lot of mental effort to make UF5 work, and trolling is the last thing I want to receive. I respect all cubers and understand my limitations and weaknesses. The wishful thinking is to get sub-10 exec in MBLD so that 100 points is possible in MBLD someday using LQ and UF5 may not work and I have to accept this fact. I am ready to accept that 5-style can tarnish my image in the cubing community as I am attempting something infeasible with no common sense and will to listen to cubers faster than me.

Currently it is waaayyyy harder to learn 5-style compared to 3-style. There have been no proven gains. It saves about 3.5 moves per comm but this does not imply that there is time saved in execution.

I hid the development and idea of the algset for 4 years as I was unsure if it was viable. Since 2018, this algset has been receiving negativity from all BLDers who are faster than me except for a few. I do want to implement this algset one day. A lot of years were spent in developing the algset though, like Letter quads, UF5, differences from 2e2e' etc.

Just to give a peek, the easiest 5-style algset is M U M' U' and there are 39 more 4 movers from one buffer. The average movecount is 10.38 STM which is a lot of move saving. The algs are more speed optimal than move optimal, although it is hard to decide clearly.

5-style is a cool concept but it is not fully executed yet :(

Docs (concepts here are provocative so take them with a pinch of salt)
https://docs.google.com/document/d/1x-gGoO4mzYkHuQoZXQcvrQLDLU_x_UYb460y58OJ8ss/edit?usp=sharing
https://docs.google.com/document/d/1b3tT8Wv18WdzFyY7FujyptwzYeWFH8K-UXTddTvwUWw/edit?usp=sharing
https://docs.google.com/document/d/1bfDsydw6pxBftd8Xwik95FNjILGkdrJMBO5EORbftII/edit?usp=sharing
 
Last edited:
Bunch of random questions and critique.



Has any research been done with different buffers? On your document you make it sound like the only reason you are pushing DF/DFR is because that is what you started with. I feel like even if difference is negligible between buffers for 5style it would make the most sense to use UF/UFR so you can fill in the gaps where you don’t know the 5style cases with the best algs.



Wouldn’t it make more sense to move the 5style algs over to a google sheet? Right now it’s a huge list that’s difficult to navigate but with sheets you could make 26 tabs (what the first target of the 5cycle is) and it would make it a lot easier to sort through. You could also enable commenting so people could suggest good algs that they find.



Also have some critiques for the 5 year plan :P



Year 1 - remove the music stuff, not really pertinent to cubing. Just practicing cubing will make their fingers more nimble.

Year 2 - recommends orozco and then 3style with UF UFR, if the method uses DF DFR then their 3style should probably reflect that.

Year 3 - no comment

Year 4 - that’s a LOT of algs, even with memoing with quads that’s 164 algs per day. Wouldn’t it make sense to spread it out a bit more and maybe combine some previous years or maybe learn letter quads as you learn 5style? I feel like the rate they have to work at is a bit too much and is going to cause the few people who may attempt to learn this set to burn out.

Year 5 - drill all algs and make them sub 1 on the gikker cube. I don’t think most of these algs are sub 1able. Do you plan fixing the algs yourself or having others contribute to your alg database?
 
Is this real?
Isn’t there already ~800 cases for 3-style, and how many for 5-style? 100,000?

Yes but I’m not sure if it’s doable/viable

378 for corners, 440 edges. Not too crazy since it’s intuitive. Full floating is 2768 which only a few people know or are close to knowing still intuitive though.

5style is almost never intuitive and has about 12600 edge algorithms and about 80000 corner algs.

Just for references on big alg sets zbll is 493 and 1LLL is 3915
 
I know that 3-style algs are pretty intuitive and short so that learning all 818 cases is not all that hard, much easier than ZBLL despite having “only” 493 algs.

But if 5-style algorithms aren’t intuitive comms like 3-style is, and the algs are longer, then 5-style is actually the equivalent of learning ZBLL times 190.

And also memory retention would be practically impossible because you run into each case like what.... in 0.00005% of solves?

If you want to make a breakthrough in blind solving, then it’ll have to be something other than 5-style...
 
Thanks for the critique, Samuel.
I do not plan on using the DF/DFR buffer for 5-Style. No way, DFR is a too ancient buffer and 5 cycle algs with it are bad.

In fact, two 3-Style corners algs are already faster than one 5-Style alg. I stackmatted and saw it. (So, the end goal is to have floating/UFR repertoire with floating 5-style edge algs and UFR corners) For the time being UF 5-cycles excluding the DF piece is not being developed.

The good thing about 5 Style is that UF and DF both share the same buffer in many cases like 20-30% of them. So we can have the same algorithm for the UF, that the DF has which involves the UF piece.

I intend to share the google sheet by next year and help other people contribute to better algs and also UF specific algs which do not involve the DF buffer, which I can treat as floating algs if I stick to UF. This is a personal endeavor as of now, and I personally prefer DF over UF although it does not make that much of a difference for 5-Style as compared to 3-Style (DF vs UF).

I just created this thread to get the opinion of BLDers on whether they will be willing to accept this idea in the future if this method gets developed and simplified. But for the proof of concept, I want to keep it personal for a year and get good official results using these algorithms and memorization techniques.

I have been making Letter quads for the last 3 years and have already improved my memo significantly. For the execution part in year 5, I think most of the algs are sub-1 able, and I will try and grind to get there.
 
But for the proof of concept, I want to keep it personal for a year and get good official results using these algorithms and memorization techniques.
That's a good approach. You'll gain way more credibility if you can prove it is feasible before you advocate it.
 
I don't know much about blind but why don't you just do 4style instead?

Old Pochmann is a 2-cycle method, meaning you cycle 2 Pieces and solve one at a time - this only works if you also permute an even number of pieces of the other type (3BLD only for simplification, Corners and Edges)
Now 3-Style up's things by cycling 3 pieces and solving 2 (Orozko and Eka left out for simplification) at a time by using commutators. Now, 4-Style would go back to the old formula OP uses which would be basically a downgrade again from 5-Style and it's commutatorish properties. Not to mention Parity is nicer to deal with (if 5-style would be adopted before full Parity)
 
Old Pochmann is a 2-cycle method, meaning you cycle 2 Pieces and solve one at a time - this only works if you also permute an even number of pieces of the other type (3BLD only for simplification, Corners and Edges)
Now 3-Style up's things by cycling 3 pieces and solving 2 (Orozko and Eka left out for simplification) at a time by using commutators. Now, 4-Style would go back to the old formula OP uses which would be basically a downgrade again from 5-Style and it's commutatorish properties. Not to mention Parity is nicer to deal with (if 5-style would be adopted before full Parity)
It’s less algs than 5style and is just as intuitive, it’s definitely more doable so I could see that becoming a thing before 5style. You could even end up using it as parity for some cases you would run into if you did learn 5style
 
It’s less algs than 5style and is just as intuitive, it’s definitely more doable so I could see that becoming a thing before 5style. You could even end up using it as parity for some cases you would run into if you did learn 5style

Less algs is true, I didn't think of that but please elaborate on how it's just as intuitive.
Using it as Parity would defenitely be possible, but we have to see if for the majority of cases doing a 4-Style alg is faster than doing one comm and a Parity algorithm
 
FInalllyyyyyyyyyyyyyy,

Sharing my 5-cycle sheet from both DF and UF buffer.

Let's keep the debate of doable or not doable aside and let's start contributing.

https://docs.google.com/spreadsheets/d/158F-jyu8ld8kbdD4I_Lqy4fsX5vvp_MQOq7Ofg8NdSI/edit?usp=sharing

Please do contribute algs, I am sharing the entire sheet, there is either DF 5-cycles or UF 5-cycles that you can find and contribute.

If contributing one alg at a time is cumbursome, join my discord server rouxnblind, and contribute there.
Your inputs are vital since this algset is huge.

The lettering scheme used here is not speffz so please keep that in mind yo!
 
Guys seriously, stop w/ all the ridicule. For all we know in 5 years this could be the next big thing in blidnsolving.
Way more than that because it takes 5 years just to learn all the algs

Let’s find something else that will revolutionize blind solving
 
Last edited:
Guys seriously, stop w/ all the ridicule. For all we know in 5 years this could be the next big thing in blidnsolving.
If someone posts an idea into a public forum, they should expect scrutiny and feedback. That is kind of the point, isn't it? Otherwise they could choose to keep their work private.

In this case, the idea of learning edge 5style by brute force and posting one alg a day on YouTube is, frankly, ridiculous. You may be right that 5cycles become a thing in blindsolving, but for that to happen, someone will have to find a more ingenious and realistic approach. @abunickabhi has been challenged to do that, but there's no sign of it yet.

(Finally, "ridicule" implies mocking and disrespect. While there may be some light-hearted mocking here, I don't see any serious disrespect and certainly don't intend any in my posts. @abunickabhi is an accomplished and knowledgeable blindsolver and I respect him for that; a respect which extends to giving him credit that he can accept criticism of his ideas.)
 
Yes agree, I have still not come up with an ingenious and realistic approach. But, by open sourcing, I am hoping some selfless contributors will help make this algset complete and then easier.

There are many challenges in the world that seem impossible (send a man to the moon, internet over the entire planet, curing cancer) but with a few years of invested time and collective effort, it becomes mainstream and widely adapted.
 
Back
Top