I wonder if there's a good reason to limit the max bet beyond just limiting the maximum profit per bet. The Kelly criterion tells me no, but when I see it happening in three different places maybe there's a reason for it.
Hmmmm. I need to run the maths, as I'm don't have an intuitive sense of if it would matter or not -- as my house odds are a bit different than a normal gambling site:
https://www.moneypot.com/faq#odds (I try give conservative gamblers a much fairer deal).
However, I don't think the typical large-gambler would play like that. If they do, that's fine -- they can run into my 10% bankroll loss rule. But for normal cases, the 1 BTC / game to be the limiting factor. If a whale came in my casino, doing hundreds of normal style 1 BTC bets -- I'd be ecstatic. If a whale came in, and wanted to make a big win or big loss and then walk ... I'd be shitting bricks.
Create your chain of a billion sha256 hashes, publish the billionth hash, then (and only then) have someone trusted pick a client seed which will be combined with each of your billion hashes in turn to pick the outcome of each game. That basically makes you commit to your hash chain before you are able to tell whether it is a "good" or "bad" one.
True, that's quite nice. Once I tell everyone the process, I'll could publish the last hash in the blockchain -- and use the block hash as the "client seed". If you don't mind, I'll send you all the details before I push the button and you can do a final review?
I'm planning on using this process for my next game on moneypot -- which is pretty similar in requirements (real time, social game)