ZZSpike
From Speedsolving.com Wiki

ZZSpike is a method for solving the Megaminx using Edge Orientation like in ZZ, which results in better ergonomics and an easier last layer.
Steps
 Solve the First Two Layers with any method.
 Build a Bálint block and the Spike Edge, which connects the block to the U layer.
 Orient all remaining edges to make the rest of the solve 4gen without destroying the Spike Edge.
 Finish the Second Two Layers by only using four sides (the layer with the Spike Edge is not included).
 Solve the Last Layer in either three (OCLL + CPLL + EPLL) or two (OCLL + PLL) looks.
Edge Orientation
For EO recognition, the following rules apply: [1]
 The two "even" faces are opposite to a each other.
 The two "odd" faces are opposite to each other.
 Even faces are adjacent to odd faces and odd faces are adjacent to even faces.
For each center:
 If the edge has a matching or opposite sticker:
 Sticker connected to center: Good edge
 Sticker pointing away from center: Bad edge
 If the edge has an adjacent sticker:
 Sticker connected to center: Bad edge
 Sticker pointing away from center: Good edge
When the layer with the Spike Edge is turned, all good edges become bad ones and the bad ones become good ones. This means that if the solver wants to orient a bad edge, it is placed in the layer with Spike Edge, the layer is moved, the now good edge is replaced with another bad edge and the Spike Edge is moved back to its initial position.
Pros
 2LLL: The last layer can be solved in only two looks without any influencing from last slot and slightly less algorithms than for standard 3LLL.
 Ergonomics: Since after EO, only four faces need to be turned instead of five and almost no rotations are required, the solver can reach higher TPS than with other methods.
Cons
 Edge Orientation: Edge Orientation can take a while to get used to and even if mastered, the solver needs to spend time on recognition and execution while other pieces could have been solved in the mean time.
 Fixed Order: Because the solver usually only learns EO recognition for four specific sides, the F2L and the Bálint block always need to be built at the same position, which can sometimes be inefficient.
See also
External links
 Speedsolving.com: Initial Proposal
 Speedsolving.com: Method in its current form
 Speedsolving.com: Walkthrough solve and EO recognition
 YouTube: Walkthrough Solve Part 1
 YouTube: Walkthrough Solve Part 2
 YouTube: Three Sub1 speedsolves
 YouTube: 55.52s speedsolve