Cubing Predictions Thread

[One Wheel]

I've been enjoying the discussion on the "World Record Predictions in the Year 2020?" thread, but I have ideas for predictions that are not world records. I can't find another thread like this, so here we go. Put dates on, or not. Support your predictions with math, hunches, astrology, or whatever you like. Here are a couple of mine:

By the end of 2018:
One or more of the big three manufacturers will offer a mass-produced line of premium magnetic cubes, much like what The Cubicle is doing now on a custom order basis. Price points will be approximately:
3x3: $32
4x4: $49
5x5: $56
6x6: $69
7x7: $74
Megaminx: $57
All world records except multi-blind and FMC will be set on magnetic cubes. Clock will no longer be an official event, most likely replaced by kilominx.

By the end of 2020:
The big three of Moyu, Qiyi, and Yuxin for non-3x3 hardware will be joined by either Seven Towns or V-cube. Whichever of those two is not a legitimate competitor in the speedcube market will be out of business. It looks like Seven Towns lost their trademark fight yesterday: adapt or die.

By the end of 2021:
Somebody will set a sub-0.6 2x2 WR average with four 4-move solutions in the same average.

 

[Rcuber123]

By the end of 2021:
Somebody will set a sub-0.6 2x2 WR average with four 4-move solutions in the same average.

Ummmm. are u sure?

 

[One Wheel]

Ummmm. are u sure?

No, it's a guess. I do believe that it is possible, and if it happens you can say you heard it here first .

 

[Rcuber123]

No, it's a guess. I do believe that it is possible, and if it happens you can say you heard it here first .

meant are u sure ur sober.

anyway this will be amazing if u predicted this correctly.

 

[Daniel Lin]

By the end of 2021:
Somebody will set a sub-0.6 2x2 WR average with four 4-move solutions in the same average.

probability of getting 4 four movers in an average=
5*(1847^4*3672316/3674160^5)=
3.19144458e-13

about 1 millionth the prob of dying from a measles vaccine

 

[One Wheel]

meant are u sure ur sober.

Well, I'm really not in the habit of doing drugs or alcohol, and I haven't done anything out of the ordinary today. I realize it seems like a crazy prediction, but what's the fun of any other kind?

probability of getting 4 four movers in an average=
4*(1847^4*3672316/3674160^5)=
2.55315566e-13

about 1 millionth the prob of dying from a measles vaccine

So you're saying it's possible? does your calculation take into account the fact that the non - 4 - move solution can occur at any solve in the average?

 

[Daniel Lin]

So you're saying it's possible? does your calculation take into account the fact that the non - 4 - move solution can occur at any solve in the average?

yeah it was wrong i fixed it

It's possible, but not very likely. It's also possible to get a 0.6 average for 7x7

 

[kamilprzyb]

I guess that at some point Gianfranco will get very lucky scramble and will set low 17 as WR in BLD, that will be almost impossible to beat

 

[Daniel Lin]

I guess that at some point Gianfranco will get very lucky scramble and will set low 17 as WR in BLD, that will be almost impossible to beat

3BLD IMO has a lot of room for improvement. Just look at how crazy memory championships are

I predict sub 15

 

[WACWCA]

No, it's a guess. I do believe that it is possible, and if it happens you can say you heard it here first .

There hasn't even been a counting sub 1 in comp though

 

[kamilprzyb]

3BLD IMO has a lot of room for improvement. Just look at how crazy memory championships are

I predict sub 15

Quoting Maskow from reddit:

I think that we are very close to that limit now. Why do I think that? I think that because it's impossible to improve our memorisation time right now and we have completely no idea how to solve BLD faster than with commutators. Yes, those comms can be upgraded a little, because there are a lot of them and nobody is able to do them all in most optimal way, you can also learn all the possible parity cases (396 of them), but it will be end. Over. What else can be done? We don't know, and I hardly believe that it's even possible to learn anything more to be much faster. So probably something like ~22,5 seconds long term average is a cap when you do everything perfect. I really want to see that I am wrong but it hard to believe for me right now.

We will see..

 

[One Wheel]

3BLD IMO has a lot of room for improvement. Just look at how crazy memory championships are

I predict sub 15

I suspect that we are getting close to the limit of "normal" human capability, but somebody with a photographic memory, quick fingers, and "Rainman" type abilities could probably do a ~10-11 second CFOP, ZZ-CT, or whatever other method blind speed solve with 1.5-2 seconds of memo time. Possibly less if they could look at 5 sides of the cube and solve based on that.

 

[efattah]

2x2 has huge room for improvement. Compared to ZBLL, 1LLL, and even OLLCP, the 2x2 algorithm set is tiny (126). If you add TCLL, TEG1, TEG2, it becomes larger, but some day an ambitious and dextrous person will memorize 2000+ algorithms on the 2x2. Fetching the algorithms from memory will not be a problem because it will happen during the inspection time. With a much larger number of algorithms, the average move count to solve the 2x2 would drop by about a third, and the times along with it.

3BLD has probably the biggest room for improvement. It would absolutely be possible to memorize the 3x3 in 1.5 seconds, close your eyes, and solve at full 3x3 speed. This would mean that 3.85 + 1.5 = 5.35 seconds would be possible. Such a person would simply visualize the cube turning at the same time they are turning it and 'see' the cube in their mind. A feat such as this is not only possible, but on a similar level to the handful of individuals who have solved the 4-dimensional megaminx without macros, a feat which seems absolutely impossible to most people.

 

[DGCubes]

Prediction: By this time tomorrow this thread will have at least five more posts.

 

[One Wheel]

2x2 has huge room for improvement. Compared to ZBLL, 1LLL, and even OLLCP, the 2x2 algorithm set is tiny (126). If you add TCLL, TEG1, TEG2, it becomes larger, but some day an ambitious and dextrous person will memorize 2000+ algorithms on the 2x2. Fetching the algorithms from memory will not be a problem because it will happen during the inspection time. With a much larger number of algorithms, the average move count to solve the 2x2 would drop by about a third, and the times along with it.

Once people start learning all optimal solutions for 2x2, at what point does it become a trivial test of dexterity and is eliminated, like magic? Here's a bold prediction I'll make: 2x2 will be deemed trivial and eliminated by 2100, in the off chance WCA is even around by then.

 

[turtwig]

3BLD has probably the biggest room for improvement. It would absolutely be possible to memorize the 3x3 in 1.5 seconds, close your eyes, and solve at full 3x3 speed. This would mean that 3.85 + 1.5 = 5.35 seconds would be possible. Such a person would simply visualize the cube turning at the same time they are turning it and 'see' the cube in their mind. A feat such as this is not only possible, but on a similar level to the handful of individuals who have solved the 4-dimensional megaminx without macros, a feat which seems absolutely impossible to most people.

I'm pretty sure it takes longer than 1.5 seconds for world-class 2x2ers to one-look a 2x2 so I don't think 3BLD will ever get that fast. I feel like sub-15 global average is the absolute limit for 3BLD.

 

[One Wheel]

I'm pretty sure it takes longer than 1.5 seconds for world-class 2x2ers to one-look a 2x2 so I don't think 3BLD will ever get that fast. I feel like sub-15 global average is the absolute limit for 3BLD.

I'm not anywhere close to even being able to one-look a 2x2 at all, but I believe that the 1.5 second figure would include figuring out the solution. 3BLD memo in that time would not: it would just be visual memo and then visualizing a speedsolve as it's performed, the blindfold at that level being merely incidental.

 

[WACWCA]

2x2 has huge room for improvement. Compared to ZBLL, 1LLL, and even OLLCP, the 2x2 algorithm set is tiny (126). If you add TCLL, TEG1, TEG2, it becomes larger, but some day an ambitious and dextrous person will memorize 2000+ algorithms on the 2x2. Fetching the algorithms from memory will not be a problem because it will happen during the inspection time. With a much larger number of algorithms, the average move count to solve the 2x2 would drop by about a third, and the times along with it.

Interesting idea, but the move count would definitely not drop as much as your saying. I know up through TCLL+ and depending on what type of subset I go for on a solve, it maybe saves 2-3 moves than the others. Algs need tps rather than move count too. And to one look more then 2-3 sides in an inspection is unreasonable
also, I bet TEG-1 and 2 algswould suck, half the algorithms already don't make any sense with TCLL. Also it is pretty Rare that one of those would be the best to start with, considering how many other options you have with better algs. I think that the sets we have now are about all that's Necessary, but maybe a few more. I've always thought that someone would learn how to do pseudo EG eventually. And the reason that the sets like ZBLL you mentioned are bigger is because they have edges.
The official average I think can be brought to the 1.2x in comp, maybe 1.1 eventually, but I think 1.6 avg at home is about the limit

I'm pretty sure it takes longer than 1.5 seconds for world-class 2x2ers to one-look a 2x2 so I don't think 3BLD will ever get that fast. I feel like sub-15 global average is the absolute limit for 3BLD.

How on earth is this reasonable to recognize that fast? You have to track the cycles around the cube for both edges and corners? Even if you could picture the entire cube in your head at that point, you would have no idea how to solve it without tracking it. World class people still need good scrambles for sub 10 memo

 

[One Wheel]

How on earth is this reasonable to recognize that fast? You have to track the cycles around the cube for both edges and corners? Even if you could picture the entire cube in your head at that point, you would have no idea how to solve it without tracking it. World class people still need good scrambles for sub 10 memo

The point is that for a very few exceptionally talented people it should be possible to track a 3x3.

 

[sqAree]

How on earth is this reasonable to recognize that fast? You have to track the cycles around the cube for both edges and corners? Even if you could picture the entire cube in your head at that point, you would have no idea how to solve it without tracking it. World class people still need good scrambles for sub 10 memo

I think Maskow's global average for memo was ~7 seconds. But I'm doubtful as well.