Thank you for the suggestion, it is something along the lines of what I was planning to do. I was looking for a way to do this without having to store 70-80 million ticket hashes every round due to the load it would put on our servers, but it doesn't seem like there is another way.
The method is pretty good, bit it still doesn't solve the problem of me cheating by creating fake accounts and tickets for them so that the hashes of these tickets fall into the positions of the winning numbers. I doubt there is any way to make the lottery 100% provably fair due to this but this is still a much better implementation that what we currently have.
I'll add this to my to-do list and should be implemented in a month I hope.
Please pm me your freebitco.in user id for a bonus
Maybe increase the price of each ticket by 10 times (to 0.00075 USD worth)? The free tickets given away in free rolls could be changed to given away by chance (e.g. if rolled lower than 2000/greater than 8000 than give 1 ticket). Assuming 100 byte per ticket data, 7M tickets would be only 700MB in size.The method is pretty good, bit it still doesn't solve the problem of me cheating by creating fake accounts and tickets for them so that the hashes of these tickets fall into the positions of the winning numbers. I doubt there is any way to make the lottery 100% provably fair due to this but this is still a much better implementation that what we currently have.
I'll add this to my to-do list and should be implemented in a month I hope.
Please pm me your freebitco.in user id for a bonus
Or make the lottery blockchain-based: Order the tickets by some method (maybe the order of users acquired them, or by user ID), and make use of a block hash (use the last digits, the first digits is always 0) to determine the winners.
P.S. Why not make freedoge.co.in have these features too?
Its 70 to 80 mil tickets every month and one entry would be about 100 bytes so 7 GB. However, the problem is not with just storing the data - that is the simple part. Looking up ticket numbers by user id or sorting them to do the draw is what would take a huge amount of processing power so I'll need to isolate this from the main servers and set up a separate server just for doing this so that it does not impact the rest of the website.
We currently store a bet history of about 700 million bets at any time and for that I have had to set up a separate server, this would be something similar although smaller but the sorting would consume resources which is something we do not need to do with the bet history.
I don't think increasing the price would be viable at this time, it would reduce interest and also distribution for multiply and referrals would become complicated.