Welcome to my tutorial for my big cube method, otherwise known as sandwich (or r4)
sandwich is my variation of the cage method. The first cage method I learnt was from here:
I have modeled my variation after this very method by Denny Dedmore.
A rough overview of my method (or variation):

1. First 2 centers

2. Corners (with ortega)

3. Finish R and L faces

4. Middle edges in 2 phases.

5. Final Centers

I do not claim authorship of this method; I believe that many have come up with a similar variations. This method is no more than an optimization of Denny's solution, a hybrid method with a touch of everything else. By naming my variation "r4", I am paying tribute to a method that I admire, otherwise known as k4, created by Thom Barlow. Someone had suggested me the name sandwich, and I liked the name so much it shall be the official name of the method =P

Many thanks to Frederick Badie for his OLL and double parity algorithms, Lucas Garron for his 2-cycle parity algorithm, Kenneth for the ELL3 algorithms of his KBCM and other cage-friendly algorithms he has provided me with, and also Marc Waterman and Francois Courtes for some of their algorithms. Credits go to Per Kristen Fredlund for his guidance as well, and also Mike Bennett for inspiring me through that one tiny post on speedsolving that ultimately led me to change my middle edges method.

Last but not least, I would like to thank Lars Vandenbergh for his wonderful Imagecube which I used to generate the 4x4x4 images, Joel van Noort for his NxN imagecube which I used to generate the 5x5x5 images, and Werner Randelshofer for his wonderful cube applets.

For the generation of algorithms, I used Ron's Cube Solver, optimal444, and ACube by Ron van Bruchem, Clement Gallet, and Josef Jelinek respectively; I would like to express my gratitude to them as well.

History of this variation

I first started 4x4ing by learning Dedmore's solution sometime in July 2007. 2 weeks after that I adapted the solution to the 5x5.

After more experiences with the 4x4, I realized that I do not need to put each edge cubie individually for the first 7 dedges, as it is really really slow, so over a long time I found all the shortest solutions I could find for solving a dedge at a time. After much practice, I had my first sub 2 (1:26.xx) in January 2008, and then a sub 2 average followed.

By May 2008 I found it hard for myself to improve; my solves were inconsistent, and times ranged from 1:30s to high 1:50s, or even over 2 minutes. My real average was around 1:50.

I posted a thread on, and Mike Bennett replied about his Roux by 4 method. I adopted his idea for the middle edges, learned most of the algs (about 11-13), and found that my timings cut by at least 10s after a day. My first average was only barely sub 2 with it, but by the end of the day I was at the 1:40s already. More practice brought me down to sub 1:30, with FU's (Hoe Tze Han) advice on slowing down and lookahead...without lookahead this method would be a complete fail.

Before the Singapore Open 2008, I decided to optimize the way I solve the last dedge of the R and L faces, and wrote out all the algorithms I can find in about an hour or less. =D Before that it was very primitive, so using this I cut about 2-4 seconds depending on cases. I still use the very same algorithms to this day. They are probably not optimal, but at least they are fingertrickable and they are quite fast anyway.

My rubik's died on me after the Singapore Open, and I gave up 4x4 solving (for like 3 months) until the new Meffert's brand 4x4 came to town! :D This was around the time I realized that centre control with solving the last 2 dedges can be done. Before that I was using mainly Kenneth's ELL3 algs from his Kenneth's Big Cube Method, which do not affect centres at all. I learned all the relevant algorithms of the Waterman Big Cube Method and also some of the algs Kenneth gave me through PM on speedsolving. This doubled the number of algorithms I know for the last 2 dedges =D.

My times are around 57s +/- 3 (with warm-up, duh. My best average, 10 out of 12 solves, was 52.69 though.), and my non-lucky record (full-step) is 42.83s with a little luck, and my next best time which isn't really lucky is 47.80s. I am still optimizing the system; and will continue to update this tutorial when I have new inspiration.

Have fun with my tutorial! email me at if you encounter any problems, or have any questions on my system. If I have time I will post a small tutorial on a system that only really needs you to memorive 5 algs (excluding intuitive ones which do not need memorization) to solve the 4x4..basically a primitive form of this system.

Some notes of my notation: small letters denote slice moves, big letters denote face moves, and moves within a parenthesis denote moves that can be done together in one turn.

I use a black body tiled Meffert's 4x4x4.

This website works best with Opera and Mozilla Firefox.

Step 1 Step 2 Step 3a Step 3b Step 4a Step 4b Step 5 r5 r5 centres