The median should be about 1.98 - yet Bustabit is running 1.96 - while it may not sound like much it is actually a HUGE difference.
The expected median should indeed be 1.98x, but the actual median is going to vary significantly over a finite amount of games.
Now, devans (aka Justin Malara) will say that it is proven and will spit out an equation or 2 that "proves" it is a 1% house edge.
That system is known as "provably fair" and contains absolutely everything you need to prove the game is fair. While i do admit it requires quite a bit of technical knowledge, it has the advantage that anyone can verify the game for everyone. So if you can find someone who understands how the maths primitives work, it should be easy to verify =)
When asked if he ran the entire 10 million numbers to see what the actual median is he stated no, it was unnecessary work.
If I was in his position, I would not do it either. If that knowledge affects your actions, I would say it's pretty unethical (as it'll either advantage/disadvantage investors/gamblers). If that knowledge doesn't affect your actions, it's kind of pointless to know. I would definitely not want to put myself in a position where I could get influenced by it.
BTW, Daniel actually makes more money when the games are good than when the games are bad. His profit is a function of the amount wagered, and players are able to turn over a lot more when they're not getting raped.
If I am advertising a 1% house edge and I am getting a 2% house edge, what would you call that? I call it cheating. A stupid ass equation does NOT change the house edge.
The house edge can be 1%, but players lose 2% .... but that doesn't mean there was cheating, it means there was variance. The provably fair system proves that the game outcomes have been decided from a probability distribution function that results in a 1% house edge. It's totally normal and expected that the games will be "good" and "bad" over a finite period of time
I have not done the math but i am pretty fucking sure that >500,000 games is a large enough sample that the median should be pretty spot on and not 1.96
I'd encourage you to do the maths, or get someone to do it for you. Variance is far more of a bitch than almost everyone realizes. In the end, bustabit has done all that is possible to allow you to verify the games and the maths and check things -- the ball is really in your court on this one.