WE KNOW THE FACT THAT If a person wants to create a transaction hash that starts with (or ends with) a particular set of characters, then they can generate a transaction without broadcasting it and see if it starts (or ends) the way they want it to. If it doesn't, then they can discard the transaction, and create a new slightly different transaction to see if it starts (or ends) the way they want it to.
OK, that's all we need.
You just admitted that you know that your game isn't provably fair for players or for investors. All we need to do now is figure out how hard it would be for you to cheat.
The house edge is around 1 to 2% for baccarat, right? So if you make just two transactions and broadcast the one which gives you the bigger profit, you'll be doing much better than expected. I tried making a transaction that starts with 888 earlier. On average it will take 4096 attempts to do. I did it using bitcoind's RPC interface, calling it from a shell script and checking the output with command line tools like 'cut' and 'grep'. The first time I tried, it took 80 seconds to find a suitable transaction. I'm able to generate and test around 20 transactions per second on a single core of this cheap laptop.
Dooglus, you should realise that we send transaction every 2 mins via our doge wallet automaticly, no matter if there are any bets. and every draw result is coming out within 10 sec after the betting time ends. Lets say even we could 'choose' a transaction ID, but the time for us to repeat the process of above actions is far more inadequate. Our wallet sends a doge transaction within 5 seconds after the betting time closed, and the draw result comes out in the next 5 second.
So you 'only' have 5 seconds to cheat? I could generate 100 transactions in that time, which would allow me almost guarantee that any large bet on your site lost.
About your game, could you possible to send me the transaction ID begins with 888 within 5 seconds after we advise you to send?... If yes, we will have to reconsider our project.
I would be happy to play the game I proposed if I had 5 seconds to prepare my transaction, yes. I couldn't guarantee to win every time, but if we played a few times I'm sure I would win. I only need to generate 4 or 5 transactions in those 5 seconds to expect to break even and I think I could do 20 times better than that. I think my edge over you would be huge.
The important point is that the game wouldn't be at all fair, and I would be dishonest if I claimed it was. Especially after someone explained to me why it wasn't fair.
I guess what bothers me about this is your claim to be "next generation", "provably fair for investors", etc. when in fact you've taken a big step backwards. Whether you're actually cheating or not isn't that point (probably you're not). The point is that you could be and we wouldn't be able to tell. That's the problem that provable fairness solves, and your game doesn't solve.