Generally I'm not so much a friend of normally automated escrow services because in case of moderation it is often too late to speak about precautions that should be taken. Moderation then might not be possible anymore really.
Yeah, all fair points. I think the point of automated escrow is to replace *not using escrow*. Pretty much every day someone on bustabit complains to me about getting scammed, and while escrow wouldn't help in all cases (e.g. delayed paypal reversals) the main reason people send money directly without using escrow is: * Escrow is too complicated or technical (e.g. multisig) * Escrow is too slow or too much friction (dealing with a person) * No one wants to pay the escrow fee so bithra is (in-my-biased-opinion) a great solution for the above. However, it will never and doesn't try to replace trustless or human-arbitrated escrow. |
|
|
|
Without public keys how can it be a multisig?
It's expressly not. It has very similar characteristics from a traders perspective as 2-of-2 multisig though (e.g. no outside arbitration) but is much easier to use, as you noted, no public key sharing and signing release transactions. =) |
|
|
|
It looks like you might have it setup so that the release transactions will get broadcast via blockchain.info's node (maybe via blockchain.info/pushtx), and although they are my personal favorite block explorer, they do have a lot of problems, especially DDoS type problems that prevent people from accessing their website (or broadcasting transactions via them).
Nicely spotted. But actually I push to both bc.info and blockr. The code if interested: https://dl.dropboxusercontent.com/spa/rmczv2tqcr196vz/oxrp5rgx.png If running a full node is too costly for your service that you are not charging anything for, then I might suggest running electrum, using the "loan transaction -> from text" feature and then broadcasting the signed transaction via electrum (and possibly also providing the signed transaction that the user can broadcast themselves in case of other problems).
I already run a full node for bustabit, so I could push it to that node with a few minutes work. I actually store the transaction (and the release signature) before it's pushed in the database, so it's a TODO: to expose that to the user as well. I might also suggest that you require some kind of cryptographic confirmation that the release secret (what appears to be a private key generated by your website) was actually backed up or that the user can in fact sign a message from the address they say they will sign from to release the funds prior to generating the escrow address. This will prevent someone from inadvertently creating an escrow address that is essentially impossible to release.
Good point It also looks like your FAQ might have a typo: ...You can it directly with our website.... I believe that you intended for this to say "...You can generate it directly with our website..." [/quote] Yup, thanks |
|
|
|
I am also curious to know what will happen after the one year period is up
If the funds are never released, they will be forever lost to bithra. They won't be sent to the destination, or returned to sender -- as picking either of them could encourage a type of scammers. It kind of mimicks what happens in a 2-of-2 escrow situation, when the parties don't come to agreement (except in this case, I keep the money instead of them being forever unspendable) |
|
|
|
I tried it out and it seems there is some issue.
Uggh. And thanks! I threw 0.05 BTC into your escrow while you're waiting. Seems like my latest change has stopped the "Advanced" release from actually sending the signature to the server. Will have it fixed in a few minutes. Note to self: Setup email alerts for errors --- Thanks! Stupid mistake, fixed. And I used the interface to release the escrow. (Posting it here, as you made the details public) https://live.blockcypher.com/btc/tx/c5d19e91a26e50a8e5ca23f6b9a8187c389dafec1bfd25126f008047b256ffd0/The fee on that is strangely small, but it's seems to be working correctly (blocktrail's api is telling me the optimal fee is 94.12 bits per KiB) |
|
|
|
Might be the wrong place for it, but I made a free escrow service and would appreciate being listed.  https://www.bithra.com/ (thread: https://bitcointalk.org/index.php?topic=1456828.0) It has no fees, and it's completely automated (thus has no arbitration). It's designed to be the moral equivalent of 2-of-2 multisig, but easier to use (initiated completely by 1 party, and no public keys). |
|
|
|
I really don't like to make enemies, but I've left negative trust based on this. I don't see how bitstars.net is able to justify taking the users money for betting too big when: * it's clear he wasn't advantage gambling * didn't derive his money from "disallowed bets" * the system never stopped him * and rule violation that bitstars.net screenshoted saying the max allowable bet is 0.025 btc isn't even on the website as of now! ( https://web.archive.org/web/20160502231900/https://www.bitstarz.com/terms ) I will happily remove if I'm mistaken or missing something |
|
|
|
Very nice site. I'm impressed.
What happens in the case of disputes?
Thanks! The two parties need to work it out. It's the moral equivalent of 2-of-2 multisig, but designed to be easier to use (no public key sharing, both people don't need to sign the transaction). If you need arbitration or something more advanced, it's not really the service. There was a good discussion of it in bustabit, but basically the summary is that it removes the upside in scamming -- but doesn't prevent people scamming. |
|
|
|
That is only one part of it though, and probably one of the smaller parts. The main reason he should be trusted is because he owns bustabit, and previously owned moneypot. Moneypot had lots of bitcoin invested in the bankroll, and I don't believe a single satoshi was stolen since after the sale, people could access and withdraw their funds at mp.bustabit.com. Bustabit has also had a lot of bitcoin wagered, and large winners are paid out very quickly, with some huge 10+ bitcoin withdraws. There wouldn't be much of a point for Ryan to create a site, just to scam people since if he did, his site (which probably makes him nice profit) would get a lot smaller volume, and there would be less profit.
Yeah, I think at a single point of time I had held >1000 BTC of other peoples money that I could've stolen if dishonest, and in total processed a little over 16k BTC of peoples withdraws. It doesn't really prove I will always act ethically, but bithra.com is just supposed to be a light weight alternative to "Send me the money first", not the be all and end all. I specifically avoided any advanced features and usage just to make it simple. The use case I'm trying to replace is: "Hey, I'm new and selling X. Please send me the money first, I promise I won't scam you. I don't have time to use a human escrow". It's not trying to replace trustless on-chain schemes, and things that offer better guarantees. |
|
|
|
PROVABLY FAIR: We provide non-repudiable proofs for all our escrow services, and is operated by Ryan Havar (owner of bustabit), on bitcointalk's default trust list, and handled millions of dollars of bitcoins. ^^What is this? So, you are basically asking for a favor for being on default trust level 2! Where is the provable fairness in it? Actually it offers a little more guarantee than what is typically called "provably fair". Because in addition to proving to yourself it's fair (like gambling) it allows you to prove to others to prove you were cheated (non-repudiation). The provably fairness is explained in the FAQ, but here's a practical example: {
"address": "1MGeKWdiUiKPBt3NUR7NsmNDsZT5LgycrW",
"message": "The address 12SbYg7CZL9pxkvJqmmFriJXBYXhTiXPep was generated on 2016-05-01T23:12:56.110Z by bithra.com, and holds funds in escrow for 1FDUdzgVhUbSpkbCKST11XoeQ3K5zxac8a and will be released with signed instructions from 14iik8PjwVKGhHPvTD3eJUCdpyR7dfjW2Y in the next year",
"signature": "IBGDmeCNLxx6KBCBjJJkvvWpQwKFAsZd3PKHmkzzPI8gL+6AgOqOxScU/uHKrJoksQ4S5FRcI2iQ9qqHGdGpLPk="
}
You need to verify three things: * The address is indeed the one we use to sign guarantees * The signature is valid ( I like this tool: https://www.blocktrail.com/BTC?verifysignedmessage=1 ) * The details are correct Now you can save that guarantee, and if we don't act according to it -- you can call us out on scamming! |
|
|
|
So with this site, you can basically become an escrow service provider yourself easily, right? (Correct me if I'm wrong)
This is absolutely amazing. I might use this sometime soon.
Note: There's no reason to not trust the site because the owner of bithra is the owner of bustabit too.
Yeah, I act as escrow for all transactions on the site (just don't provide arbitration services). The moral equivalent of using 2-of-2 multisig, just with a nicer UX but worse guarantees. But yeah, there's definitely a lot less counter-party risk in using bithra than trusting some random stranger when trading |
|
|
|
I'd like to announce bithra.com a completely free, automated and provably fair bitcoin escrow. It's something I've wanted to build for a while, with the goal of making escrow so easy, quick (and free) that there's no reason to not use it. Part of making it user-friendly, means I had to forgo some of the more fancy cryptography like M-of-N which would require users to work with public keys -- but I have managed to make it provably fair. It also comes with an API if that's your thing. Give it a try, I think you'll like it =) https://www.bithra.com/ |
|
|
|
This seems pretty cut-and-dry in worthyou's favor, and I plan on leaving negative trust on bitstars unless I'm mistaken or missing something (in which case I greatly apologize). They are breaking their own terms and conditions rules. Here it is:
Bonus was 0.25btc 30% of that is 0.075. 50 bets of 0.06 would not therefore be in breach of the terms.
Everything is so clear. However they are still confiscate my bitcoin.
Hi there, Here you go.  Mike lol. That's what I call bullshit. First of all, when I go check the terms right now ( https://www.bitstarz.com/terms ) I see no mention of it. Secondly, there's already an industry standard for dealing with bets bigger than you want: don't accept them. It's dishonest to accept a bet that you're not willing to honor. For instance, in BaB I'm not comfortable with the variance of bets bigger than 1 BTC. So what do I do? I don't allow people to make them. I don't accept them, cite some obscure ToS violation and steal peoples money. What makes it doubly bullshit, is that according worthyou he predominately made his money from from a bet less than that, and the "violation" was inconsequential and just an excuse to not pay? |
|
|
|
Out of interest, could you share some more information about the faucet abuse? I'm seeing a huge amount on bustabit.com too (which has 3 bits per hour). Do you think it's humans, or someone with a recaptcha breaker? Also, not sure why but someone is faking moneypot.com emails from what appears to be a reasonable large spam (I got two?!) campaign with emails like: The March and April invoices are outstanding, please make a paymentasap. Thank you.
Earnestine Hays
Deputy Director of Finance
(with a .zip attachment) from the headers: Received-SPF: neutral (google.com: 116.106.0.86 is neither permitted nor denied by domain of HaysEarnestine527@moneypot.com) client-ip=116.106.0.86;
I'd strongly recommend updating the SPF headers to change the ?all to -all to kill the spammers |
|
|
|
To the OP: The lag is a distraction. Read the provably fair, they're either provably fair and the lag changes absolutely nothing. Even if each roll took 20 seconds, if the site is provably fair, it wouldn't matter. Scamming despite provably fair is possible.
Example: I bet high every time, with the same client seed. The server now knows I bet high and knows my client seed. The server now choose a server seed which would give a low result. It hashes this and displays the server seed hash.
The server does not change anything after I have bet, but by assuming that I don't change from high to low and don't change my client seed, I will lose.
Primedice actually used to have more issues (maybe still does) where even if you change from bet high/low, there's a chance they can still mess with you.
That would make the server extremely vulnerable. Imagine for instance you knew PD did that, all you need to do is keep re-using the same client seed, and then always betting high. Then when you change your server seed, start making giant, low bets and destroy their bankroll. There's no way a casino would expose itself like that. Or even better, just randomize your client seed like you're supposed to  |
|
|
|
|
I would think it's pretty standard. In general it's probably better from a casinos point of view to process withdrawals as fast as possible, as you want to offer the smoothest experience (especially so to winners, so they come back). On bustabit for instance in the last 3 months we've processed 14095 withdrawals, of which 14092 were instant (the remaining three required manual intervention, due to hot wallet constraints)
|
|
|
|
Provably Fair means that they can give us a proof that they are fair. That means they maybe unfair but they have a proof that they are fair. But since there is no independent organization that oversees "provably fair" I think provably fair itself is a bulshit.
It doesn't work like that, I am independent and can (and have) looked at PrimeDice's provably fair system for instance. I am quite satisfied it is robust, and I have verified several games that I played there. However, the provably fair only proves the games that I verified were fair, not every single game. There's no feasible way for an independent organization to verify a site, and even if there was -- it would require people to trust the independent organization. Which is worse than what we have now, not having to trust anyone. Anyway, as someone who runs what is probably the #2nd biggest bitcoin casino by volume, I'll let you in on a secret. Casino's don't need to, and aren't particularly interested in cheating. Even if I could undetectable cheat, and had zero morals, I actually doubt cheating would make me any more money. Why? Because cheating would decrease the variance for players (the number #1 attractant to casinos), decrease the great success stories that bring in players, and most importantly most players are loss bound, it's just a question of how much volume they turn over before they eventually lose it =) |
|
|
|
|
What "fair" refers to, is that you when you win you actually win, and they don't tamper with the odds. Not that they have "fair odds". That's why if you bet at 2x, you have a 49.5% chance of winning, not a 50% chance. That's how they make their money.
|
|
|
|
2) Does MoneyPot charge commissions if the gambler wins?
Yeah, it's basically a fee you play to accept the bet. You pay the fee regardless if the gambler wins or loses. In the short run, the investment can reach high volatility that can either win or lose the investor a high amount of Bitcoin. Such as the previously mentioned example, in the short term, this could mean that a quick loss of 100 Bitcoin at a 1% House edge would result in as high as a 99.3 profit to bankroll. The commission would be 70% of the house edge on the wagered amount, in this rare case, a 7% commission. It does not mean that commission is 70% and that Moneypot would take 70% of 100 Bitcoin.
Where is the 7% coming from? No where, it's wrong. Even according to Dogedigitals maths it would be (50% + 20%) of the house edge (1%) ... which .7% not 7%. 5) The kelly is not actually not that confusing. I think most investors would understand if you write that "based on a mathematical formula known as the kelly criterion, the most that an investor could lose on a bet with 1% house edge is 1%, the most that an investor could lose on a bet with 2% house edge is 2%, ect"
Agree, but any sane investor should trivially be able to figure it out if it's documented with examples. |
|
|
|
Just a few questions from the perspective of an interested current investor:
1) Can Ryan explain the difference between EV (which I take to be expected value) and expected bankroll growth?
EV = Expected Value. It's the easiest to understand, it's basically the average you'd make over a VERY LARGE amount of bets. e.g. If we flip a coin, with fair odds the EV is going to be 0, even though in reality you'd either have made money, or lost money. But lets say someone offers you good odds, 2.01x if you win, and 0x if you lose. That's a great deal! You should take it, it's +EV. But the question is, how much should you bet? Well, you don't want to bet too little, because you're missing a great opportunity. But you also don't want to bet too much, because you risk losing too much money and will be very slow to recover (if you recover at all). The paper and surrounding stuff goes into great depths to explain why you should optimize for bankroll growth, not EV. So if you plot it, you'll see it probably looks like this:  So risking a 1x kelly is mathematically ideal, and what every casino should do. But reality is not so clean: a) You can't force a gambler to do it (almost most of the time, gamblers only want to bet a small amount and not what you tell them) b) It might be worth taking a few 1-2x kelly bets, because otherwise you might lose the gambler to a competitor c) Very few people actually do large bets, but everyone likes being able to do large bets. d) Lots of investors are gamblers, so maybe they want to risk a wild ride e) Variance sky-rockets with bigger bets. As I remember, Dooglus sanely dropped down from a fully kelly on JD (which had thousands of bitcoins in its bankroll) because of variance concerns etc. So when I said MP is in -expected growth territory, it means its risking even more than that 2x kelly so with an theoretical-angry-whale-attack it should expect it's bankroll to zero! Here's a toy simulator i wrote earlier: var kelly = 1;
function roll() {
return Math.random() > 0.495 ? 1 : -1;
}
var wins = 0;
var loses = 0;
function run() {
var bankroll = 1;
for (var j = 1; true; ++j) {
var bet = bankroll * 0.01 * kelly;
bankroll += bet * roll();
if (bankroll > 1e40) {
wins++;
console.log('Won all the money in the world. Won: ', wins, ' Busted: ', loses);
return rerun();
}
if (bankroll < 1e-40) {
loses++;
console.log('Busted! Won: ', wins, ' Busted: ', loses);
return rerun();
}
}
}
function rerun() {
setTimeout(run, 0);
}
run();
Sample output of running with a 1x kelly: ...
Won all the money in the world. Won: 189 Busted: 0
Won all the money in the world. Won: 190 Busted: 0
Won all the money in the world. Won: 191 Busted: 0
...
Sample output of running with a 2x kelly: Busted! Won: 38 Busted: 83
Won all the money in the world. Won: 39 Busted: 83
Busted! Won: 39 Busted: 84
Busted! Won: 39 Busted: 85
Busted! Won: 39 Busted: 86
Busted! Won: 39 Busted: 87
Busted! Won: 39 Busted: 88
Won all the money in the world. Won: 40 Busted: 88
Sample output of running with a 3x kelly: Busted! Won: 0 Busted: 191
Busted! Won: 0 Busted: 192
Busted! Won: 0 Busted: 193
Busted! Won: 0 Busted: 194
Busted! Won: 0 Busted: 195
--- So when I wrote MP, it was running at a 1x kelly for obvious reasons. Now it's apparently at a >3x. Which is fine, but I doubt investors realize what they're getting into. Probably a more honest name would be "+EV gambling", even if the site makes thousands of bitcoins (Which is very well might) it's still not (and can never be) safe with that risk level |
|
|
|
|
Show Posts
