Quote from: natri on June 20, 2014, 07:51:33 PM

Quote from: wetsuit on June 20, 2014, 07:40:33 PM

Quote from: natri on June 20, 2014, 07:36:44 PM

Quote from: Luror on June 20, 2014, 07:12:02 PM

But you said you made only 500 rolls? Or did you mean some bigger test? You need more rolls to get 3rd and 4th prizes (at least 1250-5000).

That last statement ("I've never rolled more than 9993 on any of my multiplier or regular rolls on neither faucet") I referred to all of my plays on either freebitco.in and freedoge.co.in which total way over 2000 rolls so by calculating 1-(9993/10000)^2000+ the probability should be 75%+.

Ok, lets assume prob for rolling 10k is 1/20000. But by simply checking the freebitco.in WEBSITE STATS, there was currently 538,271,966 played and 394.52490422 BTC earned, if we consider that you can ONLY win top roll price (huge undervaluation) and probability is 1/20000 to earn ONLY 0.1 BTC (undervalued) -> the expected earned value of BTC should be 2700 BTC (way more than 394).

And if we say that all ppl were extremely unlucky we can check;
variance=npq=538271996*1/20000*19999/20000
sqrt(var)=164

99.7% rule (3 std deviations): Still users should earn at LEAST: 2700-3*164= 2200 BTC, and this is with only top win of 0.1 btc.

Those rolls include both free and multiply rolls, and multiply rolls are way more in number.

Alright, that explains it sufficiently enough then

But why is there such a huge discrepancy when generating a roll. Notice the frequency of occurrences inside the 5001-6000 bracket. I am not trying to accuse or anything but the randomness procedure might well be flawed.

Posting results:

Number   Expected   Observed   (O-E)^2/E
1-1000 50 45 0,5
1001-2000   50   42   1,28
2001-3000   50   50   0
3001-4000   50   39   2,42
4001-5000   50   52   0,08
5001-6000   50   77   14,58
6001-7000   50   51   0,02
7001-8000   50   41   1,62
8001-9000   50   42   1,28
9001-10000   50   61   2,42

I am not sure to be honest, maybe you need a larger sample size ? I shall look into ways to make the client and server seed generation more random.