Edit: I mixed up my terminology. Valor is what gamblers get. Silver is what investors get.


I'm trying to avoid repeating the mistakes I made in moneypot, so I guess here's the backstory: in MP I could've provided the bankroll myself (and/or shared it with BaB), but I intentionally didn't as I thought it would be a community-building feature type thing. I also assumed it'd be less stressful if I wasn't a big investor and didn't need to worry about variance.

However what ended up happening is:  a) Risking other peoples money is actually more stressful than risking your own.  b) Almost all substantial investor logged in only ~twice. Once to invest, once to divest.

During the ~6 (?) months before I sold the project, I believe those investors made something like 30-40% return in that time period. The whole thing left me thinking that if I were to do it again, I would structure it radically differently.

--

So taking from that, my original idea was that only active gamblers could invest in BaB, or there would be certain wagering requirements. Then after generalizing a bit, I came up with the idea: for every 1 bit of investing profit you make, you need to have wagered 100 bits. But if you're an investor, and not a gambler -- you can buy that "wagering requirement" from actual gamblers (and thus subsidize their play).

So as you can probably tell by now, "wagering requirement" is known as "valor". Gamblers automatically get 1 valor for every 100 bits they wager. And "investing profit" is known as "silver". And they are freely tradable between players, with the ultimate goal of community building. To also further this goal, investors profits will be automatically divested each day. This way investors need to login to every day and interact with the community if they want to experience the magic of compounding returns (i.e. investors profits are divested in the form of valor, and they need to either trade or gamble to get silver in order to convert it to bits, to reinvest).

From gamblers point of view, the new setup is pretty great. You effectively will get a ~50% off the house edge, selling your valor to investors

 


Yeah, although if you need/want your testcoins back, just shoot me an email and your address and I'll happily send


Thanks, I'll take a look soon. Something came up in my personal life that's taken me out of action for the last couple of days and will stop me working on it for at least a few more. However, I'm still checking in on things  and making sure all the prod support tickets are handled and everything is running smoothly. But there probably isn't going to be any new commits this week.



I should probably create a different thread for the dev site, so people don't think there's a production problem. But the withdrawals there don't seem to be working at all, i think I broke something pretty badly. I'll be fixing that soon =)

So due to popular demand, I'm releasing the current dev branch of bustabit v2:

http://dev.bustabit.com

It uses bitcoin testcoins instead of real bitcoins, and comes with a faucet (you can also google "bitcoin testcoin faucet") and deposit them in your account.

You can find the source here:
https://github.com/bustabit/bustabit-client

It's completely full of bugs, and half-done stuff. The best thing is to make issues on that repository -- make note of your bustabit username, so you can be rewarded with some vanity stuff in the indeterminate future.

Also pull requests accepted  




Edit: in an unrelated note, here's a game dump of the last 10M plays: https://ln.sync.com/dl/0c20bf440#78hvmigq-xjiymwmj-vqcdt46n-is8vhdnc

Yeah, bitstarz is pretty well known as a border-line scam. It's a pretty common practice of theirs to do everything in their power to avoid paying out winners while trying to look legitimate. Asking for proof of id and address seems lately to be their favorite way to do this, because they know a large amount of players would rather preserve their anonymity than winnings (I mean, I'd rather lose my money than give a bunch of shady people my full dox). They also then have the added benefit of being able to void peoples winnings based on where they live, and messed other excuses like that.

Pretty much the only reason people play there, is they seem to have a big advertising budget and pay affiliates well to sucker people in

There's a really common misconception that provably fair means the casino can't cheat, but what it means is that cheating is detectable if you verify. A good example might be dicebitco.in which used a robust provably fair scheme, but selectively cheated some players (by skipping some winning nonces). This was only detectable because of verifying the outcomes.

And then of course, there are casinos with all together broken provably fair schemes. Normally this just is a mistake, or comes from the developers not understanding it. Most new multi-player casinos I see pop up suffer from this.


I make a habit of always reviewing and verifying games when I play on someone else's site. I'm pretty cool with losing money, but I want to be damn sure it was bad luck and not me being cheated.

His bot seems to be doing reasonably well. It's wagered wagered 501.34 BTC, but got 8.1453 BTC in bonuses. His average cash out is around 1.28x, so that puts him pretty firmly in +EV territory.

rofl, interesting play. Out of interest how much did that cost you?

BTW if you're going to blatantly break the Anticybersquatting Consumer Protection Act, you might want to at the very least use some form of whois protection. Also your extortion price is a bit silly, ~2 BTC would be much more reasonable price as that's about how much it would cost them under the domain-name dispute resolution process.

Congratulations to LosDeXibalbaMC for joining the exclusive club of getting 200+ BTC in profit!

Yeah, 1) is concerning. Not only does open source make it easier to find attack vectors, it makes it easier to abuse them. Of course, it's never a good idea to rely on obscurity -- but it makes a good defensive layer.

2) I don't think I'd worry about. I think as long as the license allows you to relicense it, I'm not sure it'd be a big deal. Big companies buy companies for the player base and brand, not the code itself.


I guess the middle ground is to open source the front end, but not the back end. It largely avoids all the security risk stuff, but unless it's easy for people to run their own dev copy it might be hard to get contributors

In hindsight, it's pretty obvious. But damn, once again I'm impressed.

I think the coolest part about it, is you don't necessarily have to match an investor with a gambler. But you can match gamblers with gamblers. For example if I was willing to bet 1 BTC @ 2x and Bob was willing to bet 1 BTC @ 2x, and both willing to pay a 1% house edge. The casino could match the two bets, and take a risk-free expected value of 0.02 BTC.

Or someone could put in "order" to do a 100 BTC gamble with 0 house edge, etc. I think it'd be pretty cool to just operate it like a bet exchange, taking and matching orders.

The part I don't like, and is dealing with the edge cases around disconnecting. e.g. where one party doesn't or can't or won't reveal the seed that created their hash. You'd probably have to do something like they have X period of time to reconnect and provide it, or the bet is decided against them. But it's going to be a pretty bad user experience for anyone when that happens, and you'd likely need a ban or reputation system to ban gamblers/investors who did that too often.

I don't mean to derail the conversation, but just FYI they're not even provably fair for gamblers, let alone for investors. They just blindly trust random.org (and oracalze). My conversations with them lead me to believe they are dishonest (e.g. tacitly admitting they aren't provably fair, or fully decentralized but wanting to leave it on their site for marketing reasons) , and the only thing going for them is they spent tens of thousands of dollars buying advertising and paid media coverage.   I'm not sure how much money their investors have lost so far, but I have trouble imagine it'll be long before it's all.

  Yeah, that fixes it. Why is that? I don't get it

Interesting. Although:

You make it sound like they ran off with money or something. What happened, as I recall was rather uneventful. Someone had their large withdrawal (1300 btc?) manually processed, but before their account had the funds deducted they gambled with the "fake" money and lost it to investors. So instead of the money getting deducted from his account like normal, it was taken from the bankroll.

All things considered, I believe it went pretty smoothly and no one lost any money or was harmed

Yeah, although you could pretty easily have two different schemes: one for micro-bets which are fast and bankrolled by the site. And the second for the larger bets which get bankrolled by a provably fair investor scheme. I'm not totally convinced that if someone is betting 10s of thousands of dollars, they care about waiting a second or two. Hell, it's just good suspense at that point 

Yeah, i can think of a few ways.

The simplest is the one I developed before, you can check out pevpot's thread which is now defunct, but it had a provably fair scheme that was totally fair for both investors and players, and used the blockchain in a way that was robust against miners too. But since the method requires waiting on blocks, the UX is going to suck pretty hugely. But it might be worth it for huge bets (e.g. >100 btc). Maybe?


The other way to make it provably fair for investors, the workflow could be something like this:

1) User sends bet to the server
2) Server publishes that bet to investors
3) Investor with at least $BETPROFIT BTC in theier account accepts the bet, and sends a hash of a random number and seed they generated
4) Server locks user/investor balance for the amount and sends the server hash to the user
5) User records the hash, and then sends a client seed to the server
6) Server sends client seed to investor
7) Investor uses the client seed to know if they they won or lost, records it, and reveals their original random number and seed
Server validates the investor random number and seed, and resolves the bet in favor of the investor or client


That should be provably fair for both investors and users.


With a bit of scripting magic, I'm reasonably sure it's possible to make the above procedure not just provably fair, but actually completely trustless. I'll avoid saying too much, incase I stumble upon a bottle of amphetamines and get the energy to try commercialize it =)

Mine was generated with:

Now that I think about it, that x-axis glitch it's probably caused by a few days of downtime shiba had. Then when shiba did go back online, and download those days the timestamp would've been wrong, so those days statistics are being grouped in the wrong day.

It should be pretty easy to confirm by checking the raw data for missing days, but I can't think of a 1-line bash command for that. Anyway, it's not a big deal as the end graph results should all be correct, just a few days lumped together.

Yeah, I was wondering the same thing. I've been meaning to look into it, but a bit lazy. Here's the gnuplot file, and the raw data if you're interested:

https://gist.github.com/RHavar/31296dd5f396bb1935a098db507509c6


(It's possible there's a problem with the data, it was provided by shiba rather than me querying the master database directly. But shiba tends to do an extremely good job of keeping in sync, as it even has logic to fetch games she missed)

The chart is a bit confusing, but the red line is profit and it's axis is on the right hand side. So it actually is starting at zero, it's just not at the bottom because for a brief time the site profit was negative.  And the blue line is the wagered, and its axis is on the left (so it also starts at zero.

Yeah, there are a few bots like that around. I suspect they're not for sale though, as bonuses are rather 0-sum.  It uses to be a lot easier to make that sort of script, but now (as of a couple years go) bustabit blinds wager amounts (only gives their amount compared to others) until the game actually starts. This makes targeting players a lot harder, and a lot easier for them to trick bots