I have watched several videos on F2L, but most of them simply show alg's or discuss how to solve specific cases. I'm looking for more of a "strategy" on how to improve scanning and pair recognition.
Right now I scan for corner pieces in the top layer first. If I find one, then I scan for the corresponding edge in the top layer, moving to the bottom layers as needed. Do you normally solve the first pair you come across, or do you look for the "easiest" pair to solve first? Any specific tips are appreciated.. Thanks.
Yesh - some good F2L 'tactics'. Firstly, learn to solve in the back slots. Mirror your F2L cases (i.e. learn how to solve them in all 4 slots) Bob Burton's Cubewhiz site has an excellent page on
solving cases from all 4 corners.
Secondly, and this complements the first point, use process of elimination. By this I mean:
> There are only 12 edge pieces on the cube
> Of these, 4 are already solved (cross edges)
> This leaves 8 edge pieces which might be the one you need (for the first F2L pair)
> 3 of these are on the F (front) face, so you can easily see both colours
> 2 more are on the UR (upper face, right side) and UL (upper face, left side), so you can glance at these to see both colours
> the remaining 3 edges are on BR (back right), BL (back left) and BU (back/upper face), so you can only see one colour on each of these edges
So what? OK, let's say you have the red, blue, white corner in UFR. You can see that the blue and red edge isn't on the F face, or in the UR or UL positions. This means it
must be one of the 3 edges on the B face.
Now there are only 3 remaining edges on the cube that will have any blue on them. These will be blue/red (the one you want), blue/orange, and blue/yellow. There is a chance, albeit very low, that the 3 stickers you can see on the B face edges are all blue. With my basic knowledge of probability, I'd say 1/8 or 12.5%. In that case, a 1/4 turn cube rotation (y or y') will show you where the blue/red edge is. However, there's a pretty high probability (87.5%) that you will be able to see a red edge. As you already know all 3 blue edges are on the B face, being able to see that one red sticker will tell you this is the edge you need. Even if you can't see the red sticker, you might be able to see the yellow sticker and the orange sticker, meaning you'll still be able to work out which one the blue/red edge is without rotating.
Now you're on the second F2L pair. This time you go for blue/orange and, once again, you find the corner in UFR position. If your already-solved pair is in the FR or FL slot, there are still 3 possible 'hidden' edges on the B face. So this is where the first tactic comes in to play - if you've placed your solved pair in the BR or BL position, this means there's only 2 potential 'hidden' edges on the B face. This means there are 7 remaining edges, of which you can see both sides of 5/7. You have a 5/7 (71%) chance of already being able to see the edge you need, and only a 2/7 (29%) chance that you can only see one sticker on it. Again, there's a chance both the stickers you can see on the 'hidden' edges will be blue - I believe this is only a 25% chance.
In other words, after solving your first F2L pair in one of the back slots, you have a 71% chance that you will be able to see your next target edge straight away. If you can't (only a 29% probability), you will still have a 75% chance that you can see at least one non-blue sticker on the 'hidden' B face edges, which will immediately tell you which one your target edge is. I.e., if you know both remaining unsolved back (hidden) edges are blue, one will be blue/yellow, the other blue/orange (the edge you're after). If you can see a yellow sticker, the edge you need is the other one. If you can see an orange sticker, you know that's your edge.
If you solve your second pair in the other back slot, you can now see both stickers on all remaining edges, except BU. If you can't find your edge among the visible pieces, that's where it's hiding. There's a 100% chance you can now find your remaining 2 edges with absolute certainty, and without rotating the cube.
Cube-rotations, although generally frowned upon, aren't as bad as all that. Reducing rotations will make you quicker, but my point is more that you can find your target piece without rotating, rather than that you
shouldn't rotate.
It seems that at the moment you're not particularily comfortable with all of the F2L cases, which can only be resolved with practice. Just do tons and tons of solves. Once you get your F2L under 20 seconds, and can do the F2L pairs without looking at them, try looking for the next F2L pair while solving the one at hand, which is called lookahead.
Agreed - lookahead is the single thing that will make the most difference to your solves, once you get to about 30 seconds. Until then, I'd work on recognising cases and finding pairs quickly. Hopefully the advice above will help a little with this.
