[Open Source] Coin Flipped Seed (coin flip, dice roll, rubik's cube mixing)

[BlackHatCoiner]

With 20 moves being enough to reach any configuration of the cube, I think my suggestion which would include 60 moves is enough to ensure sufficient entropy. Still, if you are concerned about the amount of entropy then you can just continue to scramble for a few minutes as you suggest, and even better to do it without looking at the cube/with your eyes closed.

Yes, but as you can imagine, it doesn't differ from the mouse movement way at all. It doesn't provide anything to the users, besides that cool way to derive their addresses. It is actually worst than the mouse movement, because it takes much longer than that. Anyway, I don't think that it belongs on this software, I think that it'd be better to keep it simple (and fully unpredictable).

Nonetheless, it'd be interesting to implement it in the future, since no one hasn't done it yet. *cough-cough*

[1715091703]

1715091703

[1715091703]

1715091703

[BlackHatCoiner]

Major update:

Added rubik's cube mixing option and custom number of derived addresses.

As you can see blue, white, yellow and red gives us 2 bits while orange and green only one. So the worst case scenario (for time) would be having only green and orange blocks which would give a total of 9 bits for each side of the cube. On the greatest scenario, having no greens and oranges will return 18 bits.

So if I'm not mistaken every side will return 18 - (1/3)*18 = 12 bits 15 bits (9*1.666) which means 8.5 scramblings on average.

Release (binaries): CoinFlippedSeed-v0.3.zip (1.18MB)
SHA-1: 4DA93F3D72A9EB65282650E15D4E3C288A28FD71

Code:

-----BEGIN SIGNED MESSAGE-----
I, BlackHatCoiner, publish the v0.3 of CoinFlippedSeed in 21st of May 2021.
SHA-1: 4DA93F3D72A9EB65282650E15D4E3C288A28FD71
-----BEGIN SIGNATURE-----
advance concert visit awesome neglect fire dizzy club deny danger disease sign rebel donkey tone educate dumb desert mosquito happy crane jungle grit near
HxJcCeyv9zVxoc22OmAiHJiRhBzzpiYNPfZY6KzIRZjyVONGmZwehAZpaae4UzHaY8tLQwxGh+yfxdKG8QJcbR4=
-----END SIGNED MESSAGE-----

[BlackHatCoiner]

Hi, do you mind telling me why different color have different bits? From what i know and this rubik page, each color have equal distribution.

For the same reason you can't have 3 bits on each result of a six-side dice. The dice can only return you 6 different values, while 3 bits have 9 different combinations. Same happens for the rubik's cube.

(If I confused anyone, once you submit your cube's result, you mix it again. I just chose the same result to put into my entropy because of the GIF size)


Also forgot to write, future changes:

• Deriving addresses for altcoins and testnet.
• Option for electrum seed instead of BIP39.
• Maybe deck. Not sure how well can this work in practice. Do you have any other random related items on your mind to work on?

[o_e_l_e_o]

So if I'm not mistaken every side will return 18 - (1/3)*18 = 12 bits on average.

There's a 2/3 chance of 2 bits of entropy and a 1/3 chance of 1 bit of entropy, meaning an average of 1.666... bits of entropy per square. 9*1.666 = 15 bits.

Having said that, I still wouldn't recommend this method for the reasons I discussed above. Scrambling the cube manually is not going to produce the same kind of entropy as rolling a dice or flipping a coin since you are consciously deciding every move to make as well as which face to read from. Further, each square you enter decreases the entropy of all subsequent squares you enter until you scramble the cube again.

It's a cool idea to play around with, but you shouldn't use this to generate a serious wallet.

[o_e_l_e_o]

What if you use physical rubik and mix it few times? Would it still produce less entropy?

That is what I was referring to. There are only 4 red corners on a Rubik's cube, for example. If you use one of them on the first corner, then you have decreased the chances of the other three corners being red and therefore reduced the entropy of these squares.

If someone was to use the tool without a physical cube and just click on a bunch of squares to change their color, then that would likely be even worse, similar to how we tell people not to manually pick words to make a seed phrase, regardless of how "random" they think they are being.

[BlackHatCoiner]

There's a 2/3 chance of 2 bits of entropy and a 1/3 chance of 1 bit of entropy, meaning an average of 1.666... bits of entropy per square. 9*1.666 = 15 bits.

Yep. I'll have to stop announcing releases during the night.  

This means 128/15 ~= 8.5 scramblings on average. I'll change it on my post.

Having said that, I still wouldn't recommend this method for the reasons I discussed above. Scrambling the cube manually is not going to produce the same kind of entropy as rolling a dice or flipping a coin since you are consciously deciding every move to make as well as which face to read from.

It is surely not recommended, since the result isn't completely unpredictable, as you said the user knows what he's doin'. I did it only for fun and I discourage anyone from using it for anything above a million satoshis. The dice is, by far, the greatest option. On a future release I'll warn the user about that.

P.S. i just noticed you haven't update download link for newest version on first page, you might want to change it.

Thanks. I had totally forgotten it.

Further, each square you enter decreases the entropy of all subsequent squares you enter until you scramble the cube again.

As long as you scramble the cube each time you submit a side, I don't get how that applies. Wouldn't that be true only if you mix it once and submitting the 6 sides as they were?

[o_e_l_e_o]

As long as you scramble the cube each time you submit a side, I don't get how that applies. Wouldn't that be true only if you mix it once and submitting the 6 sides as they were?

No.

There are 4 corners in each color. 4 red, 4 orange, 4 blue, 4 green, 4 yellow, 4 white. Your first corner will be one of these, a 4 in 24 chance of each color. By the time you get to the second corner, you have used up one corner block. Let's say the first corner was red, but that individual corner block will also have two other colors attached to it. Let's say it is the red/blue/white block. Your next corner has a 4 in 21 chance of being orange, green, or yellow, but only a 3 in 21 chance of being red, blue, or white. Let's say your next corner is blue, using the blue/red/yellow block. Now you've used 2 reds, 2 blues, 1 white, and 1 yellow. So for corner three you are down to 2 in 18 for red and blue, but twice as likely (4 in 18) for orange and green, with white and yellow being 3 in 18. You are no longer being random.

The same principle holds true for the four non-corner squares you are reading, albeit with different numbers since each side square is only attached to one other color.