Post
Topic
Board Pools
Re: [1450 TH] BitMinter.com [1% PPLNS,Pays TxFees +MergedMining,Stratum,GBT,vardiff]
by
organofcorti
on 15/08/2014, 03:31:26 UTC
The mathematical probability of hitting 2 blocks in the 99th cumitive distribution factor is astronomical, roughly the same as hitting 2 at <1, not quite the same as the 99th is infinite since 100% can never be achieved.

I'm not really sure what you mean, but by definition the lower tail 0.99 CDF occurs when an event is expected to occur one time in 100.

If that's for one block, it will require total work of 4.60517 times the network difficulty. If it's for two blocks, it will require total work of 6.638352 times the network difficulty.

In the first case we would expect this to happen once in one hundred blocks; in the second case once in two hundred blocks.

That's not even close to infinite.

Actually it is multiplicative not additive.  Thus it is one in 10,000.  Which if you extrapilate out at the same difficulty means it should happen about once in 25 months.  I.E. hardly a once in the life of the universe event.

No, it's not. By definition, a CDF of 0.99 is a one in a hundred occurrence, whether the event involves one block or 10 blocks or a thousand blocks. 

We are not discussing one event.  We are discussing 2 events and now a third.  Yes a CDF of 0.99 is one in a hundred.  However the chance of 2 CDF 0.99 events occurring back to back is one in Ten Thousand.  The site does not calculate the CDF of multiple blocks.

So it's two events each with 0.99 CDF? Why not just say so then?

The cumulative distribution function is a percentage probability charted with an "S" curve. It charts the probability of finding an answer with x number of tries. So 99.33% means that there is a 99.33% chance that we should have found a block with that number of shares. There can never be a 100% chance since, as was mentioned, the chance always resets with each try. So when you reach 99% you can never reach 100% probability from there making the 99th percentile infinitely long as it can go on forever if an answer is never reached.

http://en.wikipedia.org/wiki/Cumulative_distribution_function

Why are you repeating me? Are you implying I didnt explain all that clearly? You may be correct.