Peerbet.org - Play without house edge!

[Camello_AR]

The spanish one have a lot of errors (looking fast). Appears to be automatic translate like Google Trans

[smooth]

You realize the code in the first post produces slightly biased draws for ticket counts that aren't a power of two, right?

[giantdragon]

You realize the code in the first post produces slightly biased draws for ticket counts that aren't a power of two, right?

Someone already told me 2-3 month ago about this flaw and I have increased base to 28 bits, which makes skewness statistically insignificant.
Could you offer better algorithm?

[giantdragon]

The spanish one have a lot of errors (looking fast). Appears to be automatic translate like Google Trans

Please correct this translation.

[giantdragon]

My previous post seems to have gotten swallowed in the noise... Any interest in a Hebrew translation?

We don't have much Hebrew users, however you can start if you want!

[smooth]

Could you offer better algorithm?

The standard method is to discard the inputs that lead to a biased draw. If this happens, you could wait for the next block to pay off the raffle.  It is very unlikely to happen in practice so as a practical matter, you can certainly ignore it, but people tend to get a little heated about "unfair" gambling sites.

I think the top answer here is correct, but I didn't examine it that carefully: http://stackoverflow.com/questions/11758809/what-is-the-optimal-algorithm-for-generating-an-unbiased-random-integer-within-a

[AdriKGB]

I can translate to Czech language.
Are you interested?

OK, start!

I can translate to spanish. Let me know if you need it. I am native spanish.

Someone already have made website's Spanish localization. Hovewer strings.xml for Android app is still missing.

/me proud to be Italian... well Italian-American

Lets try.

Hi,

I have reviewed the spanish translation of the xml for android and it is very poor (google translator for sure).

As agreed I have corrected it (I am sending it by PM together with and excel sheet indicating the changes).

I am waiting for the 0,1BTC in Peerbet; My account is AdriKGB.

If you need more help please let me know.

[samurajj]

I finished czech translation and uploaded it here: http://www.megafileupload.com/en/file/405155/locale-zip.html
I expect 0.1 BTC at 1JohN3jM5gFjGdAxkxnSai758fnYFhFvFB
Thanks.

[giantdragon]

The standard method is to discard the inputs that lead to a biased draw. If this happens, you could wait for the next block to pay off the raffle.

Discarding blocks can let users to think that Peerbet operator tries to play against its users (i.e. buy first ticket in the 50/50 raffle and discard all blocks until admin's ticket is winning), therefore it is not acceptable solution.
28-bit base number is large enough to produce high-quality random numbers with statistically insignificant skewness (it is even impossible to detect in stats software like Statistica).

It is very unlikely to happen in practice so as a practical matter, you can certainly ignore it, but people tend to get a little heated about "unfair" gambling sites.

This can be unfair in "play against house" casino games where this skewness can be used to increase house edge, however gives no advantage to the P2P gambling site operator.

[giantdragon]

I finished czech translation and uploaded it here: http://www.megafileupload.com/en/file/405155/locale-zip.html

Thank you!
Write your Peerbet username.

[samurajj]

my username: samurajj

[smooth]

Discarding blocks can let users to think that Peerbet operator tries to play against its users

You misunderstand.  The discarding is entirely deterministic based on the formula.  There is no way for the administrator to exploit that.  i.e. still provably fair.

[giantdragon]

You misunderstand.  The discarding is entirely deterministic based on the formula.  There is no way for the administrator to exploit that.  i.e. still provably fair.

You are right, but some users may incorrectly understand why some blocks being rejected and just stop playing. Suggest another algorithm without this disadvantage.

BTW, I still don't understand why change of algorithm is necessary when skewness does not (statistically significant) affect fairness of the game?

[smooth]

You are right, but some users may incorrectly understand why some blocks being rejected and just stop playing. Suggest another algorithm without this disadvantage.

BTW, I still don't understand why change of algorithm is necessary when skewness does not (statistically significant) affect fairness of the game?

Your statements are contradictory here.  If the bias is very small (which it is), then blocks won't be skipped very often and users won't have anything to be upset about.  It's not going to happen very often, if ever.  But as things stand now you can't (honestly) say that your game is "fair."   You can say that the bias is "small" and users might be okay with that, or they might not.

But anyway, another algorithm would be to increment the nonce and generate a new hash.   Again, you do this deterministically only when the first hash falls in the bad range.

[giantdragon]

But anyway, another algorithm would be to increment the nonce and generate a new hash.   Again, you do this deterministically only when the first hash falls in the bad range.

Suggest working PHP version of the algorithm with nonce increment and I will update Peerbet scripts.

[smooth]

But anyway, another algorithm would be to increment the nonce and generate a new hash.   Again, you do this deterministically only when the first hash falls in the bad range.

Suggest working PHP version of the algorithm with nonce increment and I will update Peerbet scripts.

I'm not a PHP coder.

If you don't want to do it, don't do it.  I was just giving feedback based the incorrect algorithm you posted on the forum.

[samurajj]

Hi, I made a few security tests on your site. I didn't find any SQL injection vulnerability, but I found out a way to do XSS attack. Is there any bounty for that?

[giantdragon]

Hi, I made a few security tests on your site. I didn't find any SQL injection vulnerability, but I found out a way to do XSS attack. Is there any bounty for that?

I can add some credits to play on your Peerbet account.

[giantdragon]

If you don't want to do it, don't do it.  I was just giving feedback based the incorrect algorithm you posted on the forum.

I never said that I don't want to change the algorithm. I just thought: because you wrote about this flaw, you could also suggest how to correct it.

[giantdragon]

I have implemented new random algorithm that solves bias flaw. Please ensure that it is correct.

Code:

function getWinningTicket($nonce, $blockHash, $ticketsTotal)
{
	$randMax = 268435455;
	$remainder = $randMax % $ticketsTotal;
	$output = -1;

	do {
		$hash = hash('sha256', $nonce . $blockHash);
		$randHex = substr($hash, strlen($hash) - 7, 7);
		$x = base_convert($randHex, 16, 10);
		
		$output = $x % $ticketsTotal;
		$nonce .= '0';
	}
	while($x >= ($randMax - $remainder));
	
	return $output + 1;
}