VDW, short for VanDerWyst, is a Rubik's cube speedsolving method proposed by Alex VanDerWyst in 2015. It could be described as a mix of Corners First and Roux and allows for high TPS while maintaining a low movecount and keeping blockbuilding to a minimum. In spite of its potential, it currently doesn't have any active users.
- Lines (sometimes also "Skis" or "Twins"): Two 1x1x3 blocks are built in DL and DR, respectively. Centers can be ignored here.
- CLL: The remaining four corners on the top are oriented and permuted in one of 42 algorithms while preserving the DL and DR edges. Most CLL algorithms for 2x2 apply here, although one might also use CMLL (or any other CxLL subset that doesn't destroy DL and DR).
- LR: The four edges in the equator edges are solved in pairs so that F2B is finished after this step. (Similar to 4b of Roux.)
- LSE: The last six edges on the M-slice and the U-layer are solved <MU>-gen, usually in multiple steps. Roux LSE is most common, although different approaches (listed in LSE) can be used here as well.
Alternatively, instead of both orienting and permuting, the corners can also only be oriented in the second step with OCLL. When this approach is used, CPLL needs to be applied after the LSE step to finish the cube. Although this requires less algorithms, it's objectively worse and should only be used as a beginner variant.
- Inspection: Due to the lines step, most of the solve can be planned in inspection. Therefore, it is sometimes even possible to plan up to CLL. This prevents pauses mid-solve due to recognition (which is considered one of the biggest problems in Roux).
- High TPS: Except for LR, every other step is either planned in inspection or algorithmic, which allows for very high TPS.
- Less blockbuilding than Roux: Compared to Roux, the VDW method needs comparable or even less blockbuilding at the beginning and none in the middle of the solve. Therefore, less blockbuilding skills and intuitivity are required. This also gives beginners an easier learning curve than Roux does.
- Algorithms: Only 42 algorithms need to be learned, which is the same as Roux and less than in other popular speedsolving methods like CFOP. Since only the DL and DR edges need to be preserved, most of the very well developed 2x2 CLL algorithms transfer here.
- Very ergonomic: The first step and CLL are mostly <RUF>-gen and LSE is completely <MU>-gen, so turn speed can be very high and there are no rotations.
- LR: LR can be harder to learn as it is very intuitive and edges can be almost everywhere. This causes worse lookahead and low TPS. It also doesn't have great ergonomics, so one needs to be able to switch between R, U, F, M and E moves.
- Slice moves: Because of LR and LSE, this method heavily relies on slice moves, especially M, which can be hard for beginners to get used to and also aren't very easy to execute on big cubes. Since LSE is performed at the end of the solve, there is a higher chance to get a DNF instead of a +2.
- Undeveloped: Although this method has been proposed multiple times, it hasn't been developed as much as other big speedsolving methods like CFOP, Roux and ZZ and thus has a lot less resources.
- Semi-algorithmic LR: The inventor, Alex VanDerWyst, planned to generate algorithms to make solving LR easier but still not fully algorithmic. However, none of them have actually been made yet.
- E2L: To improve LR, Eric Fattah suggested the use of LMCF's algorithmic E2L step to allow for better ergonomics and higher tps because of the pre-made algorithms, although this will highly increase the amount of algorithms that need to be learned. The difference from semi-algorithmic LR is that this is fully algorithmic.
Methods similar to this one have been proposed very often. Examples are TICT by Feliks Zemdegs which was proposed six years prior to VDW. The first two steps are the same, although TICT finishes with T perm instead of LR and LSE, so in this case, VDW can be considered as a different method.
In 2019, however, the Skis method was invented independently by WoowyBaby, followed by Zimlit's LCE in 2020 (although the last two steps of the latter, E2L and LSE, were very likely inspired by the Skis method because they aren't contained in the proposal). In 2020, both of these methods were found to be the same as VDW, so all three articles were merged into this one.
Of the three people mentioned here who invented these similar methods, only WoowyBaby is listed as a (co-)proposer here because he was the first one to see VDW as a method that can be used to achieve fast times instead of a worse version of Roux.