From Speedsolving.com Wiki
Proposer(s): Joseph Briggs, Eric Fattah, Julien Adam and others
Proposed: 2017 (or possibly earlier)
Alt Names: Last Seven Edges
Variants: ELL, L5E, LSE
Subgroup: unknown
No. Algs: 149 (Advanced WaterRoux L7E)
Avg Moves: 17.25 (Advanced WaterRoux L7E)
Previous state: 7 Edges missing cube state
Next state: Solved cube state

7 Edges missing cube state -> L7E step -> Solved cube state

The L7E step is the step between the 7 Edges missing cube state and the Solved cube state.

L7E, also called Last Seven Edges can be used as a last step to solve the remaining seven edges in 42, WaterRoux, Waterman, LMCF and other Corners First methods.

Possible approaches

There are multiple ways to solve the last seven edges, some of which are listed here.

R edge+LSE

  1. Solve one edge in the R layer
  2. Finish the solve with any LSE variant

While this approach is very easy for people coming from Roux, other variants are more efficient.


  1. Solve EO and the FR edge
  2. Finish the solve using L6EP to permute the remaining six edges (e.g. with Roux's 4b and 4c)

This was proposed by Joseph Briggs for his 42 method.

2opp EO

  1. Solve two opposite edges (UL+UR, UF+UB or perhaps even DF+DB) and EO simultaneously
  2. Permute the remaining five edges using L5EP

This was proposed by Joseph Briggs for his 42 method.

Old WaterRoux L7E

  1. Do 0-5 setup moves and then execute an algorithm to solve UL, UR, FR and EO
  2. Permute the four midges (edges in M) with Roux's 4c

This was the first idea for WaterRoux L7E by Eric Fattah with a movecount from around 15 to 19, although it is not recommended for use anymore.

WaterRoux L7E

  1. Orient two edges and position them at UL and UR whilst bringing either the FR edge or DR edge to the D-layer. Centers must be solved or off by an M2 in [1]
  2. Using one OL5E algorithm, the remaining edges are oriented [2]
  3. Solve the remaining six edges with special L6EP algorithms [3]

This approach was invented by Julien Adam for the WaterRoux method and is currently the recommended approach. In its advanced form, it averages 17.25 moves by utilizing 149 algorithms. [4]


For one-handed solving, Joseph Briggs proposed that l and l' moves can be used during table abuse to "switch between styles" in order to access all two edges on the R layer. A better explanation is given in his video.

External links