N = 36*10+1;
It's "signals" arrays length (don't like any dynamical memory allocation anywhere).
Those code representing 36 blocks sliding window and i simply want to see period as long as 10 windows.
N can be any value (not 36 dividible may cause some problems in this code, but it's only quick overwiev)
You should mention about astatism of diff recalc.
It's relative, so algorithm don't know what difficulty need to set for hashrate, it even don't know exactly hashrate.
Algorithm operates on how need to multiply current difficulty to get target block time with assumption(maybe aprove through probability theory),
that block time reverse proportional to hashrate with same difficulty and most important - that blocktime proportional to
difficulty with same hashrate. And all of coefficients of proportions is unknown and more, random (with distribution law
depending on the hash distribution law). And i use very symplified model of most probable average time (think it's "Expected value" in english terms).
Take it from LTC calculator as i remember.
If you take in account astatism, and tests show that blocktime stabilise on 10 mins after hashrate change and hold both ways, i'm calm.
And yes, catcoin never been anonimous, but i'm about logo design. My "serious cat" not a cake too.
I'm explain. Actual logo on my opinion has expected standart advertisment design. If you think about
food or traveling company logo it's what first come to mind - light offce mainstream style. And "old" anime-style
(not fan of anime, only like "ghost in the shell" for it's cyberpunk setting and not very simple plot) kitten was what you
never expect from currency, somehow original(in this place) and rememberable. It was cute and not serious,
it was attracting, something which "CATcoin" name can be accosiated. Posche - you imagine porsche 911 design. Bmw - you remember
double front grill and sign. Microsoft - you remember windows logo. Bitcoin - BTC. It's about associations.
If those gray kitten has no name and story behind it it's the best logo. Still offer gray kitten on mars surface.
Hi strelok369,
From a mathematical perspective, what do you think about the idea of breaking up the tendency for harmonics to develop, by replacing just the "36" portion of "SMA36LIM12" with a prime number pseudo-randomly at each block retarget, to be selected from a table? For example, if we take the least significant 4 bits of the hash of the last block, and used it to look up a table of 16 prime numbers consisting of {7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67}. Why this sequence of prime numbers? It's the closest consecutive list of prime numbers that approaches the average of 36 that we are already using (average of this list of primes is 34.875 and median is 31 or 37, so responsiveness to sudden changes in hash power would roughly match what we have now at 36, actually slightly better). And why prime numbers? By using prime numbers, we eliminate any unintended/unforeseen harmonics that could happen in relation to sampling sizes. Even if there is no danger of harmonics in this sense, it seems good to spread out the numbers a bit, and using prime numbers seems as good a way as any to distribute the numbers, and gives us a nice variability range from 7 to 67.
Thank you for your thoughts.
Etblvu1
p.s. If this approach is discussed it might need a name so, I'll give it the name SMA~36LIM12 to indicate the sample size is
approximately 36, instead of exactly 36. And I guess "simple" moving average may not be the accurate name, but I don't know what else to call it.