Those different level colors look pretty cool like that :p

Seems like PD just reached BTC 2,000,000+ total wagered. Congrats

https://dicesites.com/primedice

Sorry, but this looks like a simple scam.

If you actually have BTC50,000 ($45M) - you should be able to hire at least 1 normal developer. This site looks completely shit and broken

The "deposit now and earn some easy BTC0.2" angle is pretty obvious too.

Nice update. I am not sure if you updated the provably fair implementation now or already before, but I just had a proper look again. The implementation looks now very close to perfect I happily activated your site on my site again. There are still some smaller things that could be fixed:

1. Changing the client seed manually doesn't actually work (but "generate new" button does work.) So needs some change event listener that puts the new clientseed in the JS var and localstorage.

2. Ideally the function verifyRoll() will also check the HASH which is the first part of verifying a result. So something like this:

Obviously that alert should never happen though

3. If you use the above code, you would be checking the hash that you got after the bet. But you should check the hash that you saved before the bet (just for the theoretical situation where MP tries to cheat by changing the hash too.) You put this serverseed(=hash) already in the box with insertProvablyFairHtml() before the bet. So you can simply remove the following line from updateProvablyFairHtml() to ensure you verify the hash that you got before the bet:


4. It says "Next Bet" in the second box, but that is actually the "Previous bet"



Still I believe your provably fair implementation is now superior to almost all "per-roll implementations", so these are just smaller things that should be easy to fix





PS, I only tested Dice-BTC. Perhaps I can look at Plinko and NXT some other time if you want.

I think the problem is also that it's tough to really qualify which casinos requires ID. I am considering to make a list that simply mentions KYC in TOS. But then again, there might be sites that list KYC in their TOS just to cover themselves (from potential law enforcements), while not requiring this from (practically all)/normal players. Of course I could test sites (like OP), but that is still just 1 experience and might be different with other amounts.

Same for other situations like "instant withdrawals", I know plenty of sites that are "normally" instant, but of course if HW is empty it can take more time. So it wouldn't be fair to judge it on a few experiences.




OP's list seems like a decent start though

Yes, it is. You can see xbach already mentioned it a few times: https://bitcointalk.org/index.php?action=profile;u=881412;sa=showPosts

Another source would be reddit: https://www.reddit.com/r/TREZOR/comments/5nd56x/introducing_trezorio/ (slush0, stickac, xbach all replying in there.)

This depends on what house edge you would like to bet against.

You can bet up to BTC 6.5 on 0.5% and up to BTC 13.23 on 1% (which is like most dice sites.)



The maximum house edge is 5%, so in theory you could win up to BTC64.81 per bet.

17:13:22 = loss (timestamp from you)
17:13:23 = win
17:13:24 = loss (timestamp from blockchain.info)
17:13:25 = win (timestamp from blocktrail.com)
17:13:26 = win
17:13:27 = loss
17:13:28 = win
17:13:29 = win

My point is not "this should have been a win", my point is: there is no fixed timestamps, so you could choose any timestamp in your favor without us detecting you are cheating.

Uhm, your site says the timestamp is 1484500402 (= :22) and the result hash is:



So you agree with me that there is no fixed timestamp for each transaction and you could just choose a timestamp that makes the player lose?

I don't understand why you have don't believe me somehow. I find it very frustrating that you, as "Legendary" member, are trying to defend this guy. You keep repeating that this site is provably fair, even though I explained you 1 month ago(!) that this site is not provably fair. You are effectively helping his scam.




I will repeat:

This site is NOT provably fair.

The result calculation includes a timestamp but transactions don't have a fixed timestamp. Therefor the site owner can just make the timestamp a second later and generate a completely different outcome. Remember: you are not betting on a bitcoin price here. The price is only 1 part of the hash calculation resulting into a "odd/even" result.

I could copy this site's "provably fair method" and generate a losing result 100% of the time - that will still "verify" correctly.


Example

See this relatively big loss of BTC0.05 on crazybtrade.com:


Timestamp 1484500402 =    Sun Jan 15 17:13:22 2017 UTC

If you look at the transaction on blocktrail:

https://www.blocktrail.com/BTC/tx/0ea617a32f00d73d2e6d48d12adc08e60182ae12efab336dc965bbae6861f047

Received Time: Sunday, January 15th 2017, 17:13:25

Can you see, how blocktrail got this transactions a few seconds later? Do you realize there is no way for us to verify which timestamp is correct?



Let's calculate the results with those 2 different timestamps
First result of crazybtrade, SHA-256 of:
=
Since the first number is a 0 - and the bet was on "odd", the result is a loss.

Let's see what happens if we change the timestamp to the one on blocktrail (a trusted block explorer):

=
Since the first number is a 7 - and the bet was on "odd", the result is a win. So that player could have won according to the time of the trusted block explorer Blocktrail.




Note: there is no pattern between timestamp and result, every timestamp would simply generate a completely new random result. Since 0-20 second delays are normal at the bitcoin network, there are enough tries to basically guarantee a loss.




Like I said already 1+ month ago. This site is not provably fair. The site script can just cherry-pick a timestamp to make big bets lose. There is no way for us to know. And the SHA-256 result calculation will still verify according to his method.




If he really is the same scammer as that RollinCoin guy, this is an interesting development. On RollinCoin he faked the hashes to be a losing result. Since that site had a decent provably fair method - we actually could detect he was rigging the results and therefor see that he was scamming his players (and stop it by warning them.)

However with this new "provably fair calculation", "rigging the results" literally cannot be detected - since it's not provably fair in the first place. Pretty smart in a way I guess. Or maybe people like this "Legendary" member can just stop spreading the misinformation of the claim of provably fair so people might stop putting their money to this scam site.

From the abstract:

SHA-256 has 64 rounds and SHA-512 has 80 rounds. Since this is impossible to attack, researchers reduce the number of rounds and try to attack that. It gets exponentially more difficult though. So if 24 rounds (from that paper) or even 28 rounds (based on latest papers) of SHA-256 are practical enough to attack - it still doesn't compromise the security of full SHA-2 at all.

I will repeat: There are no known SHA-2 collisions.

How about you just give me 1 example of a SHA-256 and SHA-512 collision? Since you claim dice site owners can generate these on the fly, you should easily be able to give us just 1? If you want we make a bet out of, I will be happy to bet some coins on this (with escrow obviously.)

I have had plans to make that for years now. But since half a year there is a site for that: http://spotodds.com

I am not sure how good that site is (I would do at least some things different.) But at least for me, I plan to just focus on development of other sites first now

Did you even read the title of that paper? "Freestart collision for full SHA-1".

We are talking about SHA-2 (includes SHA-256 and SHA-512) - used by the dice sites.

Lol, nice one Avirunes.

I have considered to remove them before for having bad support, but generally the issues were always solved "eventually". These latest issues seem to be very serious though and about big amounts, so I removed them from my site for now too. Note that I do still track the stats at https://dicesites.com/satoshidice I hope it will all still be resolved.

In the beginning there was a bug were the Doge JP was sometimes shown as BTC JP. Seems like this is related to that. Did you switch through those currencies or something like that?

That's called "hash collisions". Dice sites always use SHA-2 to hash their serverseeds. SHA-2 has no known hash collisions.

Note: if it were that easy to make collision attacks on SHA-2, dice sites would be the least of your worries. Governments, bitcoin, the Internet, banks, etc. would all have a huge problem.



edit: you are talking about MD5. Dice sites do not and should never use MD5 indeed.

The whole idea of the provably fair method that dice sites use, is that you get a HASH of the serverseed BEFORE you bet. Therefor you can verify (= provable) if that hash is still the same after the serverseed is revealed.

If the site changes the seed (what you call "fake seed"), the player is able to detect this, because the HASH would be different. Therefor he can prove that he was cheated (or if he was not cheated.) This is provably fair.




On your game, there is no way for the player to know if that specific timestamp is correct. Therefor that is not provably fair.



PS, you look more idiotic with every post you make, so I don't think this is helping much for your reputation as a gambling site owner.

Lol?

No, I am saying many times now that you cannot use timestamp in a provably fair calculation. What has that to do with BTCC? Did you read my message here or just ignored it?




In addition, I am sure that BTC/CNY price is just as well unreliable and unverifiable because there is no clear public history of 1 specific price per timestamp - this has nothing to with the reputation of BTCC. But I am not even talking about that here.

Your game is not provably fair. I could copy your game and generate a losing result for the player 100% of the time while it would still "verify" according to your method. I can do this by simply using a slightly different time, because there is NO fixed time on a transaction.

I have explained this multiple times now:



And in more detail starting from: https://bitcointalk.org/index.php?topic=1666167.msg17188778#msg17188778







Since you continue to ignore me and claim that this game is provably fair (while it really is NOT provably fair AT ALL), I will add negative trust too. Your game doesn't have to be provably fair, but if it's not, you should not promote it as provably fair.