Gaelium (GAEL) mining

[FP91G]

The project's Emission Schedule has a very interesting distribution of coins for miners.

After six years, the block reward will decrease by a factor of 10, and after eight years, by a factor of 20. How do you hope to retain miners in your project with such a decrease in rewards?

[nc50lc]

After six years, the block reward will decrease by a factor of 10, and after eight years, by a factor of 20. How do you hope to retain miners in your project with such a decrease in rewards?

Interesting, looks similar to Bitcoin's halving but at an extreme rate.
I've read the related parts of their whitepaper and it seems like it's capped to 5GAEL after 1B supply is mined.

Bitcoin's deflationary supply works since it's offset by the "coincidentally timed" bull run near every historic halvings
and of course, its increasing price over the years.

Maybe @gaelium is confident that the price of 5GAEL will be enough in about 12.5 years?

[gaelium]

https://pool.gaelium.io/site/mining
You have a Profit column with zeros. You just need to open the data for the last hour and 24 hours, and you don't need to create a calculator.

Fixed, the profit column is showing a value now.

The cause was simpler and worse than I expected. The pool had never been told which exchange to read, so the coin price sat at zero and multiplied the whole calculation down to nothing. The function that creates the market rows had been switched off back in February, when the pool was set up as single coin, and nobody put it back. It had been zero since day one.

There is a note under the table now explaining what the number actually is, since it was not obvious. mBTC per 1 MH/s per 24h.

The last hour and 24 hour columns you pointed at are still off. Most blocks in the current window were recorded with a zero price, so turning them on today would show a return five times lower than reality. Once the window rolls over with real prices in it, they go back on.

While digging into this we also found that solo shares were being counted into shared block splits, running since launch.

That one is fixed too, and the miners who lost out have been paid back.

[gaelium]

After six years, the block reward will decrease by a factor of 10, and after eight years, by a factor of 20. How do you hope to retain miners in your project with such a decrease in rewards?

Maybe @gaelium is confident that the price of 5GAEL will be enough in about 12.5 years?

Both fair questions, taking them together.

Small correction on the emission first. It does not stop at 1B. The reward settles at 5 GAEL per block and keeps going from there, permanently, so there is no max supply. That tail is deliberate. Ravencoin, which we forked, caps at 21B and once the last one is mined the only thing left for miners is transaction fees on whatever volume the chain has that day. Bitcoin has the same problem waiting for it, just further out. We would rather keep paying for security than find out how that goes.

On retaining miners, the reward size is not really what keeps a rig running. Difficulty is. If the reward drops and nothing else moves, some miners leave, difficulty falls, and the ones who stay find more blocks each. That balance is not something we control, it is the same mechanism every PoW chain lives on.

What we did choose is where the coins go. The emission is front loaded on purpose, the early years pay the most, and it is 1B spread over twelve years rather than everything minted at genesis for whoever got into a presale. There was no presale, no ICO and no investor allocation here.

As for whether 5 GAEL will be worth anything in twelve years, I am not going to pretend I know. Nobody does, and anyone telling you otherwise is selling something.

What I can tell you is what happens in the meantime. Wallets on Windows, macOS, Linux and mobile, our own pool, six other pools running the coin including Mining Dutch who added us yesterday, and the daemon patched within 48 hours when the Ravencoin consensus bug went public in August while several forks were still exposed. The code is a Ravencoin fork, which is a Bitcoin fork, and we adapt it where we need to rather than leaving it as it came.

That part we do control, so that is where the hours go.

[FP91G]

https://pool.gaelium.io/site/mining
You have a Profit column with zeros. You just need to open the data for the last hour and 24 hours, and you don't need to create a calculator.

Fixed, the profit column is showing a value now.

The cause was simpler and worse than I expected. The pool had never been told which exchange to read, so the coin price sat at zero and multiplied the whole calculation down to nothing. The function that creates the market rows had been switched off back in February, when the pool was set up as single coin, and nobody put it back. It had been zero since day one.

There is a note under the table now explaining what the number actually is, since it was not obvious. mBTC per 1 MH/s per 24h.

The last hour and 24 hour columns you pointed at are still off. Most blocks in the current window were recorded with a zero price, so turning them on today would show a return five times lower than reality. Once the window rolls over with real prices in it, they go back on.

While digging into this we also found that solo shares were being counted into shared block splits, running since launch.

That one is fixed too, and the miners who lost out have been paid back.

I racked my brain trying to figure it out:
"Estimated earnings in mBTC for one MH/s over 24 hours at the current network difficulty, block reward and market price, so multiply by your own speed in MH/s and divide by 1000 to read it in BTC"
https://pool.gaelium.io/site/mining
But I did, and I got 8 cents in revenue from 10 megahash.

You don't pay in Bitcoin, so maybe it's easier to write down how much 1 megahash yields in dollars in 24 hours so miners don't break their calculators.

[gaelium]

That is fair, and the fact you had to do the arithmetic yourself means the column is not doing its job 

The reason it is in BTC is that the pool software computes everything in BTC internally, so the column is just showing what it already has. But you are right that nobody mines here to get paid in BTC.

I will look at what it takes to show it in USD per MH/s per day instead, or alongside. There is a second problem to solve at the same time, the value is currently rounded to five decimals which leaves one significant digit, so it is coarse even before the unit question.

Thanks for actually running the numbers, that is more useful than the report itself 

[nc50lc]

On retaining miners, the reward size is not really what keeps a rig running. Difficulty is. If the reward drops and nothing else moves, some miners leave, difficulty falls, and the ones who stay find more blocks each. That balance is not something we control, it is the same mechanism every PoW chain lives on.

Good point, this is even used as an argument in Bitcoin difficulty adjustment's link to miners' revenue.

I haven't delved deeper into Gaelium's protocol but is there something that can defend it against 51% attack if the network's hashrate becomes too low that it'll be cheaper to attack the network rather than to aim for one or more 5GAEL block reward?
Considering the circulating supply at that point, I'm not expecting big numbers in 12.5 years to cover for it.
But yeah, it's still quite far in the future to confidently guess.

BTW, this is an interesting project to follow because its fast paced but similar to Bitcoin mining's future that we'll not experience in our generation.
(bookmarked)

[FP91G]

That is fair, and the fact you had to do the arithmetic yourself means the column is not doing its job 

The reason it is in BTC is that the pool software computes everything in BTC internally, so the column is just showing what it already has. But you are right that nobody mines here to get paid in BTC.

I will look at what it takes to show it in USD per MH/s per day instead, or alongside. There is a second problem to solve at the same time, the value is currently rounded to five decimals which leaves one significant digit, so it is coarse even before the unit question.

Thanks for actually running the numbers, that is more useful than the report itself 

Why create data per 1 megahash? A simple Radeon RX 480 or 590 graphics card mines around 10 megahash, so it's easier to record 24-hour revenue per 10 megahash in dollars.
I don't think miners with weaker graphics cards will go into mining; it's unprofitable.

[gaelium]

is there something that can defend it against 51% attack if the network's hashrate becomes too low that it'll be cheaper to attack the network rather than to aim for one or more 5GAEL block reward?

No special defence. We inherit Ravencoin's DGW difficulty adjustment and nothing more. At 12 years out, any 51% mitigation we designed today would be based on assumptions about a network we cannot see yet.

What I would say is that the tail is the wrong thing to look at on its own. Attack cost scales with the hashrate defending the chain, and hashrate follows total miner revenue, which is the reward plus fees plus whatever the coin is worth. A 5 GAEL reward on a chain nobody uses is indefensible. The same reward on a chain with real transaction volume is a different picture. That is the part we can actually work on, and it is what we are doing now rather than in 2038.

The uncomfortable version is that this is Bitcoin's problem too, just closer for us. Nobody has solved it. We at least keep a permanent reward instead of going to fees only.

Why create data per 1 megahash? A simple Radeon RX 480 or 590 graphics card mines around 10 megahash, so it's easier to record 24-hour revenue per 10 megahash in dollars.

Per 1 MH is not our choice, it is what the pool software computes and we did not want to invent a number on top of it. But you are right that it is not what a miner wants to read.

We are working on it. And separately, we are renting cards one by one and measuring them on GAEL with the same protocol each time, stock settings, 30 minutes, hashrate read from the pool.
That table goes on the mining page this week, so people can look up their own card instead of doing arithmetic.