Impressive job. Minor nitpick, on the splash page you should fill probably prefill the input box with the attribute placeholder="player" instead of: value="player"

That's a reasonable concern, and a concern you should hold to everyone before ever investing money. And it's specifically our motivation for developing an investment scheme (discussed in depth a few pages ago) that allows you to minimize counter party risk by keep your own money where its safest, in your own control.

And for all the people not familiar with money pot: the motivation for that is to provide incentives to cash out early. And when people cash out early, it provides incentives for other people to stay in the game longer to collect bonuses. This interaction provides the core of the money pot player-vs-player play. And of course, the house edge is so low, that in general you can play strategy obliviously and still end up with better odds than any dice site.


That's one way of looking at it. However, I prefer to look at it as the volumes (and limits) are too low. My goal is to make money pot the bitcoin gambling site, and having crazy low house edge with the possibility of even playing +EV is important to that goal.

They haven't been at those levels since axedcode and troll~ won 40 BTC a piece. The recent dip is primarily caused by better who has made 14.9 BTC and and indie168 who have won 11.3 over the last couple days.

Better himself is quite the success story, at one time was down 9 BTC, is now up 11.

Which represents about an 8 BTC loss to the players, which drove the miniwhales home for now. There's quite an interesting relationship between how well people are doing and how much they wager. When the site loses money, the volumes go way up. When the site makes money, the volumes way down.

For all the people interested in the skill aspect, now is a pretty fantastic time to be playing. There's some good competitive play happening with 4 or 5 max-betters, meaning you can get pretty decent bonuses.


Also rather impressively, despite 29,800 BTC being wagered the house is at a loss. It's been a good run for players, to say the least.

While people might not care, I suspect it is the biggest factor of how much people play. For two reasons, one: it hugely increases variance for players, which they love. A concrete example: whitetuxpeng might now have lost a lot of money in the pot, but at a time he had a net profit of over well over 30 BTC. I'm sure those are great memories that keep him coming back.

And secondly, I think people recycle their winnings to a huge degree. So in general, I think it would hold that if you half the house edge, people on average would gamble twice as much, by the mere fact they can.

Sorry, that is what I meant indeed. I'll edit the post to clarify

I guess MB means mBTC? I'd encourage you to use bits. Where you put the decimal place is quite important as it takes a lot of time before people internalize the value of different denomination, and few people want to learn how to convert between satoshis, bits, BTC and SI prefixed BTC.

No, not quite. It's a common misconception that there is some sort of "memory" but none exists. When the game instant crashes (1 in 101) the money goes to the house, not an explicit account for the bonus scheme. This means that house can pay out more in bonuses than it receives from instant crashes. Or conversely, it can collect more from instant crashes than it pays out in bonuses.

The bonus scheme is pure EV0 for the house, but the house does expose itself considerably. Which is good for the players, and bad for the house.

Actually as your bet increases relative to the total amount bet, the closer you bonus approaches 1% regardless of game crash. Which means the effect of the bonus scheme goes away, and your house edge approaches 0.5%. This is because no matter when you cashout you will get a bonus.

So assuming the limits were higher, money pot is pretty much the BEST site I'm aware of to do a YOLO style bet.


Edit: fix incomprehensible grammar

Good point Dooglus, a 3rd party is not required.  All you really need to do is write a script that:

1. Generates a private key
2. Displays the address
3. Asks the user to send coins to the address, and specify exactly how many coins they sent, what txid they sent it with, and what output they used.  (If you were advanced, you could programatically get it, blockexplorer makes this easy)
4. Ask the user for a receiving address and when they want the transaction
5. Send a transaction to the receiving address, using nLockTime
6. Output the transaction. The user would have to push this to the network themselves when its eligible
7. Delete the private key

You don't need to broadcast your spend transaction until the gamblank gives you the returning transaction (to do so, you'll have to provide them with the txid, outputs and amounts). Only once you have the refunding transaction do you actually broadcast your funds to the multisig address.


This reduces your risk to quite simply: "They might keep their private key, and release the funds (with your consent) before they promised to"

SatoshiDice asks you for a client seed, and then adds it to their server seed, modulos it by some number to generate the roll result. Instead of picking a positive client seed, someone picked a negative client seed, which invalidates the scheme as it shifts down the range. If the seed is negative enough, the roll will even be negative. Negative is less than any target, so the bet always won.

In short, SatoshiDice didn't validate the client seed, which their scheme requires it to do.

I did the same thing, and searched for keywords like 'offsite' and read around there. I also followed a link to a thread where there was a poll from an angry investor with a appetite for variance upset about the change to a 0.25% max profit per bet.

...and we're routable again. It looks like it was just a hiccup with our provider. I'll chase it down with them to see if they can provide me with more details. I intend to make the most of this experience and setup a warm fallback game server for issues like these.

Sorry to everyone who was effected by this brief outage.

The game server is done right now. Or rather, it's not routable from the outside world -- but still online. Looking into it right now, it's not clear if it's because of an abuse claim, or null-routed by our hosting provider because of DoSes. We're preparing the move to a new game server, please have patience while we make the move. In the mean time you can still deposit (or more likely) withdraw your money

Yeah, I never meant without multisig. It's actually exactly how GreenAddress.it already works, and they already refuse to sign a transaction unless it meets certain requirements, and give you the time locked release transaction.

Someone should contact them and ask them to create a "vault mode", where they will refuse sign the transaction releasing some money up until a certain amount of time.

Took me a while to find, but it was a very interesting thread. It seems you proposed it during the time when investors lived in fear of Nakowa and wanted to be able to configure different risk profiles. I had mistakenly believed I had some part to do with its creation, when trying to explore a good way to allow investors to minimize their counter-party risk, but it was fully explored a long time ago. I apologize about not giving you full credit when it was due.

Bitcoin supports time locked transactions, so the bank can immediately return you an unspendable transaction. This solves most of the issues. Now you don't need to worry about them losing your key, or otherwise never giving it to you.

Yes. Processed over a 1000 bitcoins of withdrawals without issue.