this method would be very difficult to make viable at a high level because doing EO and the face-1 takes probably ~20 moves which is like move count for 2 or 3 pairs in cfop then the belt seems awkward because you have to do a lot of D moves and the IPFL algs don't have very good finger tricks
Ermmm actually 🤓 I am going to reply to this year old comment cuz I am bored. Anyways, eoface-1c only takes 9 moves on average not 20. And I have regennes IPFL to be RUD gen with good ergo. And tricks like R2 E’ R2 E and R2 E2 R2 commutators can reduce belt movecount
for the last step (the 4th one where you permute both layers) you could just permute the corners (8 cases), and then solve the edges with M2 U2 M2, S' M' U M U2 S, M2 U* M2 U*' M U2 M2 U2 M, and edge only pll algs (and with some U and D moves to adjust the layers)
why would you solve belt after almost finishing the 1st side, if you can use dseudosloting. something like eo+incorrect cross and then pseudosloting would work great. the other option would be to do eo line with any edges, and then blockbuild the rest. theoretically you can also give up on eo, and just make an incorrect cross, and then solve pseudopairs while orienting edges, and then you can insert the last edge and last corner at the same time while doing WV. that would be a totally different method but based on similar idea
I find that as a CFOP solver, using F2L to do the first face and the belt at the same time speeds up the solve. I was able to get a sub-30 by doing so (I average 17 seconds with CFOP)
I haven’t tested this out myself, but I wonder if it’s only faster because of F2L familiarity. The freedom of face building might make the “normal” way faster. Plus you can just move the open corner around to insert edges since bottom layer orientation is irrelevant.
I remember finding this on the speedcubing wiki and trying to solve with it, its very slow even with cls and ipfl, It's better to just solve 3 pseudo f2l pairs after eo, insert last corner using cls. So the time go like so Eo 2s White face 7s (hard to plan due to inspection used for eo) Inserting belt edges 4s (Takes lots of u and d moves) Cls 3s Pll 1 2s Rotation .5 Pll 2 2s Total 22.5s for very fast execution Ipfl not worth it its not very fast
I'm sorry if there are any typing issues; I quickly typed this on my phone. 1. EO doesn't take 2 seconds. Also, you should not consider EO and the white face as 2 different steps. With some exercise, you can plan almost them in the inspection. 2. I agree that inserting the belt edges takes some time because of the U and D moves, but this is one thing that can be fixed by doing the first two steps (EoFace-1C and the Belt) in some kind of ZZ style. EoPseudoCross + PseudoF2L is already an improvement/suggestion that I added to the Iacob Method webpage. 3. You can predict the CLS algorithm before inserting the last belt edge so that it wouldn't take that much time to recognize the case. 4. I know that IPFL it's not very fast, but doing a z2 rotation and then doing PLL is 100% slower.
I forgot to mention that before doing PLL you can also see the IPFL case, having almost no pauses between permuting the last layer and permuting the first layer.
@@im_razvan saying that IPFL is 100% faster than a z2+PLL is only true as long as there are fast a low regrip algs for IPFL, which is not true at the moment. I'd rather do a big rotation at the start and then a 0 regrip alg than pretending that not doing a z2 but doing a trash alg with 3-4 regrips is any better
1. Have you ever tried the belt method? Try it and see how much of a "better" method it is. 2. I main the Roux method and yes, I can say that it is better. The Iacob Method is good only if you know PBL.
I don't think this method is worth learning it because it's too inefficient and has too many algs for a beginner 1) You solve the (almost entire) first face but incorrectly so you have to re solve it later with unintuitive algs and more move count. I think it's better to use the freedom you have at the beginning of the solve to correctly solve the first layer 2) You solve one piece at a time, which can be good for beginners but when you want to get faster you'll have to learn an entire new method like roux or cfop cause learning all the algs to solve both the first and last layer seems pretty hard. So instead I think it's better to directly learn cfop or roux when you're beginner, cause it offers less algs, less move count and is more intuitive.
On the Iacob Method website it's says that it is recommended to kinda make PseudoF2L pairs instead of doing that. Also making just a pseudo layer can help with lookahead