double posts ftw
Big cube BLD DNF streaks are totally awesome!
At least I am getting material to work on with my post mortems though. That does lead me to a question though, Mike how do you post mortem centers? Because of Michael's point that the centers are indistinct, then there are always multiple errors that would give the appearance of any specific DNF state. What general process do you go about to discover the nature of a centers DNF? I have found that if the DNF locations contain one of my images intact, and the first letter location after this image contains the color of the first letter of the intact image, then I did not execute that image during the solve. Other than that, though, I have a hard time figuring out what I did wrong on my center cycles to achieve the DNF state that I end up with. Any recommendations?
3x3: 3 DNF's Yesterday
First one was off by 2 fliped edges and an R2
Second one was off by 4 edges Time: 6:33
Third was a completely scambled cube
Im getting there...
The way I do them isn't real fancy - nothing as elaborate as what you've been doing for non-centers. I try to actually remember my mistakes - often I can. I will first look at where the pieces are that are wrong, and then go through my memorization looking for those pieces. If I find them in the memo, many times I can immediately remember where my mistake was (oh, that's right - I went this way instead of that way on that commutator - how stupid!). Or I will often find one of the pieces in the memo twice - then it's obvious what went wrong. If I see the mistake at this level, it's usually not that hard to work out where all the pieces would have went based on my mistake, so I can check and make sure it accounts for everything. I would guess I figure it out that way over 50% of the time.
When that doesn't work, then I reapply the scramble, and check the memo. If it's wrong, problem solved; if it's right, then I will often just start solving that orbit of pieces, trying to remember as I go exactly how I did them during the original solve, and check after each pair of pieces. I almost always figure it out that way, if it comes to that.
I admit that if I ever have more than about 5 pieces of a given orbit wrong, I almost always just give up and don't bother.
Bolded part: Do you consider the possibility that you scrambled incorrectly before doing the actual blindfolded solve? Or do you scramble very carefully such as to avoid this possibility?When that doesn't work, then I reapply the scramble, and check the memo. If it's wrong, problem solved; if it's right, then I will often just start solving that orbit of pieces, trying to remember as I go exactly how I did them during the original solve, and check after each pair of pieces. I almost always figure it out that way, if it comes to that.
The whole comment: I think I may have to try that. I don't think I've actually stepped through the centers solve like this before. I can see there being some lingering muscle memory of what you actually executed, such that a red flag would go up when/if that doesn't match the memorization.
These tend to be the kinds of DNFs I get, though I will have wings and corners solved perfectly, but have over 5-6 centers unsolved. Keep in mind that my solving order as soon as I put on the blindfold is: corners -> centers -> wings. So this would mean that corners were done correctly such as to leave the centers in the same state that they were at the scramble. I considered the possibility that one of my algs is not actually being supercube safe, and is rotating some centers, but I use strictly BH. Because of it's conjugate and commutator nature everything there is supercube safe. I don't fix corner parity until after centers are solved, so that is not the cause of any possible center rotation. So my corners step, it seems, is not affecting or rotating any center groups.I admit that if I ever have more than about 5 pieces of a given orbit wrong, I almost always just give up and don't bother.
Next I solve centers somehow incorrectly, but leaving the wings unaffected. I know this because wings will be solved perfectly at the end. So I am not making setup and setdown errors, or the wings would be unsolved as well.
The only other conclusion I can come to is transposing of my R face lettering to my old scheme without realizing it. I have on occasion caught myself doing this on a solve before starting the commutator (or sometimes halfway through). In these cases I correct, but perhaps I don't always catch it? If I shot to the R face incorrectly once or twice I suppose this could account for 3-5 unsolved pieces. I find it hard to believe that I would shoot incorrectly to the R face all 3 or 4 times that the solve would require, but perhaps I do sometimes?
At Michael's suggestion I may try to fix up a much better quality 4x4x4 supercube than my current eastsheen cube and do solves on it like a regular 4x4x4. This way I could at least compare a DNF state to the original scramble using my [(DNF)]' and [X] techniques the same way. I wouldn't do this all the time, but at least until I got the same kind of bad centers DNF and could figure out the problem.
Ryan I may actually try this as well. I don't have a video camera, just a webcam. I will see what kind of frame-rate it gets, and see if this might be possible. Thanks for the suggestion, as that could certainly pick up any error quite easily!Originally Posted by RyanReese09