Litecoin being orders of magnitude more secure than Bitcoins? Fact Check Please.

[twelph]

I would like to bring into the discussion a post on Reddit where some very strong claims were made discussing Litecoins being extensively more secure than Bitcoins. Could someone with more knowledge on this subject possibly fact check these calculations and give us a second opinion?

http://www.reddit.com/r/litecoin/comments/1cssqr/the_math_why_litecoin_is_more_secure_than_bitcoin/

[markm]

Isn't it mostly more secure at comparable hashing power but still influenced by overall hashing power?

For example if bitcoin is hashed by more than 50% of the hardware in the world capable of hashing it, whereas litecoin is only hashed by 1% of all the hardware in the world capable of hashing it, doesn't that give a 50:1or more  advantage to bitcoin that then litecoin in turn reduces somewhat by its faster blocks?

How many blocks does it take to counteract a 1:50 or more advantage? Or any other specific ratio of advantage?

ASICs will change this somewhat, as unlike in the GPU era, bitcoin will tend less and less to be able to turn its hashing power upon litecoin.

In the GPU era, litecoin had maybe nowhere near enough hashing power to attack bitcoin but bitcoin maybe had easily enough to attack litecoin?

-MarkM-

[paraipan]

Already explained:

You are misunderstanding the problem. We are trying to find the chance of getting a run of K or more successes (heads) in a row in N Bernoulli trials (coin flips). What are you solving here is the probability of a single (K=1) poison distribution. If I get 3 blocks per hour on average (50% hashing power) what are the chances that I will get 6? This is a very easy problem to solve but it's not the correct problem.

http://www.reddit.com/r/litecoin/comments/1bvgf4/did_switching_from_sha256d_to_scrypt_acheive_its/c9agapx



Btw, Scrypt is quite new hashing algorithm that was designed to solve some issue but could have created others, so user diligence is advised when dealing with systems that rely on it.

http://www.tarsnap.com/scrypt/scrypt.pdf

http://security.stackexchange.com/questions/26245/is-bcrypt-better-than-scrypt

[markm]

They are only calculating what percent hashing power gets what chance to do something nasty.

The real security though is, how impossibly is it for enough hashing power to exist that is attacking the chain rather than defending it for any given chance of causing harm?

For example, if pretty much any major miner could 51% your chain, it is not even really relevant whethr you are SHA256 or Scrypt, in either case you are screwed.

Whereas if it would take quite a few large miners all ganging up on you to 51% your chain, you are more secure.

So really the real security comes from what percent of the entire world's ability to hash that type of hashing if they chose to do so, not from whether it would be a little harder for them with less than 51% for one hash than with the other.

For random acts of mob mass lemming weird actions, the fact that any GPU anywhere could join an attack on you might make you less secure than if all the GPUs existing in the world all acting against you could not come near 51% of your hashing power.

-MarkM-