Because only one distinct CPID with payments and magnitude can actually solve a block, therefore rewarding one miner a disproportionate share would be unfair for everyone else. If anyone could always solve a block it would be fair to let the floodgates swing wide open, but think about what is happening in this scenario:
User 1 has 100 machines - And pulls 90% of the block reward
User 2-11 have 1 machine each - and pull 1% of block reward each
Now in reality over 100 blocks, user 1 is only allowed to solve up to 13 of those blocks Max (because of the distinction rules), so he pulled in 9% of the reward yet we paid him 90%? LOL, not a good idea....
Hi Rob,
could you explain this part in more detail? I'm not sure if I understand it fully.
This part confuses me: "User 1 is only allowed to solve up to 13 of those blocks Max (because of the distinction rules)"
What would stop the user from solving more blocks (if he have a valid CPID with a magnitude)? Or is this a special rule of the main pool?
Sure!
So since we require a User (who has multiple machines) to have One cpid who has been Paid in a prior superblock, that means one pool user who has 100 mining machines has copied wallet.dat 100 times, and all wallets share one cpid.
However as mining round starts, the only cpid who can actually solve the block is one who has not solved a block in the last 5 blocks, who has a signed cpid, who has been paid in a superblock (with magnitude existing), so that means your power user is in effect stealing hash power from all the small dogs, if you let him (IE if you award him shares for everything he does) while he is really using a copied CPID over 100 machines.
Please see this function : HasThisCPIDSolvedPriorBlocks
That is the main contention point - if big dog A had solved any of the prior 5 then his hash contribution for the next 4 blocks is really worthless to the group, so hes then "taking" from the little dogs for a while.
So yeah lets talk about this in the PM and figure out how we can modify our service to reward each share based on an RPC check grounded in current conditions etc.