Quote from: hozer on January 09, 2014, 12:33:05 AM

Quote from: etblvu1 on January 08, 2014, 11:56:16 PM

Quote from: hozer on January 08, 2014, 11:38:52 PM

What do you predict if we retarget every block based on the last 36 block average?

With rational-only economic calculations, I think coinhoppers will still have an advantage, but since it's a gradual change, rather than a step change, I think the psychology and market dynamics will be dramatically different.

In a sense, it's hard to apply the model directly to this proposed difficulty adjustment formula because the model I am using presupposes some kind of predictable difficulty over time - but apart from that, and in a casual application of the model and observing the coin markets - I would predict the coin would have a very randomly looking profitability graph, and it would land at random times, for a couple of minutes at a time, at the top of profitability charts - and get random spikes of major hashing power - and long-term miners would see on-going profitability of less than LTC, while sophisticated coin hoppers find ways to monitor the blockchain and know exactly when to jump in and snipe the few coins that can be had before the averaging formula move the difficulty out of rage of profitability again. There are coins like Galaxycoin exhibiting this kind of wild fluctuation in difficulty/profitability, and the reputation of that coin was not helped by being known for this type of fluctuation. I do not see this as an immediately catastrophic change like, like the suggestion to cap difficulty level changes to something narrow (which has a good chance of killing the coin in the aftermath of the first significant spike in the value of the coin in the markets) - but I do see it can drain 1) the reputation of the coin from the fluctuation and 2) enthusiasm for the coin among miners who see consistent profitability below LTC. These may then prevent the coin from gaining momentum to see any significant value spikes to begin with - and would tend to linger long-term as a low value coin without much hope of increasing in value. But like I said, this one is hard to predict because it does not fit the model I am using, and this is partly conjecture on my part. I apologize if I misrepresented your formula or intent in any way.

Etblvu1

I intend to take the coin-hoppers money by being better at math and mining. We'll see if this proves true or not.

Step 1: release 1-block retarget with 36 or 72 block averaging. So just plan on 2 more hardforks, the first for one-block retarget, the second for tweaking the averaging time
Step 2: get a consensus on https://bitbucket.org/dahozer/catcoin/wiki/Stake-mine which should clobber coinhoppers, AND make Catcoin unprofitable for scrypt ASICs. Deploy on a 1 or 2 month update window. (hardfork number 4)
Step 3: buy a big lot of powercolor LCS cards and make cat warmers. Sell a few to coinhoppers.

I had a similar idea which might work temporarily, but these kinds of ideas change the character of the coin - but the idea is to make a hybrid coin that requires solving the block solution using both SHA256 and Scrypt (where the solutions match up on both methods, which means the difficulty level would appear to be very low, but having a coincidence of a solution between both would be rare) - then anyone who wants to mine the coin profitably would need to add something like a Block Explorer ASIC to their computer, and anyone mining GPU only would inherently stay away from mining the coin. And unless other coins copy the exact same hybrid formula, the coin would inherently not be profitable for any coin hopper to mine. But people are very motivated to seek profits - so any static solution that does not specifically address the root cause will eventually be defeated by dynamic human creativity seeking clever way to defeat the static solution. Similar dynamics are at work when you see how iPhones keep getting jailbroken, or how hackers keep finding workarounds to security fixes by software vendors. Except with coins, we don't get to keep releasing new hard forks before exchanges and users get frustrated and drop the coin. That is why I favor permanently fixing the incentive over any temporary workarounds to incentives that remain active.

My acid test for any modification to a coin codebase is this - would you feel comfortable applying this same change to the Bitcoin code base itself, and bet the farm that it would not mess up an already successful coin? I am willing to apply this acid test to my own suggested changes - since my changes are designed to kick in only under circumstances that Bitcoin essentially never sees. If you were to change Bitcoin so its difficulty graphs became erratic - or changed it so that difficulty cannot adjust very quickly - I would expect that coin to do worse than it is doing right now. And if my estimation is true, we should not be applying that change to Catcoins either. On the other hand, Bitcoins and Litecoins have both done fine with very long retarget times - they have suffered very little effect from coin hoppers - so we can emulate that success by implementing long retarget times together with eraditation of coin hopping by eliminating from the code the incentive to coin hop. And that code can be back-ported to LTC and BTC fine without affecting the way they work - because they already don't have coin hopper problems so the back-ported modification would not affect their operations one iota. I challenge anyone to name a more conservative possible coin modification (so little chance of messing up Bitcoin if backported), with so much potential to improve the value of the coin (introduce unprecedented stability in difficulty for a minor altcoin).

Quote from: LeoC on January 09, 2014, 12:50:23 AM

etblvu1 - K, now go code it and come back with the results.

Despite the fact I've said I do not have familiarity enough to immediately code C++ in Linux, this is nonetheless a fair statement. If I cannot persuade by arguments or payment to cause the coding changes to happen, then no matter how valid the case for implementing the change, it will not see implementation. So if anyone has the skills and capabilities to make changes to the Catcoin code base and open to discussing with me 1) a simplified specifications of the changes I am proposing that should not be too hard to code, and/or 2) how many CATs you would charge me to implement such changes, and/or 3) willing to help me get up to speed with the programming environment (I have programmed in 20 different programming languages, including some C++, but in a Windows environment) - please PM me with how you can help/what you need.

Quote from: kuroman on January 09, 2014, 12:59:26 AM

-The loyalty system, although it's a good idea, but I believe it won't work simply because it creat more hassle to miners, that would want to join or that are mining but when the difficulty rise beyond a certain level (that they are losing money on electricity and hardware rather than making some) will just leave, not to mention that there is already the pools PPLNS penalty.
But the idea is good, so I suggest if this idea to be implemented, we should take it the other way, instead of punishing people that leaves you reward people that stay mining, this can be implemented pool side, lets people who mine for a week without non stop (a small intereption is acceptable should get a reward of extra cats depending on the hashrate, not this has to be discussed with pool owner and concensus should be made it) lets just take the example from commerce, this would be like the fidelity card for a supermarket, the more thing you buy from that supermarket the more reward you get. but again might be too much hasle for pool owners so overall, I think the idea is good, but in terms if feasability and implementation it is not. so tl,dr Good Idea not going to work

Thank you for this thoughtful response to my suggestions. I submit that since the logic would kick in only rarely (right after the difficulty dropped more than 50%) and otherwise the coin operates exactly the same way, it would only prove to be a hassle for people engaged in sophisticated profiting-maximizing strategies of coin hopping. If coin hopping people find the coin too much hassle to mine and stay away from the coin, I submit this would improve the coin rather than damage the coin. People who do not coin-hop to try to maximize profit would virtually never notice this mechanism at work - except that they seem to see more profits during easy difficulty times than with other altcoins - which if anything should attract more miners who want to keep their rigs pointed to the coin. So I ask you this question directly - if we increased the number of steady miners, and totally eliminated coin-hop-profit-maxmizing miners, do you really believe it would make the coin worse off vs. better off? And which type of miners do you believe dominate in the BTC and LTC mining scene, and how does this relate to the perceived stability of of those coins?

Etblvu1