0:00 Solving the Inverse Scramble (1 switch) 0:15 How to Reverse Moves 0:41 NISS Example Solve (many switches) 3:57 Writing the Final Solution The example solve purposely avoided experimentation (forcing easier pairs, or OLL/PLL skips) to make it simpler and highlight how NISS works. In a real solve you have 1 hour, so make sure you try many more things than I did!
I did 2×2×2 and 2×2×3 block on normal scramble and then f2l-1 with reverse scramble. Now final solution will be normal solution+ reverse of inverse solution But that doesn't solving the f2l-1 state on cube??? Why??
At first I was a little bit confused, since you explained how to use NISS, but you didn't really explain why it works. I just wanted to share the explanation I found after thinking about it a bit with pen and paper. Let A be the solved state and B be the scrambled state. The normal scramble is a function f from and to the space of cube configurations such that f(A) = B, and the anti-scramble f' is its inverse function. Our objective is finding a sequence of "step" functions such that: 1)if s is their composition, s(B) = A 2)s is as efficient as possible. First, we start from B and try to find the most efficient step towards A (e.g., we find the shortest way to do a cross). Let sb1 be this step. Then we do the same thing, but starting from A. Since we only know how to go from scrambled states towards the solved state, we put new stickers on the cube at B so that it looks solved, then use the inverse scramble to go back to A (that now looks scrambled) and we find an efficient step from there. Let sa1 be this step. If sb1 is more efficient than sa1, then _sb1_ will be the _first_ step of our solution and the new problem is finding the most efficient step from sb1(B) to A. If instead sa1 is more efficient than sb1, the new problem is finding the most efficient step from B to sa1(A), and the _inverse of sa1_ will be the _last_ step of our solution. Now we just iterate. Say we have sa1 more efficient than sb1. Since with the current stickers B looks solved, we perform sa1(f'(B)) (i.e. f' followed by sa1) to find sa2, the next best step towards B. Now we make sa1(A) look solved, perform f(sa1'(sa1(A)) in order to arrive at B and find sb2, then choose between sa2 and sb2 and so on and so on until you have a chain of sa_i 's and sb_i 's that link together to form a path from A to B. Obviously, instead of putting new stickers every time we can solve the cube and pretend that we are in the same state. This process works because at each step you effectively have double the amount of moves towards a solution, and the best of many is equal or better than the best of few.
This is a good example NISS, but what I feel is missing is an explanation of *why* this works. That part still baffles me, which means I get really confused in trying to put it all together in practice.
1. Scramble switch: What any sequence of moves does is it sends each piece to a particular spot. Doing the inverse sequence will send each piece in the other direction. If a solved piece stays solved, then the sequence doesn't affect that piece, and therefore neither does the inverse. Example: When the cross is solved after doing [scramble] + [moves], then the cross must still be solved if you do [moves]' + [scramble]' since that is the inverse sequence. NISS takes advantage of this fact while continuing to add moves on both sides of the sequence. Reversing the sequence does not affect solved pieces.
2. The final solution: Suppose you ended on the regular scramble but did moves on the inverse as well. Then that means, from a solved cube, doing [inverse solution]' + [scramble] + [normal solution] results in the cube being solved once again, since that's how NISS works. So then, why does [scramble] + [normal solution] + [inverse solution]' solve the cube? In other words, if A B C solves the cube, why does reordering it to B C A also solve the cube? Well since A B C turns a solved cube back into solved, then it does nothing. Therefore A' (A B C) A also solves the cube, as it is just A' A. Cancelling the first 2 things, we get that B C A does nothing to the cube. And that is why NISS works.
this man gives tutorial for absolutely everything like, 3x3 two handed all the moethods 4 or 5, blindfolded cubes 3 or 4, megaminx, kilominx, one handed puzzles and many more that i havent seen. this is one fo those
I noticed that if you have rotations in the normal solution, you'll have to undo them before adding the inverse solution. This is something I had trouble with when starting to use niss and i think this should be adressed in some way because i know that a lot of people, me included, do rotations in fmc.
I think I have an idea how it works. Let A be the solved state and B be the scrambled state. The normal scramble is a function f from and to the space of cube configurations such that f(A) = B, and the anti-scramble f' is its inverse function. Our objective is finding a sequence of "step" functions such that: 1)if s is their composition, s(B) = A 2)s is as efficient as possible. First, we start from B and try to find the most efficient step towards A (e.g., we find the shortest way to do a cross). Let sb1 be this step. Then we do the same thing, but starting from A. Since we only know how to go from scrambled states towards the solved state, we put new stickers on the cube at B so that it looks solved, then use the inverse scramble to go back to A (that now looks scrambled) and we find an efficient step from there. Let sa1 be this step. If sb1 is more efficient than sa1, then _sb1_ will be the _first_ step of our solution and the new problem is finding the most efficient step from sb1(B) to A. If instead sa1 is more efficient than sb1, the new problem is finding the most efficient step from B to sa1(A), and the _inverse of sa1_ will be the _last_ step of our solution. Now we just iterate. Say we have sa1 more efficient than sb1. Since with the current stickers B looks solved, we perform sa1(f'(B)) (i.e. f' followed by sa1) to find sa2, the next best step towards B. Now we make sa1(A) look solved, perform f(sa1'(sa1(A)) in order to arrive at B and find sb2, then choose between sa2 and sb2 and so on and so on until you have a chain of sa_i 's and sb_i 's that link together to form a path from A to B. Obviously, instead of putting new stickers every time we can solve the cube and pretend that we are in the same state. This process works because at each step you effectively have double the amount of moves towards a solution, and the best of many is equal or better than the best of few.
I'm willing to bet money that some of the 55 dislikes are people who just wanted to learn how to solve a 3x3, and ended up here accidentally. The rest probably just didn't get it, which is fair, even with JPerms excellent explanations, this stuff can be kinda tricky to wrap your head around.
Sorry for the stupid question, but I'm very new to NISS and FMC. For both the regular scramble and inverse, it looked like you addressed the same cubies each time. I mean, you did the orange cross on reg scramble, then tried orange cross on inverse (instead of a different color). For first pair, whichever colors they were, you found number of moves to solve them on both scrambles. Same for third pair (y/o/g I think). Should you always do this or did you just do it for a simpler tutorial? I mean, in real solving, would you maybe do y/o/g as your first pair on regular scramble but then see that r/w/b was better for first pair on inverse? Or do you always have to look only at the same pair of cubies on both scrambles? Hope this makes sense. Thanks.
Essentially: Using just the normal scramble is like doing a maze by starting from the start Using just the inverse is doing the maze from the end of to the start Using both is basically meeting in the middle from both sides
Doubt:" as you said that you got a scramble and find the reverse scramble, but in the WCA they will give a paper with 80 bars should we bring a separate book to do the scramble and reverse scramble
I saw this too soon. And my head hurts now. It's amazing tho. I am just trying to figure out how to scramble my cube when you do an example solve so I can follow. What position do I start in, and do I read the scramble forwards or backwards?
0:00 Solving the Inverse Scramble (1 switch)
0:15 How to Reverse Moves
0:41 NISS Example Solve (many switches)
3:57 Writing the Final Solution
The example solve purposely avoided experimentation (forcing easier pairs, or OLL/PLL skips) to make it simpler and highlight how NISS works. In a real solve you have 1 hour, so make sure you try many more things than I did!
F.3.
I did 2×2×2 and 2×2×3 block on normal scramble and then f2l-1 with reverse scramble.
Now final solution will be normal solution+ reverse of inverse solution
But that doesn't solving the f2l-1 state on cube??? Why??
Hey j perm! Can you help me? I don't understand because you are not told the scramble.
Ive been wondering cant u just reverse the scramble in fmc?
@@wonton9503 No bro!!!!! That was the problem!!!😂😂
FMC participantion rate increases to 100%
Utsav Manandharz But hey, learn the algorithms first!
You don't need new algs actually
To this means that it doubles
That's such a clever and subtle way to present it. Switching to left and right halves to drive home the Normal vs Inverse solutions is genius. 👏👏👏
I didn’t even notice that
I love these new FMC videos! Keep up the good work!
SKewber
yeah
Imagine practicing FMC so hard that you start doing crazy stuff on regular 3x3 solves. (Like Jack Cai does with bld)
As always: Good tutorial!
Thanks! I wish I could do NISS in a speedsolve haha
@@JPerm haha same
You should add:
“From the national record holder”
I don't think he held National record during the time of upload.
@@Usha_Prasanna that's why he said to add it
At first I was a little bit confused, since you explained how to use NISS, but you didn't really explain why it works. I just wanted to share the explanation I found after thinking about it a bit with pen and paper.
Let A be the solved state and B be the scrambled state. The normal scramble is a function f from and to the space of cube configurations such that f(A) = B, and the anti-scramble f' is its inverse function. Our objective is finding a sequence of "step" functions such that:
1)if s is their composition, s(B) = A
2)s is as efficient as possible.
First, we start from B and try to find the most efficient step towards A (e.g., we find the shortest way to do a cross). Let sb1 be this step. Then we do the same thing, but starting from A. Since we only know how to go from scrambled states towards the solved state, we put new stickers on the cube at B so that it looks solved, then use the inverse scramble to go back to A (that now looks scrambled) and we find an efficient step from there. Let sa1 be this step.
If sb1 is more efficient than sa1, then _sb1_ will be the _first_ step of our solution and the new problem is finding the most efficient step from sb1(B) to A. If instead sa1 is more efficient than sb1, the new problem is finding the most efficient step from B to sa1(A), and the _inverse of sa1_ will be the _last_ step of our solution.
Now we just iterate. Say we have sa1 more efficient than sb1. Since with the current stickers B looks solved, we perform sa1(f'(B)) (i.e. f' followed by sa1) to find sa2, the next best step towards B. Now we make sa1(A) look solved, perform f(sa1'(sa1(A)) in order to arrive at B and find sb2, then choose between sa2 and sb2 and so on and so on until you have a chain of sa_i 's and sb_i 's that link together to form a path from A to B.
Obviously, instead of putting new stickers every time we can solve the cube and pretend that we are in the same state.
This process works because at each step you effectively have double the amount of moves towards a solution, and the best of many is equal or better than the best of few.
This is a good example NISS, but what I feel is missing is an explanation of *why* this works. That part still baffles me, which means I get really confused in trying to put it all together in practice.
1. Scramble switch:
What any sequence of moves does is it sends each piece to a particular spot. Doing the inverse sequence will send each piece in the other direction. If a solved piece stays solved, then the sequence doesn't affect that piece, and therefore neither does the inverse.
Example: When the cross is solved after doing [scramble] + [moves], then the cross must still be solved if you do [moves]' + [scramble]' since that is the inverse sequence.
NISS takes advantage of this fact while continuing to add moves on both sides of the sequence. Reversing the sequence does not affect solved pieces.
2. The final solution:
Suppose you ended on the regular scramble but did moves on the inverse as well.
Then that means, from a solved cube, doing [inverse solution]' + [scramble] + [normal solution] results in the cube being solved once again, since that's how NISS works.
So then, why does [scramble] + [normal solution] + [inverse solution]' solve the cube? In other words, if A B C solves the cube, why does reordering it to B C A also solve the cube?
Well since A B C turns a solved cube back into solved, then it does nothing. Therefore A' (A B C) A also solves the cube, as it is just A' A. Cancelling the first 2 things, we get that B C A does nothing to the cube.
And that is why NISS works.
TLDR: you don't really have to know why, you just have to memorize the rules, which are pretty simple to follow.
Who just clicked with no intention of studying fewest moves?
So true!
Yes
Dude I'm gunna do FMC now
Me!
me
This is actually amazing , I think I'm going to enter FMC now
Your handwriting is amazing. Makes me happy
Like at 0:57 the way he wrote 'normal' is so amazing!
Can you make more tutorials on FMC
No
Jk
@@ferretonastick300 You're welcome!
this man gives tutorial for absolutely everything like, 3x3 two handed all the moethods 4 or 5, blindfolded cubes 3 or 4, megaminx, kilominx, one handed puzzles and many more that i havent seen. this is one fo those
Now that I have watched this again I realise how perfectly structured this tutorial is. J perm=God perm
Best NISS tutorial I have ever seen.
Me:
*Sees video title*
Nah, I don't do FMC
*Sees video creator*
Yeah, I'm gonna click on it
this is by far the best vid that i have seen on niss
like if you agree
Damn I really dont have the patience for FMC. You have to be a real trooper to sit there for an hour and do this.
You and SpeedCubeReview have the Best FMC videos on youtube!!!
I noticed that if you have rotations in the normal solution, you'll have to undo them before adding the inverse solution. This is something I had trouble with when starting to use niss and i think this should be adressed in some way because i know that a lot of people, me included, do rotations in fmc.
It’s amazing that you have nearly doubled in subs since Christmas
There's no better YouTutor than JPerm
Thanks dude! That really helped a lot! My very first FMC attempt was 36 moves just using the NISS method dat u taught!
congrats!
Wow it just happened, jperm did a fmc tutorial, I just can't believe this, it has to be a dream
Thanks for tutorial😍
I was waiting for this video from the previous one😘😘
EDIT:thanks for 3 likes😅
@@___wahid___
No I didn't
It's a coincidence
A "how to find insertions" tutorial would be nice
Never quite got that...
Until now, thank you!
it's a good tutorial eventhough it is really confusing but i get the idea after twice of repeat
can you do more of this in the future?
great job trying to keep it as clear as possible for us. the left vs right hand side of the screen helped a lot
Very interesting! I wonder how exactly inverting the moves and scramble works, and what the mathematical proof for it is.
@@charbelkassab6353 it's a shame that TH-cam hides dislikes; this comment shall have to suffice
I think I have an idea how it works. Let A be the solved state and B be the scrambled state. The normal scramble is a function f from and to the space of cube configurations such that f(A) = B, and the anti-scramble f' is its inverse function. Our objective is finding a sequence of "step" functions such that: 1)if s is their composition, s(B) = A
2)s is as efficient as possible.
First, we start from B and try to find the most efficient step towards A (e.g., we find the shortest way to do a cross). Let sb1 be this step. Then we do the same thing, but starting from A. Since we only know how to go from scrambled states towards the solved state, we put new stickers on the cube at B so that it looks solved, then use the inverse scramble to go back to A (that now looks scrambled) and we find an efficient step from there. Let sa1 be this step.
If sb1 is more efficient than sa1, then _sb1_ will be the _first_ step of our solution and the new problem is finding the most efficient step from sb1(B) to A. If instead sa1 is more efficient than sb1, the new problem is finding the most efficient step from B to sa1(A), and the _inverse of sa1_ will be the _last_ step of our solution.
Now we just iterate. Say we have sa1 more efficient than sb1. Since with the current stickers B looks solved, we perform sa1(f'(B)) (i.e. f' followed by sa1) to find sa2, the next best step towards B. Now we make sa1(A) look solved, perform f(sa1'(sa1(A)) in order to arrive at B and find sb2, then choose between sa2 and sb2 and so on and so on until you have a chain of sa_i 's and sb_i 's that link together to form a path from A to B.
Obviously, instead of putting new stickers every time we can solve the cube and pretend that we are in the same state.
This process works because at each step you effectively have double the amount of moves towards a solution, and the best of many is equal or better than the best of few.
This tutorial is better than other NISS tutorials because it's explained in an easier way. lol
This was actually really cool. I've been somewhat interested in FMC, but never had any idea of how to really get a decent solve. Thanks!
NISS= New Intelligent Solving Solution
J Perm, can you please show us your collection of all your cubes! I'd love to watch it
Will there be insertion tutorials because I am averaging 35 on fmc and I want to improve?
There already are many tutorials.
@@Aneurin_Hunt not really but ok.
TCY jayden McNeill and cubing world are two. Look it up before telling me there aren't tutorials.
I know it’s old but 2=many?
Youre a beast! I’ve been waiting for this!
Pls do commutators and conjugates also pls pls btw love u from India
I don't think he uses commutators,if you are talking about bld. He uses pochman corners and M2 edges
@@anshitsrivastava4078 I'm pretty sure he's talking about commutators and conjugates in FMC
@@ndas1114 That's an Insertion :P
@@Computteren yeah but basically they are comms
@@ndas1114 are you sure about that??
his handwriting is nice
Are you going to do a video on the Valk 4?
Moral of the Video- There are always two ways to solve a Rubix cube in Fewest Moves
Man this makes me feel like I’m just learning the cube all over again
Thank You J Perm. Your channel keeps me interested in cubing.
I love your handwriting
thank you for making a tutorial, more example solves next time?
You don't really have example solves in fmc. That makes no sense.
omg that handwriting is so satisfying :O
Finally a Good NISS tutorial ! Thank you very much!
I'm willing to bet money that some of the 55 dislikes are people who just wanted to learn how to solve a 3x3, and ended up here accidentally. The rest probably just didn't get it, which is fair, even with JPerms excellent explanations, this stuff can be kinda tricky to wrap your head around.
i saw your video on how to setup cubes. Where do you buy spare springs?
I usually don't change the springs, and all my springs are just from other cubes I own.
Thank you J Perm a lot I thought I will give up but now I did it :)
Thats alot of writing and inversing. I had to rewatch it several times to understand it haha
Thanks for the tutorial
Definitely not watching this an hour before my event
J Perm is right handed and has an awesome handwriting!
Can I write X Y Z etc. in the solution?
make a video about how to improve on 7x7
Can you do block building tutorial plz
And is NISS or blockbuilding same level or NISS is harder than blockbuilding?
QualityBlood niss is basically cfop with a few edits, but blockbuilding creates blocks of pieces, which is changing the solution method
Can you do a heise method for FMC?
I love your channel it is so amazing and I decided to start my own. I would love some support
@J Perm, can you use different cross colors in the different scrambles?
Could you do a block building tutorial?
Sorry for the stupid question, but I'm very new to NISS and FMC. For both the regular scramble and inverse, it looked like you addressed the same cubies each time. I mean, you did the orange cross on reg scramble, then tried orange cross on inverse (instead of a different color). For first pair, whichever colors they were, you found number of moves to solve them on both scrambles. Same for third pair (y/o/g I think). Should you always do this or did you just do it for a simpler tutorial? I mean, in real solving, would you maybe do y/o/g as your first pair on regular scramble but then see that r/w/b was better for first pair on inverse? Or do you always have to look only at the same pair of cubies on both scrambles? Hope this makes sense. Thanks.
If I have a sequence of moves such as D’ F F’ D; when I cancel out F’ and F should I also cancel out D’ and D?
Essentially:
Using just the normal scramble is like doing a maze by starting from the start
Using just the inverse is doing the maze from the end of to the start
Using both is basically meeting in the middle from both sides
question: what is your recommended timer for fmc
Great tutorial video, I've always wanted to learn FM, but I just don't know how to access any tutorial on the Internet
This tutorial helped me learn it! Thanks j perm
This video is 4 minutes, you commented this 2 minutes after it uploaded
Evrich Markell Robert - Speedcuber boi I watched it
I'm completely lost after 0:25
very nice explanation! I'm gonna try it!!
19 subs for 150k congrats
J perm is it better to buy the gan 356 x from amazon or from speedcubeshop?( I’m in Europe btw)
Nikola NINKOV scs
Doubt:" as you said that you got a scramble and find the reverse scramble, but in the WCA they will give a paper with 80 bars should we bring a separate book to do the scramble and reverse scramble
On my first real attempt (Not knowing EO or good insertions) I managed a 41. Pretty neat with NISS
Should I try to improve my speed with the beginners method?
does cube rotations count as a move?
no, but it's easier in the long run to avoid them and just write the correct face like I did
@@JPerm ok thanks a lot dylan :)
Great job Dylan, even though it's a bit confusing :p
I saw this too soon. And my head hurts now. It's amazing tho. I am just trying to figure out how to scramble my cube when you do an example solve so I can follow. What position do I start in, and do I read the scramble forwards or backwards?
Hi,when check your solution cube was not solved ?
Very interesting. I hope you have more FMC tutorials.
so the after building first 2 layers, we just write down alg for oll and pll?
3:59. Are you crying, mein fuhrer?
Can you make a video about cross example solves for one handed ?it take me 6s to solve it , really struggle
I don't get the "inverse scramble" How is it allowed to change the scramble?
Should I use petrus for this
Will you make a review on the gans megaminx
Can you do more terms every Cuber should know f. e. CLL, COLL und CMLL what's the difference?
Plz can you do tutorials on FMC
Could you do a video with some more example FMC solves like this?
this is just CFOP, I might do more in the future after showing more efficient stuff
Do we even have to reverse the scramble we are working on to switch scramble
Thanks for the tutorial I’m trying to get into fmc
If an Mgc 4x4 came out, would you buy it and do you think it would be good?
Amazing, only had to watch once to understand the whole concept
Great video but why so less views.It was really useful but it is hard to understand in the first time
Learning this bc there is a comp in 3 weeks :)
can you do M, E, S and wide moves??????
yes
Thanks a lot man😀😀
random question but, should i buy the WRM or GTS2M. both stocked not spring swapped or anything. pls reply jperm
I just started cubing i have a Rubik's brand i want to change yuxin little magic or gans air master for a begginer?
Yes yuxin is very good
I have a solve where i did everything on the inverse scramble. The inverse of the inverse solution doesnt work
More FMC tutorials ?
To say I'm confused would be an understatement. Lol
I’m looking for a new main and because my hands are getting tired with the mf3rs2 M is there any good cubes I should get