Quote from: zerodrama on February 03, 2014, 11:27:11 PM

It is 12%. The thing is it's symmetrical. It's a multiplication. Not an addition. To go up 5/4 you multiply by 1.25x. You don't add 25%. To go back down you multiply by 4/5 or 0.8x. 0.8 * 1.25 = 1. 112/100 * 100/112 = 1.

That's my understanding of this but someone told me it is not and he presented different numbers.

Quote from: zerodrama on February 03, 2014, 11:27:11 PM

It's not a clear formula. It's a copy pasted formula. SMA36LIM12 is specific. Why 0.7? Why 144? Why -1.228? Why power function?

NEVERMIND the possibility of a broken Floating Point Unit!

The number works? ever heard of testing process with different numbers? you are the one that's been promoting this from the begining, testnet, different numbers 12% (why 12% why not 6% why not 20% it's the same process) same goes for 36

Quote from: zerodrama on February 03, 2014, 11:27:11 PM

COPIED AND PASTED. No thinking used. No one knows WHY it works.

Why do they bother copying and pasting it? Because it WORKS

Quote from: zerodrama on February 03, 2014, 11:27:11 PM

I don't CARE if developers innovate. Developers are creating a ton of coins already. You seem to think, like A LOT of COIN coders, that ONLY developers should contribute, the rest of the COMMUNITY should just debate, hype, freak out, and vote on polls. How DARE they contribute code! I want the code to be easy enough for even a freshman in college to understand HOW TO IMPROVE IT. That is the strength of the coin. That everyone knows what it does.

Aren't contradicting your self? Am I a developper? no NO and NO it's the total opposite I can writte you a white paper of only math code, but do you see me doing that here? no so.

Quote from: SlimePuppy on February 03, 2014, 11:31:18 PM

Simply put: You did post code. It might make sense to the 7 or 8 people here that are fans of Kernighan and Ritchie (the authors/creators of the C programming language) - but that means you're not communicating meaningfully to >97% of the community. I'm fairly sure that's not the result you're hoping for. I'm trying to help you get your excellent analysis into the hands of a greater portion of the community. I hope you will do just that.

You are not putting into context the reason why I posted the code, I posted the code to answer the person who claimed it was overly complicated and impossible to implement hence claiming he already read the code and understand it, which is not the case, You don't see me posting code in every single comment do you ?

Quote from: zerodrama on February 04, 2014, 01:35:45 AM

Nope. I've edited linux modules to allow servers to be hidden behind invisible routers. I can do C/C++, PHP, Javascript, Lisp, and a few others.
This is about the rest of the community not understanding the code. You keep thinking only coders should understand the code.
Not on my watch.

As for examples and graphs, I've been in the trenches trying to help pull this code out of obscurity and into the light. All I've seen as far as Kimoto is the code snippet you posted. I haven't seen anything that talks about what EventHorizon and PastMass mean. If the COMMUNITY has no idea what they mean, that's no good. It doesn't matter if you or I know what it means. If the users don't know, they can't contribute. It separates them from the developers. That's no good for a coin's survival.

You've edited and can't understand a simple lines of code like that, I don't code C, I only used C / C++ and Fortran to solve, Mechanical problems, And I only use alghorithms for Mechanical and CFD Simulation on software like Ansys or Fluent ect. So to resume I have a basic level, but with my basic level of knowledge, I can just go to the Gethub of any coin that uses YGW, open the main.cpp and see the parameters used and search for the definition instance like blockmass or evenhorizon (and btw this word are used as kimito seems to be interessted with general relativity, and Mass, Gravity well and evenhorizon are just the basics of spacetime fabric warp and stuff like that and the number used are inspired from some of general relativity constantes and obviously I'm not asking you to know this )

Quote from: zerodrama on February 04, 2014, 02:28:14 AM

kuroman, I'm not trying to be nasty, but I feel as if some people see this project as a piece of software with peculiar switches and knobs and they think that's where it ends.

Unfortunately, it's not. The poorly structured code is inhibiting evolution and innovation. We need something that doesn't smell like 1995.

Bad code? you just claim it's poorly structured and bad code while it is just simple C commands ! how can this be bad code, you are just in self denial here, the solution is proven to work end very well, it you have some factual evidence that the solution doesn't work perfectly please put them on the table, also I'm not saying we should apply the code as it is, we can adapt to our coin, it would be stupid to just apply like that.
we can change:
-BlocksTargetSpacing
-PastSecondsMin
-PastSecondsMax
And this something other coins did to adapt the code to their coin (you claim they just stupidly copied YGW, that's an insult to other dev, that know more than you and me about cryptocurrencies and coding in general)

Quote from: SlimePuppy on February 04, 2014, 05:47:20 AM

Just for grins, here are a couple of scattered thoughts in no apparent order. Hopefully the pictures will make up for the lack of organization. Wink

A number of folks here as well as a number of coin devs are enamored with the Kimoto Gravity Well (KGW). So far, the coins I've found that use this have fairly fast block times. The longest is Anoncoin's 3.42 minutes. Others, from Megacoin through Franco, have block times from 30 seconds through 2.5 minutes. Here's Megacoin with its 2.5 minute block time:

https://coinplorer.com/Charts/Difficulty/MEC

It's pretty clear that this method allows difficulty to track the hash rate fairly well. Let's look at another coin with a 2.5 minute block time - Phoenixcoin:

Let's skip forward in time for Phoenixcoin - here's a look at performance after the Oct, 2013 fork:

This fork is when the team there moved to a 2-tail simple moving average with 100 and 500 block SMA, a 1 block retarget and a limit. I don't know if this can be adjusted to work for CAT, but I hope it shows that KGW is not the only way to.

So - all you smart folks in CATlandia - thinking caps people - we need you! Wink

Pheonixcoin is a bad example, as it was only hited Once the top charts of profitability and it was hit only once or twice with profitability pool since fork, Also during december the coin was not tradable during several days on cryptsy.

I'm not saying there can't be other solutions, of course there are, but the process to get one that will work well for out case will take a huge amount of time, between having an idea/white paper, translating the idea into code, simulating the code, testing the code (and as proven with current code testing does not represent real life conditions sadly as it disregard some parameters ....) ect ect, it takes time and energy.

I'm all for innovating our own solution and let other coins copying our solution like they are doing with kimitos, but I think for now and in the meantime to get all this process done, we can use KGW adapted to our coins (it does not take much copy/paste to main.ccp adapt values and tada! (well might not be this simple as I've never coded a coin and I'm not an expert but I think it's not far from it) and we can have our sweat time testing the values choosen and tweaken, then release a version that solve this problem.

Almost every new scrypt altcoin, and previous coin that had issues with profitabilty pools uses KGW and it has solved their problem and this is for reason