Haven't watched this thread lately...
My bets were also seeming to attract bad luck streaks much more than the healthy mix of wins and losses,
and those few wins were eaten up by tx-fees for the much too soon payouts. As I always said, my pet-peeve
was not about a troll ending my session within five minutes, but instead the system itself taking the part of the
troll ending my sessions much too soon, and doing so *always*.
Well, personally I don't really see this issue. Maybe I have to see it in another way - forget about the sessions, but just gamble and collect your earnings in your account and cash out whenever you want. Will see if I can add something like this when the optional accounts are done. But it's not the top item on my Todo-list as it is conceptual quite different from the current implementation.
At some point I wondered, if the secrets could *theoretically* be chosen such, that they'd lead to a non-uniform
distribution of lucky-numbers when used with tx-hashes. Everyone seems to believe that because sha256-hashes
look like random noise, that they'd actually be random noise. The actual entropy may however be much smaller...
To my knowledge there is no way to influence the lucky-number distribution by chosing a specific secret. As it is combined with transactionID and output index I can only control part of the input, resulting in a total random lucky number. At least this is my understanding how the hashing works
If at some point I was able to defer payouts for a week and do "greater than Lucky#" bets, I might actually come
back for some more bets :-)
I'll let you know when this is possible :-)