Glad to see that 2 accounts can forge all blocks. that seems balanced and fair.
NodeClub/RealSilk, if you wouldn't mind, please PM me the mathematics of the algo.
A while ago I was working on a similar but rather sophisticated project. The idea of PoA is novel but, IMHO, can be easy to manipulate. Of course, your take on the solution could have solved many of the problems that I discovered. Regardless, I think I could make a few suggestions that could improve the algo if I had a reasonable understanding of the process.
I understand that you guys are wanting to stay quiet on the details for fear of copycats but check my account-- I'm well invested in NODE, so it's success is quite important to me.
Thank you.
I sent You PM.
As I said earlier, according to our calculations possibility of manipulation will disappear automatically with the growth of the network and capitalization.
If You send 10,000 (fee is 35) and gain highest PoA weight then You can get a profit only when the next block will be 35+1, - respectively, next transaction should be 10000+1. But it can be 10, 100 or even 1, - so then You get less than You give. In this case Your unnecessary transactions will bring You nothing but losses
Now, on the initial stage of network forger's number is small and there is a high percentage of large transactions 490,000+ so You can get some profit sending money between own accounts, but the overall low level of capitalization will make it low. So I repeat again: in the future this possibility will disappear automatically with increasing of total number of transactions and a decreasing in the average transaction amount. Even now there is enough around 15-20 small & middle transactions per hour to overlap this possibility.
In the February release we try to make the PoA coefficient as open and transparent as possible.
