It shows in the bot the nonces are out of order (which are weird) but according to the site, they're not:
https://safedice.com/bets/157520191 <-25
https://safedice.com/bets/157520202 <-26

Can you send me your settings/script so I can try to reproduce it please?


Edit: I verified your rolls just to be sure (I see you've generated a new seed since then, so you've probably already checked), but they do verify correctly. It seems just the nonce in the bot got out of sync with the one in the site.

Nonce roll
24   25.8531
25   87.2483
26   90.4042
27   42.6134
28   59.9948

There's not missing nonce between the bet before and after that bet, so nothing really happened, but I see the chance is 100. Are you using a script or settings that could possibly some-how set the chance to 100%?

The invest button and function only work on sites that allow investments and allow the bot/API to do the same. At this time, it's only JD, betking and safedice that allows it, so it won't do anything on any other site.

edit: Forgot to actually answer your question. The invest button invests the specified amount into the sites' bankroll

What do you mean it stopped working? Are you using your API key to log in?

That post stunna mentions basically highlighted 3 things:
1: Ambiguity in the provably fair method
2: The site makes it difficult to get your server seed hash before every roll because you need to click the button, and the hash for the next roll isn't sent with the result of the last one.
3: the poster got banned and asked not to play at 999dice anymore when he wrote a bot that gets his seed hash for every bet.


Now, in 999dice defence:
1: the description and the code is still ambiguous, but when using their code (which is type sensitive), it works perfectly.
2: After that post, the site updated their API to send the server hash for the next bet, which, if they were cheating by changing seeds, removes their ability to do so.
3: The poster provided no proof of this and I find it hard to believe. What most likely happened is he spammed the site with requests and got IP banned for an hour, much like JD and PD does with login attempts, and the "Too many requests, slow down!" error many users get from PD when running a bot. Before the API update, my own bot was getting the server seed hash for every bet before the bet (and that made it hella slow). None of my users ever had an account permanently banned or was asked to stop playing there because of it. I was never asked to change it or remove it. In fact, Jake emailed me about the changes to the API to let me know I can update the bot to improve speeds.

999dices rolling is provably fair, I've made quite a lot of bets and verified them in an attempt to catch them cheating, because of the many people saying the many people claiming they were cheating. None of my rolls haven't verified.

The only thing that's actually somewhat sketchy about 999dice is their max profit compared to their house edge.But there's always the possibility he actually has that much money, or that max bet is simply ~1% of his bankroll

Certain features can only be enabled if the site allows me to implement it.

All of your questions are already answered:
1: The list wasn't made public because people bitch about not being paid when they didn't qualify. Micro said he'll make the list public when hes home from vacation. Talking out of experience, it's really hard to access a computer that is turned off, possibly unplugged if micros anything like me, behind a firewall and probably not connected to the internet at the moment.
2: Because no-one met the requirements to win, so no-one to credit. Thus, as soon as it was over, it was credited.
3: He didn't. everything that needed to be credited was credited.

This sounds like a job for...... TRANSLATOR MAN!

too bad I'm not him though. Good luck

Can you please test it on another site as well with the smaller amounts?

I don't mind suggestions, but please listen to my responses and requests as well.

I can't use the original balance - new balance as profit, as on many sites I have to calculate the new balance manually and deposits/withdraws/tips/bets made outside of the bot influence balance and will thus influence profit as well to display incorrect profit.

Seriously, I have general idea of where the issue is and the tests I suggested will narrow it down a lot.

I have not yet released a new version of DiceBot with the fixes for BetKing, I will do so ASAP

I have not yet released a version with the fixes for betking! I will do so asap

The latest version is 3.2.0, not 3.1.7. Bets are committed to the DB once the result is received. At this stage, the bot is not capable of committing a bet to the DB before sending it to the bot (the betID and site name are used as primary keys, committing a bet before the betid is known is going to crash things). This is a good idea though and will be looked at for v4 of the bot.

The balance will be the same since it's received from the site after every bet. The profit for each bet I have to calculate manually, thus rounding errors can occur. If you scale your betting and the difference scales with it, it can help identify the problem. For example, on the best shown on your screenshot, the bot could be calculating the profit as 0.00000019295 but the site could be rounding it to 0.00000019, OR the other way around. over 8 bets, this is a difference of ~5 satoshi. If the bets are scaled and the difference scales with them, it's something else. If the difference doesn't scale, it's a rounding problem. Please test this.

Testing on another site could also help identify the problem, if you don't mind testing on PD or BetKing.

Alternatively, you can browse settings on https://bot.seuntjie.com/settings.aspx or programmer mode scripts on https://bot.seuntjie.com/Scripts.aspx

Hmm I am unable to reproduce this error. It might be a rounding error resulting from the odd bet sizes you're sending. this is likely for 999dice and one or two other sites, since I receive very little actual info back from the site about the bet. This means I have to use the info from inside the bot and not what the site received, and due to rounding, this isn't always 100% the same. Can you try that same script but with 10 or 100 sats instead of 1 and see if the same thing happens? (Doge works well for these tests)

On which version of the bot was this? I know there was an issue similar to this in an older version of the bot but it's been fixed for a few months now.

Also, it could be a connectivity issue. That's the exact reason I fought against having the bot retry bets if something goes wrong, but people kept complaining and pushing me to do so, so eventually I did.

Can we get chat API documentation for the websocket? Since the chat API doesn't show tips or private messages

Reviews are subjective. Very very subjective in the case of casinos.

Some users win a bit and know when to stop, and leave a GLOWING review of this amazing site where they made hundreds of dollars.
Other users win a bit and don't know when to stop, and lose. Or just have bad luck right off the bat and lose, and leave a massively bad review about the exact same site as the guy above.

Did the site cheat? was it unfair? did did cheat one player but not the other? No. Some were lucky and others not. A big problem is that people are more vocal when they lose than when they win. You'll see this especially when someone accidentally loses a big bet. Like when someone accidentally clicks on max bet while hunting x9900 (0.01% chance to win) and they lose, they're furious and many times blame it on the site and ask for a refund, but if they won that bet, they'd just say "Look how lucky I was! I won all this money!" no-one would ever say, "oh, I accidentally won this much, I only meant to bet 1% of what I actually bet, here, take back the extra bit I won since that isn't what I meant to do."

This can easily be seen by these two posts (and others) in this thread even:


999dice doesn't cheat players, and if you feel they do, you have a way of proving it. I've made thousands of test bets at 999dice to address the claims of cheating and tried to prove that they do cheat. In fact, all I proved was that they didn't. Not one of my rolls at 999dice didn't verify, lost when it was supposed to win, was different to what I submitted or anything suspicious at all. 999dice's bad rep comes from some ambiguity in their provably fair documentation which has since been addressed and I haven't been able to find anything wrong with them.

That feeling of them cheating you is a placebo effect. You can make the exact same bets and get the exact same results at 999dice and, for example, primedice.
 You've read reviews that say that 999dice cheat (no-ones proved as much though), so as soon as you lose a bit there, you remember the reviews and go, oh, they were right, the sites cheating.
But at PD, for the exact same rolls, losing the exact same amount in the exact same manner, you'll think of all the people saying that PD is the most trusted casino and they've never cheated anyone, and you'll just think that you had a bit of bad luck, which is the case for both scenarios.


The reason for the advanced technical stuff is so that you don't need to rely on subjective reviews. You don't even need the advanced IT knowledge to verify your rolls, most sites have third party verifiers available. Dicesites.com has a big list of them and my bot also has them built in. This means you only need your server seed, client seed and nonce range (if applicable) to check your rolls and see whether the site cheated you or not.

True, but their wagered stat is skew. Look how much it dropped (https://dicesites.com/primedice) during the time they limiting the faucet. Whats the worth in having the largest daily wagered amount if 60% of the wagered bitcoin is your own? *they'd still be mostly #1 even if they lost around 60% of their daily wager

Edit: I'm mostly just sucking numbers out of my a.... thumb here. I have no insight into what PDs actual wagered vs faucet wagers etc are. Would be interesting to know though.

How to crush faucet abuse in 3 easy steps:
Step 1: Remove faucet
Step 2: Ignore people complaining about there being no more faucet.
Step 3: Lose tons of wagered volume and players, but see no real difference in profits (maybe possibly even increase in profits).


People that play only on the faucet, even those just playing for fun and not trying to farm it, need to realize that there's a cost accompanied to processing and storing each bet (server,internet,db +storage space,etc costs). Even if you claim the faucet only 5 times a day and make a few 100 bets for each claim, you're costing PD money to play it. PDs losses because of faucet isn't just the occasional farmer that gets it to a point where he can withdraw, it's much much more. Many sites have minimum bets far above 1 Satoshi because it's more expensive to process and store a 1 satoshi bet than the site theoretically makes from the 1% edge on that bet. Now add to that the fact that the money the 1 sat bet is made with is PDs in the first place and the site makes NO money from that bet, win or lose, and the cost just compounds. Faucets are a much bigger expenses for casinos than most people realize. You're not entitled to a faucet The site's doing you a favour, at their expense.

I'm certain PD can take the money they spend on faucets and processing/storing those bets and allocate them to more useful things, like rewarding 3rd party developers (especially HUI and nahtnam), having more giveaways and competitions and possibly even paying some of their mods for their hard work.

But that's just my not so humble and probably extremely unpopular opinion, besides, what do I know of running a casino?

Also, ETA on chat API documentation?