Big Cube BLD Discussion

[MatsBergsten]

I think you meant: What is the minimum number of centers you can solve on a 4x4 with optimal reorienting?

Yes I think so.

If you just want an answer, I'd say 4 or 5 (from experience combined with some thinking, but no proof)

 

[Zane_C]

I think you meant: What is the minimum number of centers you can solve on a 4x4 with optimal reorienting?

Yes I think so.

If you just want an answer, I'd say 4 or 5 (from experience combined with some thinking, but no proof)

Yeah that sounds reasonable, thanks.

 

[Sakarie]

I think you meant: What is the minimum number of centers you can solve on a 4x4 with optimal reorienting?

Yes I think so.

If you just want an answer, I'd say 4 or 5 (from experience combined with some thinking, but no proof)

Yeah that sounds reasonable, thanks.

I remember Istvan saying it was 5, don't know if he was correct thou.

 

[Slash]

Sorry if someone have asked it before me, but I'm too lazy to read back.
So, how do you do reorienting on 5x5? I mean, the parity cases. I know the M'EME' types and the EM2E'M2 types, and I know an alg for void cube which swaps 2 edges and four centers on M slice, but it messes up my centers (MU'MU'MUMUMU2MUMU)
It's enough if you give me a link of the earlier post(s) about it.
Thanx

 

[masterofthebass]

I don't think anyone reorients to anything other than a non parity case.

 

[Mike Hughey]

I don't think anyone reorients to anything other than a non parity case.

That's certainly true for me. To me the big breakthrough that led to my using reorienting on 5x5x5 was the realization that you could quickly and easily determine what the non-parity cases were, and then only consider those cases when reorienting.

 

[Slash]

I don't think anyone reorients to anything other than a non parity case.

That's certainly true for me. To me the big breakthrough that led to my using reorienting on 5x5x5 was the realization that you could quickly and easily determine what the non-parity cases were, and then only consider those cases when reorienting.

Thank you. I thought that too because you've always said that reorienting is very easy. Although I think I'll try to think about parity algs on sooo boring lessons in the next 1 week I still have to go to school

 

[Mike Hughey]

I have a question:

Does anyone have a good method (other than just being good enough to remember what you've used) to find the last cycles for wings where there are multiple cycles? I have a good counting system based on my locations that allows me to easily figure out whether or not I'm missing any pieces, but then I sometimes have a terrible time actually finding those cycles. This past week I had a 23 minute 5x5x5 BLD attempt (normal for me is about 15 minutes) and two 12 minute 4x4x4 BLD attempts (normal is about 7 minutes), all because I spent 5 or 6 minutes just memorizing the wings. I really need to find a way to fix this.

 

[trying-to-speedcube...]

This might be a bit far-fetched, but how about breaking down the 24 edges into for example 6 groups of 4 edges, and remembering which pieces in the group are have been memorized and which haven't with a hexadecimal system?

This probably would take a lot of time to master and would cut your usual time down, but it would also get rid of your very slow attempts.

 

[MatsBergsten]

I have a question:

Does anyone have a good method (other than just being good enough to remember what you've used) to find the last cycles for wings where there are multiple cycles? I have a good counting system based on my locations that allows me to easily figure out whether or not I'm missing any pieces, but then I sometimes have a terrible time actually finding those cycles. This past week I had a 23 minute 5x5x5 BLD attempt (normal for me is about 15 minutes) and two 12 minute 4x4x4 BLD attempts (normal is about 7 minutes), all because I spent 5 or 6 minutes just memorizing the wings. I really need to find a way to fix this.

Nah, that is a real hard one. When faced with this problem (not finding an unused edge) I scan through all edges visually (first upper layer, then middle and last down) and skipping those I have used already. Doing this I rely on a sort of combination of visual and "pure" memory (when this does not work I have to go through the memory cycle for each candidate and that takes time). So I'm also interested in a better "sieve".

 

[cmhardw]

I have a question:

Does anyone have a good method (other than just being good enough to remember what you've used) to find the last cycles for wings where there are multiple cycles?

I would also be interested in this. I personally have a hierarchy of wing pseudo-buffers. Whenever a cycle ends with my main wing buffer I go to a secondary pseudo buffer and memorize a new cycle. Daniel has been trying to convince me to break into cycles, but I have not yet tried it. Probably laziness on my part, as from hearing Daniel talk about it I think it could definitely be fast.

Anyway, when I complete a cycle, but by my count I know a small cycle is left, I visually trace the cycles from each of my pseudo buffers until I encounter "new stuff". So I guess like Mats said I just use pure memory or visual memory. Since it's very likely that a lot of my pseudo buffers are contained in one cycle, or at least other cycles, sometimes I can narrow out whole groups of them quickly. Either way, though, it usually involves me visually tracing cycles on the cube multiple times before I discover the "left over" cycle(s). Not exactly super efficient in my opinion.

Chris

 

[trying-to-speedcube...]

How about tapping the pieces you haven't memorized yet instead of the pieces you have memorized? That just takes a little time and makes it a lot easier to find the last cycles.

 

[MatsBergsten]

I personally have a hierarchy of wing pseudo-buffers. Whenever a cycle ends with my main wing buffer I go to a secondary pseudo buffer and memorize a new cycle.

Sorry Chris, but I probably don't understand what you mean by this. If you choose a new buffer you must have the exact problem of (quickly) deciding if that one has been used in earlier cycles?? How does a hierarchy of buffers help? Could you please explain a little more.

 

[blah]

How about going through every edge from A like normal people do? It really isn't that slow. In fact, it probably won't take more than a minute to find a wing that you haven't used...

Edit: Yes, especially good for you letter-pair people. You need to trace 3 wings max to know if you've already used it.

 

[cmhardw]

I personally have a hierarchy of wing pseudo-buffers. Whenever a cycle ends with my main wing buffer I go to a secondary pseudo buffer and memorize a new cycle.

Sorry Chris, but I probably don't understand what you mean by this. If you choose a new buffer you must have the exact problem of (quickly) deciding if that one has been used in earlier cycles?? How does a hierarchy of buffers help? Could you please explain a little more.

Hey Mats,

I guess I never thought about it really. When testing a new pseudo-buffer I visually trace the cycles through and see if what I'm tracing seems familiar at all. I sometimes think about the images that would come up when using those letters, and sometimes I just focus on what "area" of the cube am I tracing. For example, am I mostly around the UFR corner, or the DFL corner? Or am I mostly closer to the U layer, or D layer?

Again, I'm not arguing for my method. I think my method is inefficient, and I am also interested in new approaches for this. Chester, I will have to try the alphabetical approach, as that seems even more systematic than my approach.

Chris

 

[Mike Hughey]

Chester, it seems like your method and Chris's are very similar; it's just that you're more systematic about it than Chris is. I do something very similar to Chris, except that I actually break into new cycles instead of changing the buffer. I have standard pieces that I go to first when looking for a new cycle. The worst solves are ones where all my most common backup pieces get used and I'm casting around looking for more - I don't have a complete hierarchy that goes through all the pieces; instead I have a few that I always use.

I should probably really start using the alphabetical approach, at least for bad solves. It should at least prevent the 7 and 8 minute wing memo times.

 

[cmhardw]

Mike, slightly off topic, but I've been thinking about this a lot as well. How do you memorize the central edges and corners on a 5x5x5? Do you use the exact same method as what you use for 3x3x3? I recently read that you visualize your images for 3x3x3, but not within a journey location. I might have to try this as well, though I think it might feel a bit weird to not be using my journeys at first.

I have been using my single syllable words for both central edges and corners when doing a 5x5x5, but I end up with a phrase that is longer than the 10 second auditory memory loop - and this defeats the whole purpose of using single syllable words in the first place.

After reading your recent posts I've debated trying to visualize my edge images in black space, pretty much what I think you're doing for 3x3x3, and then continue to use single syllable words for the corners.

Does that sound like a viable approach to you, based on your experience? Of course you don't have to answer if your approach is secret, but I figured I would ask

Chris

 

[LarsN]

I've had a few DNF's or long solves because of trouble finding the last wings too. It hasn't happened a lot lately, but I think that's because it's faster to go over the cube visually and find wings that aren't memo'ed yet.

I've just started to use a special person/action memo for the first 8 wings, and I've experienced that it is harder for me to find the missing last cycle because I don't have the first 8 visually in my memo. But it helps my solves because wings are the first part I memorize and the last part I execute.

I don't have an idea for a good system that helps with this problem. Just wanted to share my experience

 

[Mike Hughey]

Mike, slightly off topic, but I've been thinking about this a lot as well. How do you memorize the central edges and corners on a 5x5x5? Do you use the exact same method as what you use for 3x3x3? I recently read that you visualize your images for 3x3x3, but not within a journey location. I might have to try this as well, though I think it might feel a bit weird to not be using my journeys at first.

I do corners the same on 5x5x5 as on 3x3x3 - auditory loop (like you), memorizing at the end and solving at the beginning. For central edges, I put them in long-term memory locations, same as with wings and centers. I memorize the central edges first and solve them last, so they're actually my longest-term memory of all. I even do this on 7x7x7, so it's often 40 minutes to an hour between memorizing and solving them!

I have been using my single syllable words for both central edges and corners when doing a 5x5x5, but I end up with a phrase that is longer than the 10 second auditory memory loop - and this defeats the whole purpose of using single syllable words in the first place.

It's funny how I just slowly moved into an auditory solving method for corners. I didn't really set out to do it, but it just sort of happened. I have still never set up a single syllable word list, but I use abbreviated versions of most of my long images, so it works almost as well; probably half of my images are represented by single syllables when doing corners. I should probably stop being lazy and memorize a proper list - it would probably save a second or two off the average solve for corners, which would be significant for 3x3x3.

After reading your recent posts I've debated trying to visualize my edge images in black space, pretty much what I think you're doing for 3x3x3, and then continue to use single syllable words for the corners. Does that sound like a viable approach to you, based on your experience?

So as you can probably tell by now, that's pretty much exactly what I do, except that I don't have a properly-prepared single syllable word list. So yeah, I think it's viable. For me, I make a single story for the edge images, all run together. The biggest problem for me is that sometimes that's pretty long, but I find that it cuts down so much on memory recall delays during the solve that it's worth doing. This was one of my biggest breakthroughs in multiBLD; I store a full story of all the edge images in a single location when doing multi (but at a proper location in multi, whereas it's with no location in singles).

Of course you don't have to answer if your approach is secret, but I figured I would ask

Don't be silly - I don't keep method secrets. I like talking about the methods far too much to ever do that.

 

[cmhardw]

I memorize the central edges first and solve them last, so they're actually my longest-term memory of all. I even do this on 7x7x7, so it's often 40 minutes to an hour between memorizing and solving them!

Wow, I didn't realize that you memo'd and solved in that order for central edges! Are you using an ABCD memo and DCBA solving, palindromic I guess?

I should probably stop being lazy and memorize a proper [single syllable word] list - it would probably save a second or two off the average solve for corners, which would be significant for 3x3x3.

If you do, avoid the mistake I did. Don't memorize a single syllable word list, but rather memorize two lists. For each letter pair, have an object and a verb. This way you can combine the phrases as object-verb-object-verb. I wish I had not made my list a hodge-podge of objects and verbs, and converting my single syllable word list to an object-verb list is my biggest future blindfold cubing goal. Sadly I am not very well on track for this though *cough*laziness*cough* :-s

So as you can probably tell by now, that's pretty much exactly what I do, except that I don't have a properly-prepared single syllable word list. So yeah, I think it's viable. For me, I make a single story for the edge images, all run together. The biggest problem for me is that sometimes that's pretty long, but I find that it cuts down so much on memory recall delays during the solve that it's worth doing. This was one of my biggest breakthroughs in multiBLD; I store a full story of all the edge images in a single location when doing multi (but at a proper location in multi, whereas it's with no location in singles).

That's reassuring, I might have to try that as well. I'll probably practice it on 3x3x3 until I get a bit more used to it, then transition it to my 5x5x5 solving as well.

Don't be silly - I don't keep method secrets. I like talking about the methods far too much to ever do that.

Good to know, and yes I also share the same thought. I would much rather debate methods than keep them secret. The method I used to use before Daniel and I began discussing and sharing ideas on the way to BH was a lot less efficient than what I use now. Without collaborating with Daniel I would certainly be slower than I am now. Also, Mike I consider you the King of BLD, as your craziness and dedication to it I think are not closely matched by very many people. So, thanks for the advice, as I think you have more actual experience with BLD solving than most everyone, including me.

Chris