Memory Methods

[MatsBergsten]

We learn from you masters that it is good to memorize in pairs.
And I agree to that, it is good to have two letters or cubies as one
memory piece, that amounts to have to store only half as much information

But on the other hand we learn to memorize in pairs <person> <action>.
(Now I do not two's but "four's" like <person> <action> <object> <adverb>
but the same objection applies). Now Mike for instance describes how he
memorize CB as cab. But do you let go of the <person> <action> then?
Or do you have two sets of images, the first with <persons> and the second
with <objects>. (But both sets with pairs of letters).

So MHCB would be Mike Hughey (goes in a) CaB?

Chris just describes that two letters (KL or whatever they were) stands for a
friend who is a drummer. Would KLCB be the drummer going in a cab or....

I tried to compile a list of persons for every letter combination but only managed
60% or so. It is a big difference between 24 persons or 24x23...

 

[Mike Hughey]

Well, actually, I have 4 different sets of 24 "persons", each of which gets a single letter. I use them for my buffer pieces, since I don't have fixed buffers for centers. It's also nice because even where I have fixed buffers, I can use a different set of the "persons" for each cube when doing multi, which helps me keep from getting different cubes mixed up. (My four groups of persons are people I know, movie and TV stars, musicians, and cartoon characters.)

For most of my memorization I have a single person (representing the buffer piece) acting on up to three images (which represent two pieces per image). Usually it's three, but if it comes time to change buffer pieces or I run out, then it will be less. So I have a single location representing a buffer piece going through 6 other pieces.

From Ben Pridmore's video description of his method memorizing a pack of cards, it sounds like he uses a similar thing with the three images per location (although he doesn't have a person acting on them). So three images seems like a logical choice. (I decided to do that before I found out it's what Ben did - it was coincidence that I wound up doing the same thing.) Of course, the difference is that Ben has a lot more images than me.

Mats, I would think that coming up with a set of images for all four cases (person, action, object, adverb) would be pretty spectacular, and it might give you a big advantage on the rest of us, but I can see where it might be daunting to do. If you really decide to go for it, good luck to you. You might become nearly unstoppable - I might have to do the same if you start to regularly beat me someday. You might really be onto something powerful there.

 

[MatsBergsten]

Well, actually, I have 4 different sets of 24 "persons", each of which gets a single letter. I use them for my buffer pieces, since I don't have fixed buffers for centers. It's also nice because even where I have fixed buffers, I can use a different set of the "persons" for each cube when doing multi, which helps me keep from getting different cubes mixed up. (My four groups of persons are people I know, movie and TV stars, musicians, and cartoon characters.)

Mats, I would think that coming up with a set of images for all four cases (person, action, object, adverb) would be pretty spectacular, and it might give you a big advantage on the rest of us, but I can see where it might be daunting to do. If you really decide to go for it, good luck to you. You might become nearly unstoppable - I might have to do the same if you start to regularly beat me someday. You might really be onto something powerful there.

No no, you misunderstood me. I am not intending to memorize some 21x18x15x12 images, let alone 23x21x19x17 or some other daunting number. I still DO NOT use images, but I have to understand what you (and Chris/Daniel and other image/memo guys (Sorry Sarah ) actually do. What I do is memo Person Action Object Adverb/Adjective as a "Silly Word Sentence". And I do that on the fly, which takes time. Like you I have several (but only two) sets of persons (men / women) which I use for x-centers and +-centers. But I have been toying with the idea of some 400 persons (that is not a daunting number) so as to use 2 letters for person and 2 letters for action/object. My problem is that memo time variates (is that a word?) too much. In good cases I make a 3x3 in 30 seconds but sometimes up to 2 minutes (as in last comp ). I guess if you have two ready sets of 400 persons and 400 action/objects the average memo time would be lower.

As to beat you or the others we can dream of that. But I can also be satisfied with
beating you and Chris when you dnf:s as in last weeks comp. That was a nice 5x5BLD result list

Thanks for your answer btw.

 

[Mike Hughey]

No no, you misunderstood me. I am not intending to memorize some 21x18x15x12 images, let alone 23x21x19x17 or some other daunting number. I still DO NOT use images, but I have to understand what you (and Chris/Daniel and other image/memo guys (Sorry Sarah ) actually do. What I do is memo Person Action Object Adverb/Adjective as a "Silly Word Sentence". And I do that on the fly, which takes time. Like you I have several (but only two) sets of persons (men / women) which I use for x-centers and +-centers. But I have been toying with the idea of some 400 persons (that is not a daunting number) so as to use 2 letters for person and 2 letters for action/object. My problem is that memo time variates (is that a word?) too much. In good cases I make a 3x3 in 30 seconds but sometimes up to 2 minutes (as in last comp ). I guess if you have two ready sets of 400 persons and 400 action/objects the average memo time would be lower.

No, I think maybe you misunderstood me. I was saying it sounds like you might be planning four 400-image lists - one 400-image list for persons, another for actions, a third for objects, and a fourth for adverbs. I can see how that could be very powerful - you could store 4 images at each location as a sentence. It might be really nice.

As to beat you or the others we can dream of that. But I can also be satisfied with
beating you and Chris when you dnf:s as in last weeks comp. That was a nice 5x5BLD result list

Hey, that's how I beat Chris for US Nationals! And you won't get that lucky this week. At least not on 5x5x5; I'm not sure about 4x4x4.

 

[spdcbr]

I'm adding alternitives:

YO: yo-yo
OY: Oily
YG: Ignite
GY: Guy
YR: You
RY: Rock
OG: Ogre
GO: Gong/Exloplosion/Loud noise
OB: Boat
BO: Box
BR: Brick/Brain
RB: Root beer
RG: Rage/Regret
GR: Grape/Grass
OW: Owl/Ow
WO: Wok/World
BW: Eyebrow
WB:Website
RW: Write
WR: World Record/Rotten
GW: Game/Wish
WG: Wig

 

[spdcbr]

Sorry, for double posting, but I figured thast this must have a seperate one. where can I find a link for hex decimals?

 

[rjohnson_8ball]

Sorry, for double posting, but I figured thast this must have a seperate one. where can I find a link for hex decimals?

Google "hexadecimal" and also look at binary notation. I use hexadecimal to memorize edge orientations for 3OP blindfold method. Each edge is either correct (0) or needs to be flipped (1). See cubefreak.net to understand how to figure out if an edge needs flipping. Each layer has 4 edges. You might memo 0 1 1 0 on the top layer, 1 1 1 0 on the middle layer, and 0 0 1 0 on the bottom layer. Each of these 3 binary numbers 0110 and 1110 and 0010 are respectively 6 and E and 2 in hexadecimal. Just memo "6E2". Sometimes you can dream up something to help you remember it, like "6E2" sounds like "sexy too". During execution time, decode it to identify the bad edges.

 

[spdcbr]

Sorry, for double posting, but I figured thast this must have a seperate one. where can I find a link for hex decimals?

Google "hexadecimal" and also look at binary notation. I use hexadecimal to memorize edge orientations for 3OP blindfold method. Each edge is either correct (0) or needs to be flipped (1). See cubefreak.net to understand how to figure out if an edge needs flipping. Each layer has 4 edges. You might memo 0 1 1 0 on the top layer, 1 1 1 0 on the middle layer, and 0 0 1 0 on the bottom layer. Each of these 3 binary numbers 0110 and 1110 and 0010 are respectively 6 and E and 2 in hexadecimal. Just memo "6E2". Sometimes you can dream up something to help you remember it, like "6E2" sounds like "sexy too". During execution time, decode it to identify the bad edges.

I don't get how to read the binary numbers. An what about L and R?

 

[rjohnson_8ball]

I don't get how to read the binary numbers. An what about L and R?

Did you use Google yet? Did you read the 3OP section at cubefreak.net?

Google binary, and look at a Wikipedia description. In base 10, the last digit of an integer is known as the "ones" column. The next digit to the left of that is the number of "tens". Next to that is the number of "hundreds". And so on. Each digit to the left indicates a value 10 times larger. In base 2 (binary), each digit in the left direction indicates values 2 times larger. So to count from 0 to fifteen in base 2, you would write 0000, 0001, 0010, 0011, 0100, 0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110, 1111. For example 1011 in binary means (1 times 8) plus (0 times 4) plus (1 times 2) plus (1 times 1) which totals to 13 in decimal.

Determine ahead of time a consistent way to order all edges. For example, the first edge might be UF, and the 2nd, 3rd, 4th might be clockwise from that on the U face. Then the 5th through 8th edges are on the layer between U and D. The 9th through 12th are on the D face.

Using the 3OP method mentioned at cubefreak.net, a "properly oriented" edge can be rotated into its proper position and orientation by using R,L,U,D,F2,B2 twists only. The edge would require flipping if a single F or B twist would be required to orient it into its proper position.

 

[hehehaha!]

I have a problem.
I use 3OP. But whenever I am doing set up moves for corner permutation. I will mess it up. I just cant reverse my moves in the right order. Is there some way you remember the set up moves. It also takes me a hell lot of time to figure out the set up moves for corner permutation. I think I need to learn a few more algos for corner permutation. I currently use only the (143) and (412) for CP.

 

[Sakarie]

You can easily learn how to cycle any three corners in top layer (pre-oriented). Learn (143) multi-mirrored and backwards, and you have it!

 

[trying-to-speedcube...]

I decided to explain my corner memory method. I use classic Pochmann for corners, but I'm sure you can use this memory method for every method.

Don't see the cube as a cube, but see the cube as a hexagon by holding it so your buffer piece is invisible and you see a hexagon with 6 corners and one corner in the middle (the polar opposite of the buffer position) Now, in this hexagon you visually memorize the permutation of the pieces, which is really easy because you have a somewhat 2-dimensional shape in your mind, and with letters you memorize the side you have to shoot to, giving each side a letter that is common in your language. Those are A, E, M, R, S and T for me. Because these letters are so common, you should be able to make easy words out of them.

During solving, you combine the letter with the visually memorized position and then you know what sticker exactly to shoot to.

 

[CuBeOrDiE]

Hmm...that seems nice; I might use it. I've only been blindfold solving for about 5 days, but I'll tell people what I use. It's just a basic number system. I give every corner sticker and every edge sticker a number from 1 to 24. Then I simply see what sticker on what piece must be placed in the up layer to be swapped with the buffer using a T perm. I simply memorize its number. I do that until the cube is solved. Between corners and edges, I memorize Y for yes, there is parity, or N, for no parity. However, your system for corners seems interesting, so I might use it.

 

[danjon]

I use BH method now.last month I changed my memory method many times,such as:
1,Tradition code method.I used letters.numbers and Chinese characters at the same time,same codes for edge and corner.Finally I found I couldn't remember them clearly.
2,Codes for edge,Roman Rooms for corner.I could remember them clearly,but not fast.
3,Different codes for edge and corner.There are 43 different codes(22 for edge,21 for corner),Remembered nearly clearly,but I feared forget them when I solving the cube.
4,Same edge codes with method 1,and 7 codes for corner.This method is unique for Chinese,because each Chinese character has a tone besides the pronunciation.There are 4 different tones,I chose 3 of them in the codes.So each corner have a pronunciation,3 orientation have different tones.For example I used 妈（mother）,马（horse）and 骂（condemn） for UFL,FLU and LUF,all the pronunciation of them are Ma,but the tone are different.Used this method I could remember the codes clearly,and very fast.But there was another problem,when I solving the corner,I had to use much time to think which formula should I use:fp
5,Same edge codes with method 1,and same method with Haiyan Zhuang in corner.This method use different arrows to show the relationship of two corner.Finally I found if I have time to remember these arrows,I should remember the corner and the orientation directly!I trained this method about 15 days.it is not very fast now,but I think after one or two month when I familiar with all formulas of corner,it may faster and faster.

 

[badmephisto]

The best posts from this thread should be documented on the Wiki I looked but noone tried. Instead I see people explain the same thing over and over and over again
And very good posts by very fast bld cubers are stuck in the 26 pages of chaos

 

[Chuck]

The best posts from this thread should be documented on the Wiki I looked but noone tried. Instead I see people explain the same thing over and over and over again
And very good posts by very fast bld cubers are stuck in the 26 pages of chaos

I'm so agree with him.
(Wow Badmephisto :rolleyes: I learned BLD from his video tutorial)

I'm starting to read this thread seriously now, page by page. There are so much useful informations, yet I haven't read all the 26 pages. Maybe I'll share my method too someday.

 

[F.P.]

I'm starting to read this thread seriously now, page by page. There are so much useful informations, yet I haven't read all the 26 pages. Maybe I'll share my method too someday.

Your own/current method seems to work very well for few...are you planning on changing to a method just because it works for someone else?

Just wanted to mention that the memory methods which one has created (by a created method I mean of course a modified version of already existing methods) him/herself are usually the best ones (for this certain person).

 

[Mike Hughey]

I'm starting to read this thread seriously now, page by page. There are so much useful informations, yet I haven't read all the 26 pages. Maybe I'll share my method too someday.

Your own/current method seems to work very well for few...are you planning on changing to a method just because it work for someone else?

Just wanted to mention that the memory methods which one has created (by a created method I mean of course a modified version of already existing methods) him/herself are usually the best ones (for this certain person).

I quite agree. I still don't understand how Tim Habermaas's strange memory method (using images for single letters instead of pairs) could possibly get good results. But considering his 24/24 in just over 2 hours, I'd say it must be pretty close to perfect for him - I hope someday I come up with a system that works as well.

 

[tim]

I'm starting to read this thread seriously now, page by page. There are so much useful informations, yet I haven't read all the 26 pages. Maybe I'll share my method too someday.

Your own/current method seems to work very well for few...are you planning on changing to a method just because it work for someone else?

Just wanted to mention that the memory methods which one has created (by a created method I mean of course a modified version of already existing methods) him/herself are usually the best ones (for this certain person).

I quite agree. I still don't understand how Tim Habermaas's strange memory method (using images for single letters instead of pairs) could possibly get good results. But considering his 24/24 in just over 2 hours, I'd say it must be pretty close to perfect for him - I hope someday I come up with a system that works as well.

Strange? I completely disagree. It's the most obvious memory system with images. And it's extremely simple, too. One image for one sticker. That's it. I don't see anything strange in it. And it's probably not the best system for me, but i love simplicity more than anything else, so i won't switch.

 

[Mike Hughey]

Strange? I completely disagree. It's the most obvious memory system with images. And it's extremely simple, too. One image for one sticker. That's it. I don't see anything strange in it. And it's probably not the best system for me, but i love simplicity more than anything else, so i won't switch.

I knew you'd say that.

And how can anyone argue with it, considering your results?