Quote from: dooglus on November 02, 2014, 03:29:33 PM

Quote from: Cyrax89721 on November 02, 2014, 02:35:16 PM

Quote from: dooglus on November 02, 2014, 08:09:14 AM

Quote from: kleineaap on November 02, 2014, 01:40:05 AM

Second, this is a max profit flat betting (sorry if I'm not correct on the max bet, but I can't publicly find PD limits).

It's weird from the screenshot how 3x pays out differently whether you roll hi or lo. He's winning 24 BTC when he rolls low, but only 23.99988 when he rolls high. Is this some kind of rounding error?

The "Over" Chance is actually 33.01000000000001%. I might have misplaced the first and last '1' but that's apparently where "Over 66.99" sits as opposed to the straight 33% that 'Under 33.00" is.

To quote the provably fair page at PrimeDice:

Quote

The resulting number 0-999,999 is applied a modulus of 10^4, to obtain a roll number 0-9999, and divided by 10^2 to result a 0-99.99 number.

So there are 10,000 possible different rolls, each with the same probability. 00.00 through 99.99.

Of the 10,000 possible rolls, 3300 of them are under 33.00 (00.00 through 32.99 inclusive) and 3300 of them are over 66.99 (67.00 through 99.99 inclusive) so they both have a 33% chance of winning and should both pay out 3x (99/33) for a 1% house edge.

So the payout is wrong. It is being paid out as if the chance of winning was 33.00011%, which it isn't:

>>> 99/(23.99988/12+1)
33.00011

So it's only a small error in this case. It would be good to understand why it is happening, and whether the error is more significant in other cases. Multiplication by 3 should be relatively easy to get right.

I, as well as others I'm sure, have tried bringing it up to Stunna multiple times and I've never gotten a response on the issue. I presented it as a bug at first but they don't seem to think the same.

Quote from: MICRO on November 02, 2014, 03:27:42 PM