Sigh, my estimated earnings are still randomly decreasing and jumping about.
Is the score calculation still flawed?
I'll do my own calculations to determine how accurate it is.
Edit:
50(13535/441462)=1.533 BTC
My earnings are showing as 1.203 BTC

I am trying to iron-out the cheat-proof estimate. It's crazy logarithmic math. Payouts are also no longer 100% proportional. They are weighted towards miners who spend even time in the round. So, if you're not trying to pool-hop, you will be paid slightly more than those who are.
Even if your estimate is off: the final payout will be correct. The final payout calculations are too complex and strenuous to run often enough for estimates.
Could you explain the entire formula for calculating how we are paid? Then we can review it and fully understand how the reward system works.
Also, with the current method, doesn't that mean the lower-power contributers (below 500 MH/s) are unfairly being disadvantaged relative to the high-power hashers, so long as the high-power hashers are present (or accepting shares while the low-power hashers are too)?
No, hashrate has nothing to do with the scoring weight. It's weighted towards when you participate.
Think of it as proportional scoring but also proportional to the length of the round you participated in.
The actual proof and discussion is here if you want the technical details (
http://forum.bitcoin.org/index.php?topic=4787.0)
So if i understand correctly, a simple situation would be this:
If Person 2 joined the pool at when the weighting was 5.
And the number means how many shares have been submitted at that weighting (would decrease within a given time period).
Person 1 Person 2 Share Weighting (Decreases with time until block is solved) 1 10
1 9
1 8
1 6
1 3 5
1 3 4
1 3 3
1 3 2
1 3 1
Person 1 submitted 10 shares. Has a total weighting of 55.
Person 2 submitted 15 shares. Has a total weighting of 45.
This means that Person 1 minimises their losses (or even equals or gets more than) Person 2 as he started mining earlier in the pool, and so the weighting starts canceling out the fact that Person 2 has a higher share of the reward.