If the server seed was changed at a particular nonce, the bets after that nonce should all verify using the latest server seed. If this set includes the losing streak, and unibtc was playing pretty unpredictably it's unlikely any foul play was involved.

It should be pretty easy to figure out at what nonce the server seed was changed (just work backwards until the bets doen't verify). And if the site has the original server seed then the bets before that nonce could be verified, which would make sure that the server seed change wasn't hiding unfair bets (that's assuming that unibtc was betting in a reasonably difficult to predict manner) and that it was just a silly mistake  =)


Or even if the site doesn't have the original server seed (which is a concerning red flag), it would still make sense to verify as many of the last bets as possible using the latest server seed and check if a losing streak fell under the latest server seed. If the losing streak happened to fall on unverifiable period, that would be another red flag.


(Although to be honest, I suspect the whole thing is just a silly mistake but the whole point of provably fair is that we can verify not trust)

Oh nothing interesting like that. Just more annoyance type things, like denial of services and having all blocks being withheld (despite having literally hundreds of peers)


Oh no, sorry it's not a website just a bitcoin peer in the network. It's really only useful if you run a node or spv client and want an honest node to connect to

BTW I've been having a lot of problem with malicious bitcoin peers, so I decided to setup a little firewall of my own using full validating nodes. So if anyone wants a to connect to a safe full-node that's not out there to kill, log or deprive you of blocks or transactions I've setup a public: bitcoin.moneypot.com

Well, this has been our busiest month by far!   

Over the last 30 days, we've been averaging over 800 BTC a day and processed thousands of bitcoins in deposits and withdrawals =)

Actually, this isn't 0-risk for investors =)

There's actually a 97.2% chance that it lands on red or black. If it does, the investors need to give the gambler their money back, but also have to pay the app owner 50% of the house edge. So this means that the investors have a 97.2% of losing 0.054 BTC, and a 2.8% chance of making 7.94600 BTC. A good deal by any measure, but certainly not without risk.

(But like my earlier roulette example, there are cases when investors don't actually have risk)
 

Yeah, I understand where you're coming from and I think your point is valid. But just remember, one of the reasons MP accepts investors is to try get buy-in from investors, and that they'll want to promote MP (and it's apps) as they have a financial interest in seeing it do well. But all this bickering and in-fighting over the crumbs has become a bit toxic.

Nah, it's very different.

Let's imagine you have 10 BTC that you want to invest, and want to do so at a full kelly (the ideal). But you also want to minimize your counterparty risk, so you'll want to deposit as little as possible on the site.

With offsite, what you'd do is deposit 1 BTC and specify 9 BTC offsite. With leveraged investments, you'd deposit 1 BTC at a 10x kelly. Assuming a 1% house edge, in both cases you'd be risking 0.1 BTC a bet ... which is exactly what you want (a full kelly). So perfect.

Now, lets say your investment does well and you make 2 BTC profit.

With offsite investment you'd have 3 BTC + your 9 BTC offsite, and would be risking 0.12 BTC ... which is exactly what you want (A full kelly). With leveraged investments you'd have 3 BTC + your 9 BTC offsite, but would risking 0.33 BTC ... which way too much (to the point of landing you in negative expected growth territory).

And the exact opposite happens when your investment loses money, the leveraged investment you'd start to risk too little and would be under-utilizing your money.

Fair enough, PM me your bustabit username and i'll credit you as well =)

While investors always have the risk of trust me with their money, it's indeed possible to place a bet that poses no-risk to investors. For instance if you play that roulette game, and put a 1 bet on every number (including 0) its impossible for the investor to lose (or the player to win).  Not really a common case for single-player games, but it's easy to imagine in a multiplayer bet  (Bob wins 1.99x if it's heads, Matt wins 1.99x if it's tails) where they both bet the same amount.

Currently investors will make 50% of the house edge for that bet (even though they shouldered no risk, so it's effectively free money) but it's really such an edge case at the moment

This is not possible, and not something I've currently planned (or understand the usecase)

Under what circumstances would you want this feature? The only one that comes to mind right now is a punitive one?

I love hearing proposals, but as profit-share is inherently a zero-sum game it's pretty draining to hear: "On X site investors get Y, we should get it too!"  and  "If i privately bankrolled i'd make twice as much!". Both statements are true, but it doesn't really help me strike a balance between apps, investors and myself.




I'm not sure if you're being obtuse on purpose, but as it's been repeated over and over in most cases the new scheme is better for investors, and in all cases the new scheme is better for you. In the case of a 0% risk bet, it went from:

JackpotRacer: 50%
MoneyPot: 50%
Investors: 50%

to:

JackpotRacer: 50%
MoneyPot: 5%
Investors: 45%

In fact, it should be a real positive for you as it'll encourage more investments which will allow you to place bigger bets.

Thanks for the chart Dooglus, once again another fantastic explanation of things =)


Yeah, it's already live


A zero risk bet would constitute 0 on the x-axis of Dooglus' graph. So both apps and investors will get exactly half the house edge

NLNico beat me to it: http://dicesites.com/moneypot and as you can see, betterbets is really doing a phenomenal job.

I believe TTM stands for "ToTheMoon" another bitcoin gambling site (http://tothemoon.me) based on the bustabit source

Sorry, you're right. I deleted my rather unprofessional post. Just sad watching such a legend go this way. They do seem totally functional, you're just  a little SOL if you need their support though (I too emailed support a few months ago, and never heard back from them).

If any (developer) is still using a-chat-server can you please migrate off ASAP to socketpot ( https://blog.moneypot.com/introducing-socketpot/ ), as we'll soon be shutting off a-chat-server.

The chat API is virtually identical, and supports more features (persistent history, moderators etc) so it's worth upgrading anyway.

Thanks, noted. Both very reasonable, and added to our issue tracker.

Offline investments (which I think is a fantastic idea) ended up getting implemented in just-dice with clams. MoneyPot went for a bit of a different scheme, and for the moment bustabit is just privately bankrolled, which helps keep things simple.

Sounds like fun! Count me in: 1LB9EkKSpRVFKYH97HSYqy2czM9H8Rhx8i  (Note the first three characters of the address, and behold my manual vanity address generating prowess)

Yup! It's a great way to think about it.


_{After I properly wake up, I'll check if that maths works though. I believe it's going to result in a slight underestimation of investor profits though, as my intuition tells me that if investors gave 20% of the house edge, on a 0.8x kelly bet it would subject them to slightly more than 1x kelly of risk (which the MP system goes to great pains to avoid ever happening). But that formula you gave should be either very close or accurate.}