Oddbean

the purpose of having proof of work is to act as a rate limiter so there has to be a difficulty adjustment, and the reason why you might think it won't be necessary is because you aren't thinking about malicious users or erroneous software

it's for maintaining a given rate of solution (hash with enough zeroes) discovery within a specified time period btw, "zeroes" is not really the best way to implement targets, better is to use a large integer library and compare the values as scalars bitcoin has 32 bit precision for compact representation in blocks, you can have full 256 bit targets in a nostr based system since it's not an immutable long living chain... the difficulty adjustment formula will produce a proportion, a fraction ie 0.8... or 1.1... and you just multiply the previous target by this value to get the new target i'm not really sure that proof of work is fit to the task though i think that lightning and marketplaces to discover prices and enable users to trade artifacts and adjust prices according to demand so you don't screw up supply of artifacts in my opinion, proof of work belongs in only one place in the whole world: bitcoin chain everywhere else there is all kinds of game theory problems with maintaining security

i acquired it partly while working on a bitcoin fork that was started in late 2014 called parallelcoin... it had a hard limiter percentage on the difficulty adjustment and this caused it to become time-dilated from a 5 minute target to 12.5 minutes circa 2020 i made an elaborate, multi-schedule multi-block type difficulty adjustment that had about 50% variance in my tests, ie, it usually got 2.5-10 minutes per block since that time i learned more and PID controllers are very relevant, as are restricting allowable timestamps and making the time divergence quite tight, like no more than half of the target interval PID controllers note that you need to devise a low pass filter that ignores the upper half of the frequency of the derivative factor or you end up disrupting its accuracy

i have considered it and my opinion is that if the game artifacts are worth money the money chasers will buy beg and steal hashpower to outcompete the players it has no place in either nostr or a game system, because it has historically, after some 25 years, proven to only be secure in the bitcoin use case and everywhere else it's a mess

it is relevant to the economy of in game artifacts but i think Lightning could do a much better job by making it concrete value and following the same kind of probabalistic pattern used by gambling computer games (pokies) if you are talking about highly time constrained, limited participant, competitive use cases for proof of work then you will see that the most determined players will equip themselves with higher powered hashing devices and dominate games by this, it will lose the chance aspect and that's gonna ruin RPGs at least none of that will take long to manifest, so again there is no use for proof of work outside of the bitcoin consensus every other possible use case that might consider it is better served by simpler probabilistic distribution algorithms, and weightings based on gamer's financial inputs in the end that's always going to be how it plays out so you need to consider that it doesn't level the playing field like you wish it would

I'm just not convinced that we know with certainty that POW has no other uses. What if bitcoin is simply the first — not the first AND last? ancient Chinese alchemists: there is no use for gunpowder outside medicine

I think it’s a fun idea to use it as a roll system. Imagine for the game to accept the item the nonce tag has to be less than a certain value. Then you would be guaranteed that a high pow roll was really lucky.