He has some other motives behind his accusations, that is clear.

Because we don't use Bustabit v1 code. The game on https://betking.io is our own.

Usually immediate, and I had immediate withdrawals (the one without the extra fee) in the next block, too. But never nothing which could drain the hot wallet which I think is >100 BTC.

When people talk about house edges, what they mean is the average amount a player would lose if they made the bet a lot of times. So let's say we flip a coin, and if you get it right you double your money -- otherwise you lose it. We'd say that has a 0 house edge, because if you played a lot of times the amount of wins would be pretty much the same as the amount loses.

But all bets in bustabit are slightly worse than that. So if you want to x2 your money, you'll only win 49.5% of the time. So if you played a lot of games, the amount you'd lose is 1% of the amount you bet.

So to answer your question, the 1% edge is taken because the distribution of busts was predetermined from a distribution such that that property holds. i.e. It only goes to >= 10x an average of 9.9% of the time. But if there was no edge at all, it'd be 10% of the time.

No. Daniel knows all future game results, so it could be undetectable abused to cheat investors.  It's probably not happening though, considering investors are ~400 BTC ahead of expected value. But no real way to be sure, you have to 100% trust Daniel, basically.

Bustadice however, has actually a pretty neat system you can look at. There the game results are created by a combination of: (server seed, auditor seed, client seed) so it would not be possible to undetectably cheat investors without the auditor noticing (or participating).