I love this thread. Someone should do a short movie about this 

If you had a million PCs (each with a similar GPU to the one you were using on your single PC) then you could divide your hashes per second per PC by a million.  Then, instead of 9 million billion billion billion billion hashes per second per PC, you'd "only" need to do 9 billion billion billion billion hashes per second per PC to find the secret in 23 hours.  And those hashes are now sha512 hashes, not sha1 hashes.  I expect they take longer, but I don't know.


Theoretical answer: "no". 

But what makes less sense than writing a general dice-botting program is to keep writing new ones each time a new dice site appears. 

It would be cool if there was a betting script where you could chose the betting logic (martingale, random, reverse martingale, whatever) and the site (coinroll, satoshidice, primedice, etc) at run time.

If each betting strategy and each site was separated out into separate functions then it would be relatively easy to add new sites, and new strategies, without having to reinvent the wheel each time.

The only thing that puts me off is that running a bot against sites with negative expectation is kind of a silly thing to do, and writing such a bot could make it look as if you were encouraging people to partake in such silly activities.

Hey,

We highly doubt anyone was cheating, we made the switch to Sha512 as it was very simple to implement and more practical in the long run.


Also we've restored the maximum bet payout to 5BTC

-Stunna

I'm aware of that blog post, but it's entirely unrelated to primedice.  That post is about how some Chinese researchers allegedly discovered a method that makes it 2048 times easier than was previously thought to find a collision in sha1 outputs.  But we don't care about collisions in primedice.  All we care about is how hard it is to reverse the hash and find the daily secret.  The Chinese discovery doesn't help in that regard at all.

Even if the blog was relevant, and the findings it referred to make it 2048 times easier to find the primedice daily secret, just replace the '9 million' with a '4 thousand', but leave the 'billion billion billion billion' in there.  It's still too big a number to comtemplate. Even if you convince every Bitcoin miner to switch from mining to trying to steal Stunna's Bitcoin stash (and good luck getting the ASIC miners to figure out how to switch from sha256 to sha1...)

I think the natural conclusion here is: nobody was cheating; it's just too hard.

I voted for 0.5% because online gambling operations, especially ones done "on the chain", incur little cost for the operator. The only operational cost for a satoshidice clone is a cheap VPS running a bitcoin node. You can probably even run it on your home computer.