Choosing a Speedcubing Method
This guide will continue to be updated with more methods and in future hopefully link through to pages with more detailed explanations and analysis of the methods, if you have any questions or suggestions message Alpha Cuber (me) on our discord server. You can also click on the method name link to show some good resources for the method.
Due to the Immense complexity of the cube there are a large number of different methods that can be used to solve it. Obviously not all of these are good and not all are achievable by humans, but there are a variety of methods used by top cubers to solve the cube and it can be hard to decide which one to use. That is what this guide is for. I highly recommend you at least try CFOP, ZZ and Roux as they are all good methods, which have been proven at the top level of Speedcubing and will suit different people. Although this guide is long I think the help it will give you to choose a method is vital for any cuber looking to move on from the beginners method.
The big 4:
CFOP (also known as Fridrich)
CFOP is the most popular 3x3 speedsolving method and was used to set the current 3x3 and one handed WR singles and averages. It has also been the dominant method in Speedcubing for the last 15 years. It is more algorithm based than most methods and has less intuitive elements than other methods, which allow the solver to turn extremely fast - the worlds best can have an average turning speed in a solve of over 10 Turns Per Second (TPS). It has been developed far more than any other method, so there are lots of resources for learning techniques and lots of techniques for you to learn. This means that there is a clear improvement path and the structured process of the method leads to the possibility of fast improvement to a very high level, as it is not necessary for the solver to develop new techniques as pretty much everything needed to be world class can be learned from tutorials online. The problem is that cfop is less efficient than other methods (average move count of about 55 for top solvers) and this means you have to turn considerably faster to get the same times. You also have to learn more algorithms and techniques to try and reduce the move count and make it more competitive with other methods (the base method uses 78 algorithms and top solvers often learn hundreds more) and the extra moves along with more rotations than any other method result in it being worse for OH. On the other hand it has the best big cube variation (Yau) and currently all top solvers for big cubes use CFOP.
Roux is the second most popular method and is widely accepted to be the best for One handed solving. It is very efficient (average move count of ~45), is rotationless and primarily uses RrUM after the first step which leads to very ergonomic solves. On the other hand it is block building based and is highly intuitive which means you have to think more during solves, and this makes it harder to turn fast. It finishes the solve with MU moves which can be done very fast and you only need to learn 42 algorithms to be fast (much less than the other methods in the big 4) with the method but this causes the problem that there is not always a clear path for improvement. Last year (2019) Sean Patrick Villanueva used Roux to become the 3rd ever person to get a sub 6 official average in 3x3 and he also managed to place 2nd at the world championships, despite immense pressure and it being his first ever major championship. It has also been very effective at the top levels of OH with Kian Mansour, Iuri Grangeiro and Tudor Lin all using it to get top in the top 10 of the world for OH average and Kian Mansour is currently the only person ever to average under 9 seconds at home in OH. The largest problem with Roux currently is that there is no good big cube variation with the most popular one (Meyer) being extremely unproven at the top levels of Speedcubing and is generally considered worse than the CFOP equivalent (Yau).
The ZZ method has many different variants but they all start with a step called edge orientation (EO), this allows the cube to be solved rotationless with only RULD moves and most steps have less cases which allows for a very fast and efficient finish to the solve if you’re willing to learn lots of last layer algorithms. There are 2 main ways of starting the solve after doing EO: EOline and EOcross. The creator of the method (Zbigniew Zborowski) originally proposed eoline, where you solve the DF and DB edges during/after EO and then use blockbuilding to create 2 1x2x3s to solve the F2L. This is pretty efficient with an average movecount of 45 if used with ZBLL (generally considered the best way to finish a ZZ solve) but lookahead and ergonomics are considered worse for 2 handed solving so it is generally only used by ZZ solvers for OH. The other way people start is by solving an EOcross which involves solving all the D layer edges while doing EO. Although less efficient (52 move average with ZBLL) it is much more ergonomic and lookahead is easier which allows for higher TPS and normally leads to a better solve. The key to being good at ZZ is being able to plan a fast and efficient EOcross in inspection, as the rest of the solve is very fast and fluid. Due to the fact EO is solved there are lots of good ways to solve the last slot and last layer, of these variants ZZ-a is currently considered best. In ZZ-a you solve your last pair as normal and then do 1 of 493 ZBLL algs to solve the rest of the cube. Other good variants include ZZ-C++ and ZZ-d with CPLS. For people who don’t want to learn lots of algs OCLL+PLL can be used to get very fast with only 28 algs, but the majority of ZZ solvers only see this as a stepping stone for people who haven’t yet learnt any other variants. For a full list of variants click here. ZZ is better than CFOP for OH because of its lower movecount and rotationless solving but still worse than Roux because the ergonomics are worse and the movecount is higher. The best big cube variant (4z4) is considered worse than the equivalent for CFOP (Yau) but it is definitely still possible to be world class in 6x6 and 7x7 using reduction with ZZ. If you're interested in a rotationless method with lots of RU and LU turning and don’t mind learning algs, this method is the best for you.
Petrus is a blockbuilding method that prioritizes efficiency and when done well can get extremely good times. Due to the unusual style of solving and lack of resources the method lost popularity and got to the point where basically no one used it. However in the last year or so it started to gain popularity and more people are starting to use it. The problem with the considerable lack of use is there are very few resources and it is unproven at a high level, so we don’t know how fast you can get or how to get to that level. This means that if you want to pursue Petrus you have to do lots of experimenting and learning stuff yourself. Having said this if you want a unique solving style and enjoy an efficient blockbuilding method, you will have great success with Petrus. Petrus usually has an average movecount of about 50 when speedsolving but some users can get this down to low 40s with lots of blockbuilding practice and ZBLL. Because Petrus does EO mid solve you can use any of the ZZ last slot last layer methods, although most people try to learn ZBLL. This means that it is more efficient than most other speedsolving methods so a lower TPS is needed. Having said this the last two steps are a 2 gen right block and ZBLL where you can get very high TPS. This allows the solver to focus on efficiency in the first 3 steps. Petrus is pretty good for OH due to the efficiency and fast finish, although the poor ergonomics in the first 3 steps mean it is not as good as roux. There is no viable big cube method for petrus currently, but it can be used with reduction to get fairly good times on 5x5, 6x6 and 7x7 but not as good as CFOP and it’s equivalent big cube methods.