Bitcoin Forum

With the leak of the code I'm asking myself whether it would be a good idea to setup a bitcoinica clone.

Pros :

• Would make some good money
• There is demand for gambling margin trading
• Half the profits could go to the people who lost money because of Bitcoin Consultancy/Bitcoinica/[insert scapegoat here]
• The app itself has never been broken into (even though it's quite surprising when reading the code)

Cons :

• Needs moar reverse-engineering/fixing
• Needs moar testing
• ...

Thoughts ?

I would suggest that everyone runs their own clone localy.

Could you post what Mac IDE you were using in your original post as well as any other steps required to make it run on a freshly out-of-the-box Mac?

I think that way many more people would join you in reverse engineering.

The IDE is Rubymine, it's the bomb.

If you're a Rails coder you won't have much trouble getting it running, you basically need to install the dependencies with bundler, launch a couple moving parts (bitcoin client, redis, resque, clockwork) and you're all set. Oh, and you need to create a MtGox API key and feed it to the code.

I am concerned about security holes in the code. Web can be a scary attack-surface. I was thinking it might not therefore be a good idea even if I disconnect it from the money, using it just as a glorified calculator to tell me how much of what to move where and do the actual handling of the funds in Open Transactions.

However I am many years out of date on secure coding of CGI scripts and never did move on from securing CGIs to securing PHP. (And have never touched Ruby.) If someone who is really really up to date on how (or is that still "whether") PHP or Ruby can be secure can fix it up for security I expect to see lots of little MyBitcoinica sites spring up.

The suggestion of using it as a local app behind one's firewall ("everyone runs their own clone locally")  sounds cool but presumably loses the ability to bucketshop the players off against each other thus losing what is likely a huge part of how it makes money. Actually come to think, if you are the only player on your instance who are you to bet against?

I was afraid the thing would be too easily attacked thus maybe not worth in effect paying $350,000 for the code in the form of making whole the victims of Bitcoinica.

If the code is really good or can have its holes plugged pretty certainly then maybe Gox themselves might find it a good springboard to adding the functionality to their own suite of services.

-MarkM-

Hmmm maybe altcoinica so peeps can get the margin devil out of their system without risking too much?

Code is not inherently insecure.  Coders practices create insecurities.

There is absolutely a need desire for a bitcoinica clone.  However based on recent events I believe the community would demand (but maybe not!):

1.  Clear declaration of ownership
2.  Proof/validation of security practices
3.  Regular third party financial audits.

That wouldn't be inexpensive however the evident profit margins of such a business would certainly warrant the expense.

I'm not "Rails coder" and most of the people here aren't either. But we are all curious and capable of coding or at least reading code. I would suggest that your target audience shouldn't be a "Rails coder", but any curious hacker not afraid to code.

Since you've mentioned Mt.Gox API key I have another question: how difficult is to stub out the portion that queries Mt.Gox and instead plays back a stored ticker tape in the csv or any other spreadsheet-compatible format.

Many people are interested what the secret pricing algorithm did to the bid/ask spreads on Bitcoinica. This could probably be of paramount importance if any litigation comes to fuition.

And thank you for the Rubymine reference.

I can assure you this one is.

Yes indeed. Being able to short any coin type against any other coin type would be very interesting and is where I will be heading with Open Transactions if I can actually figure out how shorting is woth implementing at all. I am still not clear on why anyone would loan an asset to someone else knowing the someone else is going to use it to lower its value so as to ideally be paying back less value than they borrowed...

Also I worry about how to ensure there are enough on the market to buy come payback time. It seemed to me there is a double cost in shorting if secured by a marketmaker who puts aside enough of what someone borrowed from someone to ensure there will be enough left on the market at payback time for the borrower to buy to pay back the loaned assets.

The more "alt" the asset the more need, it seemed to me at first glance at least, there is for a marketmaker like that because if there are none of the borrowed thing on the market there is no price available to determine how much value needs to be paid back in some different type of thing.

(If you short 1968 Coupe de Villes, and there are not Coupe de Villes on the market come payback time, how can anyone figure out how much dollars or gold or something you'd need to pay to make up for having defaulted on paying back the loan of Coupe de Villes?)

-MarkM-

I know NOTHING about ruby... but I don't doubt your assessment.  Yahoo Voice stores passwords in plaintext so don't event begin to think that any larger entity's online security practices are any better.  The bigger the entity the more money that have to throw any security breach when it happens and for now it seems that is cheaper for them then rolling out fully secured code into production.

Understood, if you want to read the business logic it is in the app/models folder (each model is linked to a database table), and in the app/workers folder (this one contains all the periodic running tasks) the pricing logic probably resides in one of the workers.

I don't think it would be extremely difficult, Ruby is awesome for this kind of things, you can transparently replace code at runtime.
See pastebin below

I've added a couple comments, but this is pretty much the raw TickingJob http://pastebin.com/4Ej858a8

I love it, and JetBrains gave me a free open-source license for developing Bitcoin-Central.net :D

No, for reasons like this. (https://bitcointalk.org/index.php?topic=82581.msg911141#msg911141)

From my understanding there are actual trades happening when users place orders, if they have a long position actual BTCs are bought for example.
I'm not really far in understanding how it actually works, until then, calling it a bucket shop (or a legitimate trading platform for the matter) is IMO a little premature.

Hmm so they were a bucket-shop but just not a totally self-contained one? The use of MtGox was just a way of helping keep the bucket shop from getting caught out a little longer by at least sometimes actually buying or selling actual assets?

I have been testing Open Transactions for a year or so now and am very pleased with its progress so I am very interested in figuring out a "right" way of implementing leverage and shorting. But if "doing it wrong" is what the customers want, as evidenced by their flocking to those who "do it wrong" to avoid the costs involved in "doing it right" maybe having floods of MyBitcoinica sites springing up all over the place is going to be inevitable?

-MarkM-

Doesn't mean it's desirable :)

Ah, OK then. I was worried that you are going to waste your skill and time on doing the security audit on the cadaver.

Zhoutong's key invention was the secret semi-random pricing algorithm. I recall that somebody reverse-engineered his original pricing algorithm and posted a Python script to reliably scrape small profits (maybe that person was "macbookair"?). Then he added some randomness and it worked like catnip on all the gamblers here.

I would let other people do the necromancy and create Frankencoinica, Litecoinica, Solidcoinica and whatever else.

Actually it will probably be Биткoиницa.

How was the pricing algo an invention ?
From what I thought I understood it was simply the mtgox price at a certain depth.

The original code surely belongs to someone, regardless of whether it's been leaked.

Though given the leak it looks trivial to reverse engineer and write a clone from scratch.

Yep, so that's basically a non-issue :)

NOO! You must use emacs!  >:(

This +1. Waste your damned altcoins on insecure code, and wait for a secure alternative to be built and audited before using it on bitcoins.

I don't know the details either. From what little I understand his modified algorithm would sandbag the spreads in such a way that the loss was always on customer's side not on the house's side. The amount of sandbagging was somehow keyed with volatility.

I distinctly remember somebody somewhere posting a Python code to whipsaw bitcents from the original algorithm and then Zhoutong came up with a fix for that.

I think more services = better, so you should go for it. There is always space for alternatives.

I understand that you are French and the bucket shop moniker is very much rooted in the American tradition.

But the line of reasoning is really simple. You don't really have to disassembly any perpetuum mobile to say that it can't work.

1) Long trade has a range from unlimited gain and 100% loss.
2) Short trafe has a range from 100% gain to unlimited loss.
3) To honestly allow shorting you have to do a credit check on the customers.
4) Bitcoinica wasn't doing credit checks and didn't attempt to collect on loses.
5) From the above Bitcoinica was a gambling establishment.

Call it "bucket shop", "bunga-bunga shop", "spread betting", "contract for difference" or whatever else the gambling regulations allow throughout the world. The logical outcome is really simple.

Interesting.

My understanding is that they did not require credit checks on people because they would reimburse themselves by force-liquidating your position once it was getting close to net_worth = borrowed_amount

In my understanding how it works is the following (haven't checked much in the actual code so far) :
 - For a short position you borrow BTC from the house. If your own BTCs are N your leverage 2, you end up with selling 2*N BTC at price P1, you're now holding 2*N*P1 USD, if price falls to P2 you can liquidate your position with a buy order, you buy (2*N*P1)/P2 BTC. If P1 > P2 you get keep a profit, if P1 < P2 you're at a loss. Your position is force-liquidated when P2 rises so much that re-buying would leave just enough to reimburse the house.

 - For a long position it's pretty much the opposite. You're just shorting the USD.

And since (always in my understanding) you liquidate positions with market orders there is always enough to pay profits.

People can approach this as gambling, or trading, doesn't matter, as long as the rule is clear I see no malice in that. BUT my understanding of margin trading concepts is still fairly limited, I'm genuinely curious about it, and ready to stand corrected.

I don't see a credit check being a necessity if you hold assets of theirs. Basically their credit is equal to the amount of their assets you have in hand to draw from to cover their losses.

I thought that is what margin calls are all about: any time they seem to be running short of collateral you tell them you will close the position unless they deposit more collateral.

You can limit the losses by never lending more than half your stockpile of an asset.

For example if some LiTeCoin bull wants to short DeVCoin, you only lend him, at maximum, half of your DeVCoins; the rest you put on the market at a high price to guarantee there will be DeVCoins for sale for him to buy come time he is to repay the loan.

Or, you can stretch more by buying what he is shorting. Loan him DeVCoin to short, and buy those same DeVCoin from him as fast as he puts them on the market. If you put back on the market, at a higher price, each and every DeVCoin that you loan him, you can loan him every DeVCoin you started with while still ensuring he will be able to buy enough to pay you back at any time.

So provided you have enough DeVCoin to loan him, and never loan him more than can be bought back with the number of LiTeCoins of his that you have in escrow, all should be good. You get to choose, by placing on the market as many DeVCoins as he owes you, the limit of his losses; you simply do not loan him more than you are willing to buy his escrowed LiTeCoins for.

Leverage of course complicates matters but even with leverage you know how much collateral he has and you know the max price he will have to pay to buy back the asset that he borrowed (because even if no one else will sell him any you stand ready to sell them to him yourself).

Thus far I do not see a big problem, as market-maker you are always in a position to know exactly how much you will charge him for what he is short if no one else will sell it to him, thus you are always able to limit his loss.

-MarkM-

To my understanding, when you can force liquiditations, there is much smaller risk to the service runner than in traditional lending. The money in the table is the funds that the customer has.


I agree with that. However I don't think that credit check is necessary for this kind of site, if you create the necessary precautions (forced liquiditations).

Since the code was leaked, it would probably not be legal to use the exact same code to set up a new site. I am not a lawyer, but I don't think it would be legal to use proprietary source code retrieved through non official ways to set up a bitcoinica clone.

However, if there was a team of information security professionals (with actual degrees and stuff), such a site could be a good idea. I hope that everyone by now has learned that a site created by a 17 year old in 4 days (ignoring the fact that it, isolated speaking was a good achivement), isn't something that we should put our trust in.

But of course, seeing as nobody seems interested in calling the cops, who knows, perhaps somebody setting it up could actually get away with it ?

As far as I'm concerned I downloaded code from someone who voluntarily published it, and there are no copyright notices anywhere in it.  ;D

I'm finished with margin trading myself but I say do it,  and from my view in the stands it seems you're a more capable programmer so hopefully the security issues could be prevented with more competent admins.

I think that everyone here should first try to add margin and shorting priviledges to your regular stockbrokerage account. The forms you'll have to submit have a very detailed explanations for the collateral requirements and how that affects your allowable leverage. There also will be the explanation of charges you'll have to pay for the capital you've borrowed to execute your margin or short trade.

In an honest trading establishment even 1:1 leverage requires credit check for shorting. If the lever is more than 1:1 then by definition the customer didn't put enough capital up front.

In fact, even though the code doesn't seem very secure it is still a fact that the only flaws that led to the thefts were organizational and not technical.

Oh, and BTW, if someone actually contacts me with a C&D notice we'll know who actually owns the IP :D

I like this project.

My understanding is that Intersango LTD ( or Bitcoin Consultancy LTD, Bitcoinica LP, Core Credit LTD) owns the code, maybe they will be kind enough to Open Source it?


I agree with this comment. Forced liquidations would reduce the risk for the exchange/broker and make credit checks unnecessary.


I agree.
If, for whatever reason, the owner(s) refuse to release the code, you can always make a better clone including the options to adjust the authorized leverage per contract (BTC/USD, BTC/EUR, BTC/AUD, etc.) and add/remove contracts based on liquidity and/or market demands.

Hey I didn't know that if you run a website, the server side code is automatically visible to everyone..

I mean, please. Not only is the notion of IP completely ridiculous but how are they ever going to know if someone changes the frontend a bit but keeps the server side code exactly the same? It's impossible to enforce.

This genie is out of the box and no one is ever putting it back.

For me, this is the most interesting part about having their source code now, we actually get to see exactly how their exchange was operating:


Interesting.. can you find out how P1 and P2 are determined?

Don't you know that everything is automatically copyrighted?

Aren't copyright notices required or else information is in the public domain? If there are no notices then from my limited knowledge I would guess it was in the public domain legally.

brb getting server

According to US law...

Q. Do I have to register with your office to be protected?
A. No. In general, registration is voluntary. Copyright exists from the moment the work is created. You will have to register, however, if you wish to bring a lawsuit for infringement of a U.S. work. See Circular 1, Copyright Basics, section “Copyright Registration.”

From http://www.copyright.gov/help/faq/faq-general.html

Nope everything is IP automatically. For example if I come up with a new word that word is my property and if you use it without my permission I'm going to sue you for damages so that a gang of men in blue costumes can come and force you under threat of violence to pay me royalties you should have paid me in the first place, because I said you should because I was the one who thought up that word.


 ::)

I'd love a C&D notice, I'd publish it here so people know on which door they should knock to get their money back :D

I support this method. Well played, Monsieur.

Placing a copyright notice is not the same as registration. I found this for US law:


http://www.copyright.gov/help/faq/faq-definitions.html

UK law: http://www.copyrightservice.co.uk/copyright/p03_copyright_notices


So I was wrong but it had nothing to do with registration.

What is the main difference between Bitcoinica and other trading platforms for you ?

Recently we finally opened futures trading on http://icbit.se , so I would like to know what features are missing there.

How is it not obvious to you? Do you not see all these people waiting for their money? Do you not see all these people who put their trust into an organisation getting burned?

Can you really not figure out that the feature you're missing is transparency and security and some sort of contractual obligation?

I like what you did there.  ;D

Margin trading itself isn't gambling in my opinion (yes, we can argue this forever), I like to use certain references when I have doubts:
1. The Prudent Man Rule (http://en.wikipedia.org/wiki/Prudent_man_rule): Harvard College v Amory (1830) 26 Mass (9 Pick) 446
2. Internal Revenue Code (IRC) (http://www.irs.gov): Section 165(d),income-tax deductions of gambling losses.
3. Various US Federal & State laws distinguishing games of chance versus games of skill.

With that out of the way...

I believe margin trading offers the best means (so far) to establish a bank offering interest on deposits for the Bitcoin community. Where the establishment's profit marging is (margin interest rate) - (deposit interest rate). Notice from the posts on Bitcoinica threads how people were using the site as a bank.

Yes, I agree, I brought that up in a "Shorting" altcoins (https://bitcointalk.org/index.php?topic=93356.0) thread in the altcoins section of the forum.

-MarkM-

In addition to hazek's comment, I would like to add:

Options & Futures are derivatives (http://en.wikipedia.org/wiki/Derivative_%28finance%29) (offered at icbit.se).
The Bitcoinica platform offers margin (http://en.wikipedia.org/wiki/Margin_%28finance%29) not derivatives.

As far as I know, margin is required to trade derivatives but derivatives aren't required to use margin.

Yeah, let's do it with your liability  :D

I like the way you think! :)

Although I doubt they would actually serve a C&D letter. I understand the new Bitcoinica owners found so many holes in the current code that they most likely decided to not to maintain and and instead develop a new platform from scratch. Unfortunately Bitcoinica came tumbling down before the new platform was ready. (not that it would have mattered because their non-code related security practices were lacking/non-existent)

The more I read the code, the more I see it is very very messy. I think the reason they'd want to rewrite it from scratch would have been for maintenability, not security.

In addition to the above comments, I would like to state one very important fact which makes futures trading platform like icbit.se superior to a bucket-shop marginal trading one like Bitcoinica:
When you trade futures, the market is open and decentralized. There is no single point which calculates spread and decides what price you can close your position at. The market decides, its hundreds and thousands participants. It's way more honest and reliable method. Everyone could be market maker and earn money by providing liquidity to the market.

So why recreate a clone of something which is a "previous step" already?

As for margin. Trading with leverage is an option but not a requirement. Margin trading must be used with caution, if you're provided with 1:4 leverage it doesn't mean you have to buy 4 BTC having only 1 BTC in your account, otherwise you risk loosing all money due to a margin call situation.

Well, partially. I understand that losing money is not cool, but how deploying a clone of that system can return their funds ?

Because it would generate profits that could be, at least partially, allocated to the compensation of the victims.
That was my original idea.

I'm not sure however that I like the way Bitcoinica actually works (based on my first analysis of the actual matching code see http://pastie.org/4257541) but the reverse engineering task is not finished yet.

You just don't get it do you? What I was trying to get across is that your exchange with regards to features it offers is probably equally appealing as was Bitcoinca and that if you really want to entice more people to use your service, new and better features aren't what you should be thinking about given the current post Bitcoinca fiasco environment but rather transparency, security and contractual guarantees.

Maybe profits maybe losses. Now that the previosly secret algorithm is in the open traders could tune their strategies off-line with no capital layout. It isn't immediately obvious that there is no algorithm that could push the resurrected Bitcoinica into red.

You are one of a handful of people I would trust with non-trivial amounts of Bitcoin.  Cannot quite remember the reasons why though.  After a few hours of piddling with Bitcoinica I lost interest.  If I do such gambling again, I have enough stock to risk my own funds and not pay the premium so that's what I would probably do (vs. using some Forex style platform.

As for trying to make Bitcoinica's victims whole, I see utterly no point (and I'm one to the extent that I planned, BTW.)  I'd rather see MyBitcoin folks receive a little relief as they were just trying to find a convenient way to store their wealth vs. trying to capitalize on one another's misfortunes.  It seems to me to be much more in the interest of Bitcoin to designate any profits toward future legal and lobbying efforts and that sort of thing.  Were that happening reliably and transparently, I might be tempted back into the casino should I find myself with some time and money to kill.

I picked up the domain Litecoinica.com just recently, and I think its a worthwhile idea to not only resurrect Bitcoinica, but to do something like it for Litecoin as well.

Would it be possible to hold bitcoins and use them to trade in a way like how blockchain.info has their wallet service setup? This would put everyone who deposits money in control of their own money in case the site goes down. If it is hacked, the private keys could be revealed so you would need to do the interaction with the wallet client-side. The downsides to this are that if the site goes down, the money in open positions (which would have to be moved from the clients wallet to a communal one) could be lost, and it would lead to massive bloat on the blockchain. To be effective (if ignoring bloat) the system would have to rely on 0 confirmation transactions.

lol well played sir.

No.

FTFY

Exactly - I really enjoy the price stability of Bitcoin since Bitcoinica is gone. These days, Bitcoin can actually be used for payment.

this. the hell with bitcoinica and all its progeny

^^^^ This.

Second:  why not explore a GLBSE company?  Allow people to buy in plus that way there is an easy way to motion and what not. We would need to agree here how many shares you would get for the work you have already done and the hosting and whatnot.  I am also a rails guy, but not sure what time I would be able to commit to yet. 

What about the algorithm being public? Would this affect profitability?

I thought this through, and came to the conclusion that the concept needs to be re thought.  We should try to adopt a new principle here in Bitcoin Land: If the service can be run without the vendor holding your coins, it should.

I thought about whether there was a way to do this with escrow.  Everyone long bitcoin sends margin to an account, each short does the same.  The accounts do some kind of escrow thing, where they send their coins in equal proportions to holders on the other side.  You would need a trusted person to do this, that individual is sort of like a market maker.   Their job is to police the margin and settled closed contracts.  <insert details here>

In the end the worst that could happen, were the vendor or market maker to go poof, would be the coins defaulted to all the contract holders as if the trades didn't happen.  That way the liquidation goes to all the participants and any lost funds relative to the true position value only affects people trading through that market maker or "desk."  You could have competing desks, and new desks that let you trade things like oil and gold.  Commission would be split between the market makers who manage escrow of the funds, and the parent company (exchange) itself simply provides the signed code/server services.

Is there some one who gets the protocol enough who can figure it out?

For a better explanation of how it worked, check this tube: https://www.youtube.com/watch?v=kdvTkddp1F0

I think it was that:

Instawallet is/was one of the few designs which I trusted as it made no pretenses of mining for more personal data than it needed to do it's job.

JAV seemed like a straight-up guy, and (iirc) handed things off to you.

Last I checked, my 50-ish BTC was still there.  But that was many months ago, and I'm only going by what the UI told me.

Thank you!

I'm more and more puzzled what makes people think Bitcoinica profited from force liquidation.
Force liquidation just protects them against losses.

The business plan of Bitcoinica seems to be quite obvious: they earn the spread. As long as people continue trading, this generates a constant stream of revenue, with a very small risk. (The only risk was that the market might move to fast for force liquidation to protect Bitcoinica from losses). And, moreover, the spread was so large as to make day trading not really feasible. You could make quite nice revenue though, when staying in for the more long term moves.

Thus, additionally we may conclude that the business plan of Bitcoinica was to attract every newbie in the bitcoin community, just to learn the hard way that the spread is to high for day trading. The fact that a lot of people hated Bitcoinica and constantly warned about "being zouthonged" just tremendously helped with that business. It helped hightening the thrill.


IMHO we should take Bitcoinica for what it is/was, and not condemning it for what it was not. As far as I can see, they indeed forwarded the neted amount of all positions into the real market. So it was not betting against their customers, thus not a bucket shop.

Yep.

It just so happened that they benefited strongly from force liquidation. For all its life BTC valuation oscillated back and forth making force liquidation very profitable for them.

Valuation swings remarkably less now without BTC, it's a fact. A rally like this one would consistently correct itself with Bitcoinica in play, I doubt it's a coincidence, it happened many times.

I don't think it's a stretch to say the market is not liquid enough to bear a leverage trading platform the size Bitcoinica ended up having, without serious disruption.

I agree that Bitcoinica made bitcoin more volatile, but it only profited from forced liquidations if you assume the volatility made people trade more.  Force liquidations only recovered their lent funds since the loss had grown to the size of the available margin.  Bitcoinica profited only from the spread.

Trade more? You mean gamble more. People trade less if Bitcoin is volatile. Merchants are not in the gambling business. You don't sell stuff for Bitcoin if there is a real risk that Bitcoin looses 20% on the way to the exchange - even if there is a chance that it gains 20% at the same time.

Volatility is poison.

No, it was for security.

Also Bitcoinica was losing money because its hedging algorithm was too simple, and during the last few months the site had negative profit. My role in the new Bitcoinica was to design a new hedging algorithm. The old algorithm simply had barriers that when reached would begin trading. These values were set manually on the site. I was designing better ways to assess the risk, and reduce it to acceptable levels (3 sigmas or more).

And this is why:

https://i.imgur.com/rmd6Z.png
From: http://bitcoinmedia.com/market-volatility-has-gone-down/

Local volatility in the market has gone way down. In the beginning, Bitcoinica could have a non-optimal hedging algorithm and still profit. But now things are getting harder as the Bitcoin economy gets sharper.

Thanks for posting here genjix.  I noticed you didn't address either the 'should it go back online?' nor the intelectual property questions.

Saying it was unprofitable kind of suggests not putting it back online unless davout likes losing his money.

Provided he was in the middle of a total rewrite and the hedging algorithm would be different, why carry the Bitcoinica name? it's a drawback at this point.

I'd probably run a completely new margin trading platform with my own people, once the liabilities are settled.

Of course the Bitcoinica name would be changed for a clone to come online, but first I want to :
 - finish the reverse-engineering completely (which means understand every subtlety in the algorithms),
 - understand exactly how it can be properly monitored (the business side, not the technical side),
 - do an actual security audit (the things I tweaked already were pretty obvious, now I need to go more in depth),
 - have a solid plan for a GLBSE IPO,
 - have a plan for the C&D letter case,
 - start writing my own platform (because the code is not written in a maintainable way, and because I'd have a good incentive to write a fuckton of tests)

If you think you can manage the financial exploit side of things (not allow anonymous accounts to have up to 10:1 leverage [leaving them with negative acc balances in some cases], and forbid 5:1 and 10:1 in general), then I am all for it. Please do it.

We need shorting back.

Aye. Could care less for trading on margin and interest on deposits and open positions.

How come mt-gox does not allow short selling? Financial regulations?

Uhhh, shorting in this sense is trading on margin. Interest was there to attract capital, but it shouldn't be needed for a well-funded setup.
Without margin, you can "short" by simply selling into the market.

I am not sure that I would agree with that.  Many people may have a desire to have a negative balance of BTC which is perfectly legitimate.  In that case, borrowing from your own stash is not possible.

At this point I _want_ to own N BTC.  I can short by borrowing (from myself) and selling to the point where I have fewer than N and plan to buy them back when the price falls.  If it doesn't fall, I lose...if I my N target remains valid and I buy in to achieve it.

I believe that 'margin' commonly implies 'leverage' (using fewer notional value of one asset to control a higher value of another) but I am not certain of that.

Back to Blitz's original; who is 'we'?  You could probably find someone to loan you some BTC if the price is right.  If I were to do so, I would want pretty solid protection up to about $100/BTC (where I plan to sell a certain fraction of my holdings anyway) and I don't have any automation set up (to buy in in an emergency) so I'l probably want, say, $10,000 to loan you 100 BTC.  Kind of negative leverage if you will.  I would be surprised if you jumped on the deal...and disappointed since in reality I have no interest in dealing with all the hassles.

Right, because you are borrowing (on margin) in order to use a ratio-based leverage.

Anyways, it's what the market wants, so it should be done even if some people don't like it.

There is the potential for leverage any time you sell other people's currency with none of your own to back up the loan.

Example:
Sell 1 BTC @ $9.25 with $10 margin.  BTC rises to $12.  You now owe 1 BTC (worth $12) and your margin is only $10.  You are now leveraged.

I suppose a system could be set up that force liquidated before allowing leverage, but I doubt you would have any customers.

FWIW, I personally am fine with any trading platorm/scheme which anyone comes up with.  In a non-regulated market (of which Bitcoin is a prime example...and happily so) there is room for significant price manipulation.  But I feel that over the long haul, the markets will find their natural place and I happen to be in it for the long haul.  In the interim period, a person who is reasonably adept at analysis, has some observation skills, and knows what they want can capitalize greatly on things.

I don't buy the assertion that manipulated value fluctuation detract significantly from Bitcoin's usability as an exchange currency.  It just opens possibilities for people to invent platforms which mitigate the impacts.

I think:

1) That in a typical trading platform, 'leverage' (and 'margin') is achieved in a much more straightforward manner.

2) That in a typical trading platform, not only is it common for the operator to protect himself/herself against loss by using forced liquidation, it is part and parcel to a successful and profitable design.

2-comment)  Bitcoinica had absurd fees in part to (attempt to) achieve #2 given the fierce volatility which appears in the Bitcoin economy from time to time.

I would personally improve on the current system with better ideas. Most anything online is available to the right person with "abilities". Taking away the temptation is a better idea, to me, anyway.

Go ahead, https://bitcointalk.org/index.php?topic=93583.0

"We" is us, as in Bitcoiners, who would benefit from more efficient price discovery and better liquidity in case a shorting mechanism becomes widely available again with a low barrier to entry.

Yes!!! We need a new bitcoinica, because I wanna short these things so I can make some money...

If someone is interested in trying out what I've been working on, send me a PM.

You've sure been quiet lately, I'm assuming you have put a lot of hard work into this. Beta testers, let us know how it goes! ;D

Do not forget to have enough funds to push the market either way to make people who are leveraged lose their money. Then return the BTC to the market price.

AKA The Manipulator

The house doesn't make more money when the liquidation of a position is forced as opposed to triggered by the user.

Say I am shorting 10 BTC at 10:1 at the price of $9/BTC.

At $9.5 I am forced to liquidate.

If the price jumps up to $9.4 then returns to $9 I am still sitting at the same 10 BTC.

If the price jumps up to $9.6 then returns to $9 I have zero BTC.

Who gets those 10 BTC when it jumps up to $9.6 then back?

The house simply gets its loan back by buying back when the price rises when liquidating a short.
Actually, the more money you make, the more money the house makes, there is definitely no incentive to force-liquidate positions.

I do not think bitcoinica actually sold a bitcoin for every bitcoin shorted. They would internally say they did and only sell when necessary.

Is this the source code used under bitcoinica.com exchange: http://www.reddit.com/r/Bitcoin/comments/w6xen/bitcoinica_press_release/ ?

Cheers!
Thiago

Yes absolutely, in the previous example I assumed there was only one open position, but with multiple open positions it's a different story.

if bitcoinica was "just a shady bucket shop", then they'd just pretend to open positions, but in fact they only trade for their own profit and bet against their customers. Many people assume that was the case.

On the other hand, if bitcoinica was a typical Contract for Difference offer, then they would automatically forward the net balance of all positions to the real market. In this case, they would not bet against their customers, rather just earn the spread. Zhou Tong allways claimed that this was the case.

The leaked mails show that this claim seemed to be mostly true, while at times Zhou Toung played with the "hedge factor", i.e. only reflected a fraction of the net position to the market (if I understand correctly). For example, when seeing that the majority of the positions was short, but the market moves up, the operator of the platform could reduce the hedge factor (i.e. start preliminarily to buy back some positions). To that fractional part, this would then be an additional gain of the platform, and go against the own customers. But it could also be a protective means to prevent slippage.

I briefly reviewed the hedging code a couple of days ago. The way it is written it indeed opened sell/buy positions on Mt. Gox if the mismatch between existing Bitcoinica positions crossed a certain threshold.

There is a hardcoded variable, I think it's called confidence level (too lazy to look at the code now, but it's the same name Zhou mentions in the leaked emails), that influences the threshold for opening a position on Mt. Gox.

The hedging algorithm as well as broader hedging design is very simple and I can understand why it, according to Genjix, might have caused fluctuations in profitability.

How about piratecoinica?  You can trade on leverage AND earn 7% per week on deposits!