Simply put, shares just before the block is found are the only ones that are relevant, to discourage pool hopping.
The function is an exponential weighting, so the effect is extremely strong: if you mine many shares, but disconnect just before the block is found, your payout will reduce drastically.
https://bt.irlbtc.com/view/1976.msg50002#msg50002I'll try to explain it with an example ...
I (= the pool) have a huge field (= round) in which a treasure (= bitcoin block) is buried, but no one knows where (=when the round will end), so I need to dig until I find it.
I need diggers (=miners) to help me find that treasure and will divide it between them instead of paying fixed salary, so they will get more for their work, but ...
I have paid for that field (=pool infrastructure and support), so I will keep 2% of the treasure for me (=pool fee)
I don't have an excavator, but it's faster with it.
There is a risk that the some diggers will come for a day or two, with the hope that they will find the treasure fast and leave before we find the treasure.
There is a risk that the some diggers will stay on the field pretending to work
So ... I need a strategy to find such with:
own excavators if possible, but even hand digging will do + to keep them with me until the treasure is found + not pay to those who are not actually working + divide it fair and quick, so we can move on the next field
To motivate those with excavators - I will need to pay based on the cubic meters of soil excavated per hour (=pool diff)
--- and now Slush specific strategy ---
Those who are lazy or just pretend to work will not bring soil often, so they should get less (=exponentially reduced reward)
To reach the goal I will write down the cubic meters brought at some location on a sheet of paper, but won't keep all of them - just the last one to divide it as fast as possible, so because of that ...
When it's full (=score reset) I will start a new one with a short summary of the previous, based on the same rules as if the treasure was found.
When the treasure is found I will divide it based on the cubic meters of soil excavated and brought at the location in the recent hours as:
Those excavated during the last hour are counted in full
Those in the hour before - in half ... and so on each hour reducing the volume paid in half until it falls below the minimum piece of the treasure (= 0.00000001 BTC)
So if you stop working - you will loose your payment fairly quickly
Outcome:
+
If you have an excavator you will bring more and will be paid more, but if you are lazy you will bring it less often, then you will be paid less ... at the end you will be paid fairly for your work
If you have no excavator, but hardworking - you bring less, but regularly ... you will be paid fair for your hard work
-
If you have an excavator, but it got broken (= internet connection down or miner down) and while fixing it, you missed the last hour of when the treasure was found ... you are paid less unfortunately
If you have no excavator, but hardworking - you may still miss bringing soil in the last hour, because you need 1.5h per cubic meter (=low hasrate), but still be paid for all your work during the hours before ... you are paid less, because you are slow even it is not because you are lazy
Those who complain are the ones affected from the negatives of the strategy, but it's not mine (=the pool's) fault that your excavator got broken or you are too slow to keep up with the others.
You are free to dig for someone else (without complaints) if your excavator brakes too often or you are too slow compared to other workers and miss the payment of every digging campain
