How much stake is required by a pool such that they produce 1 block per epoch on average?
My guess is 25k to 250k Ada, but i might be wrong.
Currently about 3m ish. When the d parameter is changed to 0 later this month, it’ll then be in the 1m-2m range.
I know you asked about getting 1 on average, but just for more info, I have seen 150k get 1 block every once and awhile.
The minimum stake for producing one block per epoch is …
totalStake / blocksPerEpoch => minStake
There are 5 days in an epoch and one slot per second. That gives
5 x 24 x 3600 => 432000 slots per epoch. Only 5% of those slots are currently used to produce blocks.
432000 x 0.05 => 21600 blocks per epoch.
The total stake is currently at 22.92b. Therefore you need
22.920.000k / 21.600 => 1.061k
This would not give you a guarantee, but on an annual basis (i.e. 73 epochs) it can be said that 1m will give you 73 blocks p.a.
100k will give you 7.3 blocks p.a.
For every epoch that you don’t mint a block, your chances will (slightly) diminish to mint a block in the next epoch. This is because 21600 blocks were minted and got rewards, which will increase the total stake slightly (i.e. ca. 16m)
I think you mean 100k will give 7.3 blocks per year.
This is helpful though. Thank you. I did not consider total staked in the system. That’s interesting.
You can have a mental picture of a gigantic roulette wheel with 22.920.000.000 slots. For every ADA you hold you can place your bet on one of the slots. 21.600 slots will be winning, the others loose.You can win with a single ADA, but on the long run you will need more ADA to win consistently.
100k blocks will get you around a 10-12% chance of a block per epoch. It doesn’t necessarily translate to year since each epoch is 5 days. So in bout 7-8 epochs you’ll earn a block.
You’d need a little over 1M in total delegation for a good chance to mint a block every epoch. Not 100% because protocol still has some randomness in it but 1M-1.1M around that is where you need to be. In the 800K range is more or less a 50/50 chance.