Author

Topic: A deck of 52 cards vs. a 6-sided die (Read 772 times)

legendary
Activity: 1792
Merit: 1087
August 04, 2015, 04:57:40 AM
#4
Anyone know of the perfect die?

Search "casino dice" on google
legendary
Activity: 1764
Merit: 1002
August 03, 2015, 03:40:30 PM
#3
Anyone know of the perfect die?
legendary
Activity: 2126
Merit: 1001
July 31, 2015, 05:28:39 PM
#2
Yes, I'd agree to your findings.
A die could be cheaply manufactured and be biased.

Also, depending on what you want to create, you won't need 256 bits (can't remember what made use of effectively 192bits only, I think?).

Either way, anything remotely like those are going to be fine. I wouldn't trust the pseudo-RNG of some tiny embedded device (raspi included), but other than that and faulty RNGs, this doesn't seem to be a plausible attack vector.
Not until every single brainwallet was emptied, more than half of all computers are infected with bitcoin-targeting malware, and the first hacks of "offline" wallets are reported :-)

Still, never wrong to err on the save side here.

Raalok
pf
full member
Activity: 176
Merit: 105
July 30, 2015, 10:16:31 AM
#1
I'm planning on using a 6-sided die to create entropy. That's 100 throws to get a 256-bit entropy.

But many, including goatpig (Armory developer), advocate a deck of 52 cards instead.

Can someone elaborate on this? Why is this a better strategy?

In fact, what's the strategy? Shuffle the 52 cards in an opaque plastic bag, draw one, put it back, and repeat 45 times (for 256-bit entropy)? Is that it?

I guess I just don't see the benefits because, how random can a human shuffling a plastic bag of rather big cards that get in the way of each other in the bag really be? At least a die is spinning like crazy in the air (and banging against walls and floors and ceilings) without human interference after the human has tossed it out of his hands. But then on the other hand, a die can be biased, so that's a drawback of course. Hmm.
Jump to: