The Funniest Rubik's Cube Record In History

Some epic unintentional trolling

There’s a new speedcubing single solve record which is by far the funniest one ever. No, it isn’t because it’s by an 11 year old kid. He’s the best speedcuber in the world. It isn’t even because he fumbles the solve at the end at the end and still shatters the world record, on a solve which isn’t even all that lucky, although that is very funny. What really makes it funny is that there are multiple videos breaking down the whole solve, by people who are themselves good speedsolvers, going over every single turn done in painstaking detail, which completely miss a key part of what’s happening.

The background to this is that there have long been pissing matches within the world of speedcubing about what’s the best solving algorithm. On one end there are methods which minimize total number of moves, with Petrus being particularly good at keeping that number down, and at the other end are methods which optimize for turning as fast as possible, with CFOP being an extreme of that. The problem with the turn optimizing algorithms is that they require a lot of analysis while the solve is being done and the finger placements to do the turns required tend to be awkward, in the end rendering the whole thing very slow. The speed oriented methods by contrast require only a few looks at the cube and then doing memorized sequences between them. (Speedcubers call sequences ‘algorithms’ which makes me cringe but I’ll go with it.) Good speedcubers have not only memorized and practiced all the algorithms but worked out exact finger placements throughout them for optimal speed. This is the approach which has worked much better in practice. CFOP has dominated speedsolving for the last two decades.

What seems to be happening to the speedsolvers who aren’t initiated is that this is a fairly lucky CFOP solve which happens to yield particularly good finger placements. While this is technically true it’s missing that those setups aren’t accidents. While this solve is a bit lucky, several of those things are guaranteed to happen because of some subtle things done in the solve. This isn’t actually a CFOP solve at all. It’s EOCross.

A quick summary of CFOP: First you solve the bottom four edges, which is done intuitively (really planned out during the inspection time provided before solving begins). Then a ‘look’ is done and one of the bottom corners and the edge next to it is solved. This is repeated three more times to finish the bottom two layers. Then a look is done and an algorithm is used to orient all the top layer pieces. Finally there’s one last look and an algorithm is done to position all the last layer pieces. Yes that’s a lot of memorized algorithms.

This is how EOCross works: First you solve the bottom edges and all edge orientations, a process which absolutely must be planned out during inspection. The meaning of ‘edge orientations’ in this context may sounds a bit mysterious and it’s subtle enough that it caused the accidental trolling of the new world record. If you only rotate the up, down, right, and left faces of a Rubik’s Cube the edges don’t change orientation. Literally they do change orientation in the sense that they rotate while moved, but whenever they go back to the position they started they’ll always be in the same orientation they were at the beginning. The solve then proceeds with doing corner and edge pairs from the first two layers but with the restriction that the front and back faces aren’t turned. Finally all that’s left is the last layer which happens to be guaranteed to have all the edges correctly oriented, and a single algorithm is done for those. Yes that’s an even larger number of algorithms to memorize.

That may have been a bit much to follow, but the punch line is that to an uninitiated CFOP solver an EOCross solve looks like a CFOP solve where the edge orientations happen to land nicely.

Technically EOCross is a variant on a solving method called ZZ but it’s sufficiently different that it should be considered a different method. It was invented several years ago and devotees have been optimizing the details ever since. There have of course been claims that it should beat CFOP, to which the response has mostly been to point out that no EOCross solvers have been anywhere near the best speedcubers. The rebuttal has been that the top solvers haven’t tried it because it’s so much work to learn and if they did they’d be faster. Given how handily the world record was just broken that rebuttal seems to have been correct. Good work EOCrossians getting the method optimized.