How does kuit pair 5X5 edges?

[Mike Hughey]

Guys... please get some sources or actual knowledge before claiming stuff.
Matyas does NOT use AVG nor AVG without buffer. He does do 2 at a time mostly but more freestyle and he tends to complete a semi-edge and creating another one rather than just making semi-edges all the time. He showed me in Sweden in 2007.

That's why I asked Rama - I figured he would know. Thank you for clearing it up - you answered my question, so now I hopefully know for sure. (Of course, I got in trouble before because I trusted what I thought was a good source on Haiyan Zhuang's BLD method, and it was apparently wrong. I still feel bad about that.)

Now how about you Erik, do you use AVG for 6x6x6 and 7x7x7? Obviously you're the best person to ask about that.

 

[Erik]

Yes I do but in my opinion the bigger the cubes get the worse AvG is and the better Bigcubes pairing becomes. Still I have to do AvG since I suck even worse at Bigcubes pairing. I'm barely sub-5 on 7x7 with it and like 3:30 on 6x6 compared to 4:45 with AvG on 7 and 3:05 on 6

 

[Mike Hughey]

Yes I do but in my opinion the bigger the cubes get the worse AvG is and the better Bigcubes pairing becomes. Still I have to do AvG since I suck even worse at Bigcubes pairing. I'm barely sub-5 on 7x7 with it and like 3:30 on 6x6 compared to 4:45 with AvG on 7 and 3:05 on 6

Thanks for that - now I know for sure.

I've been trying to get better at Bigcubes pairing for 6x6x6 and 7x7x7, but I'm still about a minute slower with Bigcubes pairing than with AvG. But it is getting better.

 

[AvGalen]

Yes I do but in my opinion the bigger the cubes get the worse AvG is and the better Bigcubes pairing becomes. Still I have to do AvG since I suck even worse at Bigcubes pairing. I'm barely sub-5 on 7x7 with it and like 3:30 on 6x6 compared to 4:45 with AvG on 7 and 3:05 on 6

Thanks for that - now I know for sure.

I've been trying to get better at Bigcubes pairing for 6x6x6 and 7x7x7, but I'm still about a minute slower with Bigcubes pairing than with AvG. But it is getting better.

I have discussed this a bit with Erik, but I haven't really evaluated this. I would like to know how the bigcubes methods scales when puzzles get bigger. I cannot even do it on a 5x5x5 so I cannot so for sure if it would be better than AVG on > 5x5x5. For AVG scaling is simply a matter of this formula (indication only): (number of edges to pair / 2) * 7.
So for a 5x5x5 it would be (24/2)*7 = 84, for 6x6x6 it would be 126 and for 7x7x7 168.
Untill this weekend I assumed bigcubes would be better because you can take better advantage of lucky cases, but I am really not sure anymore.

 

[Mike Hughey]

Untill this weekend I assumed bigcubes would be better because you can take better advantage of lucky cases, but I am really not sure anymore.

That's a very mysterious-sounding comment - what happened this weekend that made you not sure anymore?

I also figured bigcubes would allow you to take better advantage of lucky cases. If the methods are equal, in the situation on a 7x7x7 where all 4 pieces are in the freeslice (and even better when they're in the proper direction on the freeslice), you almost get an entire free edge. That may not happen often, but it seems like even 3 of 4 would make a huge difference. Does it not work out that way somehow?

To me, it seems like it's only situations like this that give bigcubes an advantage. I really don't see any reason for it to be any better on 5x5x5. Although I know Dan says bigcubes is better than AvG for lookahead, to be totally contrary to your viewpoint.

 

[blah]

Although I know Dan says bigcubes is better than AvG for lookahead, to be totally contrary to your viewpoint.

I've asked Dan about it. And I think it's a different sort of lookahead altogether (if there were lookaheads of different sorts in the first place).

IMHO, AvG is better in terms of "short-term" lookahead whereas bigcubes.com is better in terms of "long-term" lookahead. With AvG, you always know what your immediate next move is, that's what I call short-term lookahead. With bigcubes.com, you trace and remember semi-pairs that are already formed so that you can use them to your advantage later (not immediately), that's long-term lookahead. Correct me if I'm wrong.

 

[AvGalen]

@Mike: What changed is my incredible improvement on 7x7x7 last weekend where I did a 1:30 pairing

@Mike+blah: Dan just really dislikes AVG and is amazing at bigcubes-method so he isn't the best one to ask. I really like AVG and only know the concept of bigcubes (haven't ever done it) so I am not the best one to ask as well.
For AVG you have to find 1 wing at a time. After you found it you immediately know what the next wing will be and you can start looking for it while solving the previous wing. That is why I think lookahead is so good for AVG. Also, you only move a couple of pieces so keeping track of previously semi-solved pieces will be easy
For bigcubes you need to find 3 pieces and while solving them you have to find 3 more pieces at the same time. That means lookahead isn't as good, BUT because you can store/remember already formed tredges at fixed locations lookahead does become increasingly easy.

 

[Mike Hughey]

@Mike: What changed is my incredible improvement on 7x7x7 last weekend where I did a 1:30 pairing

Ah, so that's it. Now I understand. Anyway, I can't see where AvG is more than just a little less inefficient that Bigcubes on even the larger cubes - it's only the lucky cases that will help you. I think quite a few can often come up, but it's still not enough that I can believe it affects the solve by very many extra moves.

Anyway, Arnaud, you should really learn the Bigcubes pairing method. I'm sure you'd find it fun to play with. You can just do the "hybrid" method, pairing the first 8 and then switching to AvG for the rest; that way you can do it without learning any new algorithms. I suspect you'd really enjoy it for an occasional change.

I don't remember all the other algorithms anymore, but I still occasionally do the Bigcubes/AvG hybrid for fun on bigger cubes. I just did a 7x7x7 average of 5 with Bigcubes/AvG pairing this morning, and it was (barely) under 10 minutes. I keep hoping I'll eventually get close to my AvG speed, although I doubt I'll ever get faster.