Well in the last week of tying out this site I lost the 15 btc that I put in, I wasn't expecting it to be drained so fast but maybe that was because of the type of bets that I made, with no house rake I'd expected to be around, up & down, for longer - anyway as it's obviously not so lucky for me atm I shall leave it alone, if it's still around in a few months & getting good feedback then I may try some more - I like the concept a lot & all my bets were of the 'provably fair' type, but as yet I'm not totally confident re that, no offense intended.
What happened here is not because of fairness, but rather because of strategy.
There are strategies in Peerbet that are almost guaranteed to lose one their money. These strategies effectively convert Peerbet to a low-rake lottery. One of the common ones is the proportional bet strategy, which many people subconsciously adopt.
For example, imagine a hypothetical person with 1000 XBT to bet. He always bets 10% of his balance at 50/50 odds. You may expect him to have a 50/50 chance of being in the green. But let's examine the actual case.
After one bet, your expectation would be accurate.
50% 1100 XBT
BREAKEVEN 50%/50%
50% 900 XBT
But after two bets, it becomes clear that this isn't the case.
25% 1210 XBT
BREAKEVEN 25%/75%
50% 990 XBT
25% 810 XBT
There's a short return to the mean at three bets:
12.5% 1331 XBT
37.5% 1089 XBT
BREAKEVEN 50%/50%
37.5% 891 XBT
12.5% 729 XBT
But that is a short-lived return. If you noticed, the chances of each option is also a binomial coefficient. Now, we just need to model the median to determine what the most likely scenario is. For zero or one bet, the median is 1000 XBT. For two or three bets, it drops to 990 XBT. Because in theory betting with 1000 XBT is no different from betting with 990 XBT, we can now deduce that for every two bets, the median drops to 0.99× of what it was.
This means that starting with 1000 XBT, after 1000 bets, you should expect to be at ~6.57 XBT, effectively losing most of your money. So where did the "no house edge" go?
It turns out that this picture is not all bleak. The missing money is in the unlikely rewards. There is virtually no upper limit on the amount you can earn (the practical upper limit is that nobody will bet that high if you get lucky) if you win every game. But if you lose every game, you can lose at most 1000 XBT. This is where the lottery analogy comes in: you play a small amount, and you have the small chance to win a large amount.
There are, in fact, strategies that avoid this problem. Although they have far lower reward, the reward occurs often. However, when they are not rewarded, they can lose a significant amount of money; effectively reversing the lottery. Peerbet effectively lets you either play the lottery or run the lottery without a house edge, which is its greatest advantage. But you must educate yourself on probability theory before you take the plunge, or you will come up with an undesirable strategy.
