ICBIT Derivatives Market (USD/BTC futures trading) - LIVE

[Fireball]

newminer is one of the traders whose short position was partially closed due to non-paying long-position customers, according to https://icbit.se/margincall worst case scenario

However, I would like to bring attention to the fact, that despite the USD/BTC fall was rather big again, ICBIT handled it much better due to high liquidity, and only small portion of OI had to be forcibly liquidated. Everything else was done using traditional margin calls against the market, so it went rather smooth.

I understand that it always makes traders sad, however, it's much better than the previous time, so we are moving along the good direction. Also, icbitr2, boomerlu and other traders chat participants are providing very useful (and emotional) comments which are going to be wrapped into proposals for actual implementation.

Thank you for trading with us!

[1713973702]

1713973702

[1713973702]

1713973702

[1713973702]

1713973702

[1713973702]

1713973702

[boomerlu]

After getting some distance from the events, I can boil down my issues with it to the following:

Fireball made decisions about liquidation:
1) Unilaterally
2) Without consulting the members with positions potentially at stake.
3) Without clear documentation as to procedures
4) Most critically, with no downside at stake.

The actual problems with the liquidation itself we addressed which are basically two things
5) Too long of a wait to actually liquidate (a full session)
6) Unfair assignment of liquidation to "winners"

5 and 6 we addressed. 3 I hope will be addressed with pseudocode here.

1, 2, and 4 are outstanding issues for me.

I don't think Fireball was being malicious, but the lack of downside (worst thing that happens is he liquidates everybody and walks away) can make for negligence. As a short, there is NO way I would have accepted a full session of time to make margin call - it's a free option to the long (if the market goes up he wins, if it goes down, he gets closed out at a guaranteed price). This is immediately OBVIOUS to me because I have downside risk.

To me, if Fireball continues to maintain (without appropriate downside risk) unilateral control over settlement, clearing, etc procedures more mistakes like these will happen. Now we can't completely avoid mistakes, but we can avoid many of them.

For instance, this one could have been avoided say... if Fireball called one of us when the counterparties were in default and asked for our opinion. Instead he acted unilaterally with no real consultation and no say from the traders. Again, I don't think it was malicious, but he can't lose money on his decision so how can it be precise?

The root problem is we have the decision maker disconnected from the downside of his decisions. There are two possibilities - move risk onto Fireball, or move some decision making power away.

I don't think it's right to impose risk on the exchange itself, so the solution to me is to have some kind of membership structure where the primary stakeholders in the process make the policies. Ideally, interests from both long and short side are represented.

[newminer7950]

I brought the calculations (above) showing that for 3 sessions ago that do not fullfilled margin reuirments of  Exchange. But contracts are not covered. Since then, the price dropped significantly by the time the contracts are closed.
4 sessions to close contracts - this is a very long time. It is not normal that the loss of such a delay shifted by a few traders.

[maqifrnswa]

I don't think Fireball was being malicious, but the lack of downside (worst thing that happens is he liquidates everybody and walks away) can make for negligence. As a short, there is NO way I would have accepted a full session of time to make margin call - it's a free option to the long (if the market goes up he wins, if it goes down, he gets closed out at a guaranteed price). This is immediately OBVIOUS to me because I have downside risk.

I believe the free options given to accounts not in good standing are going to be eliminated.

However, the existence of that free option introduces a trading exploit strategy:
Make two accounts
Sell/Buy from/to yourself from the two accounts.
Remove everything except the minimum margin requirement from the account
Get margin called
One account gets a free option, but hope that you get liquidated.
One account gains linearly with price, one account losses is capped at a price
You make free money (at other trader's expense)

I am of the opinion that if an account goes below 100% margin, they get a warning to bring it back into good standing. If it goes below 75%, it is liquidated at market to prevent this exploit

[Fireball]

However, the existence of that free option introduces a trading exploit strategy:
Make two accounts
Sell/Buy from/to yourself from the two accounts.
Remove everything except the minimum margin requirement from the account
Get margin called
One account gets a free option, but hope that you get liquidated.
One account gains linearly with price, one account losses is capped at a price
You make free money (at other trader's expense)

This was discussed already a number of times, however the answer is no: At best, you loose only trading fees and end up with non-market neutral position. It is impossible.

[random_cat]

I don't think it's right to impose risk on the exchange itself, so the solution to me is to have some kind of membership structure where the primary stakeholders in the process make the policies. Ideally, interests from both long and short side are represented.

This is the crux of the matter.   Those taking the risk must be able to measure and have some semblance of control over it.

[random_cat]

However, the existence of that free option introduces a trading exploit strategy:
Make two accounts
Sell/Buy from/to yourself from the two accounts.
Remove everything except the minimum margin requirement from the account
Get margin called
One account gets a free option, but hope that you get liquidated.
One account gains linearly with price, one account losses is capped at a price
You make free money (at other trader's expense)

This was discussed already a number of times, however the answer is no: At best, you loose only trading fees and end up with non-market neutral position. It is impossible.

Of course this problem reappears if available leverage is too large.

[xeverse]

newminer is one of the traders whose short position was partially closed due to non-paying long-position customers, according to https://icbit.se/margincall worst case scenario

If i get it right, the position was closed of his contract counterparty who got a mc
and there were no other buyers. Exactly who he bought/sold fut from? How can this be?

This must not be done under all circumstances. Again this is not a casino but a hedging market instead.
Suppose if it's hedging position and the trader holds another pos in spot market
which in fact he always does because of the init margin and everything is in bitcoins.
This would immediately incur losses. This is not gonna be a futures market.

The position can be forcibly liquidated  only in case of not meeting mc requirement. Period.
You can not just deal with this by transfering the risk to the otherside. Can you?

Is this really what's happening? Or did i miss something?

[picobit]

This is probably really what happened.

Unlike futures market in the "real world", where the operator knows your identity and can come after you, your losses are limited to the amount you deposit on your margin account.  On an ordinary futures exchange, there is in principle no upper limit to your losses, as the exchange operator can drag you to court to get what you owe him.  The downside of this limitation of your loss it the introduction of counter-party risk: If your counter-party's loss is limited by his available funds, so is your profit.  And yes, this introduces a risk to arbitrageurs.  On the other hand, the profit is also above the normal in arbitrage.

We cannot have it both ways, and in a quasi-anonymous bitcoin setting this counterparty risk is probably unavailable, unless the fees are so high that the exchange can afford to assume that risk - but that would probably scare us all away.

[picobit]

Follow-up to my own post:

It appears that icbit keeps track of who is in each end of all the contracts.  This means that if I go broke, the poor guy who was unfortunately enough to trade with me on the exchange will take all the loss.  This random distribution of the risk in big lumps is in my opinion unfortunate, it would be better if the liquidations were distributed as evenly as possible among the people holding the opposite position.

[Fireball]

Follow-up to my own post:

It appears that icbit keeps track of who is in each end of all the contracts.  This means that if I go broke, the poor guy who was unfortunately enough to trade with me on the exchange will take all the loss.  This random distribution of the risk in big lumps is in my opinion unfortunate, it would be better if the liquidations were distributed as evenly as possible among the people holding the opposite position.

picobit: Your previous post is correct, however this follow up is not. There is no matching of one-to-one (neither in big futures exchanges, nor in ICBIT). The distribution of risks is (more-or-less, with certain exceptions) even to people holding the opposite position.

[picobit]

OK.  Is the loss also evenly distributed, i.e. it is better that everybody loose 1% of their positions rather than a randomly picked 10% loose 10%?  From your post, I assume it is.  I just got the opposite impression from chat and posts here, but that was of course just an impression.

[Fireball]

OK.  Is the loss also evenly distributed, i.e. it is better that everybody loose 1% of their positions rather than a randomly picked 10% loose 10%?  From your post, I assume it is.  I just got the opposite impression from chat and posts here, but that was of course just an impression.

It is (and always was) this way. Only very small (1-5 contracts) positions are excluded, and also position is not touched if trader lost money trading that contract.

Otherwise, it would be too unfair.

[maqifrnswa]

This is probably really what happened.

Unlike futures market in the "real world", where the operator knows your identity and can come after you, your losses are limited to the amount you deposit on your margin account.  On an ordinary futures exchange, there is in principle no upper limit to your losses, as the exchange operator can drag you to court to get what you owe him.  The downside of this limitation of your loss it the introduction of counter-party risk: If your counter-party's loss is limited by his available funds, so is your profit.  And yes, this introduces a risk to arbitrageurs.  On the other hand, the profit is also above the normal in arbitrage.

We cannot have it both ways, and in a quasi-anonymous bitcoin setting this counterparty risk is probably unavailable, unless the fees are so high that the exchange can afford to assume that risk - but that would probably scare us all away.

rather than have higher fees, just require larger margin and perform margin calls earlier (before the account would be forced to go negative.) This would help ensure that no one goes negative. It reduces people's ability to leverage and decreases the risk of an account going broke. In fact, you can force margin calls before the account would go negative AND force sell below market which would "hurt" the account that ran too low and "help" the accounts that are in good standing. This is similar to what a brokerage would do in the real world as an incentive to keep your accounts in good standing.

It is (and always was) this way. Only very small (1-5 contracts) positions are excluded, and also position is not touched if trader lost money trading that contract.

Otherwise, it would be too unfair.

I think arbitrary rules are unfair. The fair thing is for everyone is treated equally, but that is just my opinion. The best scenario is for such situations to not happen at all (stricter margin enforcement so contracts are force sold at market rates).

[xeverse]

This is probably really what happened.

Unlike futures market in the "real world", where the operator knows your identity and can come after you,
your losses are limited to the amount you deposit on your margin account.
On an ordinary futures exchange, there is in principle no upper limit to your losses,
as the exchange operator can drag you to court to get what you owe him.

Have never heard of such before. Are you for real? This wouldn't work out.

If i were to establish my own hedge fund and manage my investor's money,
no way i could risk of my hedging position being closed because of somebody else
defaulted on his margin requirements.

The maximum loss in a corporate market is limited by the the margin requirements for the contract,
all these terms are set in the specification and can not be changed by the exchange afterwards.

Those sharks (operators) would not care about legal procedures and stuff,
they always prefer to have their risks covered in advance and you wouldn't have it any other way.

It's just that i can not believe my eyes that you even consider possibility of closing a client's position
without margin call like acceptable, for whatever reason. That's why i don't know what to say for now.

For sure you guys have to deal with marketmaking then first,
so you could always close the defaulting position at the market price,
or would rather make the first transparent and open source exchange in the world.
With all stats and anonymous account cash flow publicly available to everyone.

[picobit]

The problem is partly that Bitcoin is so volatile, that the exchange operator has to choose between exposing his customers to this kind of counterparty risk, only offering very small leverage, or take a huge risk himself.  Icbit tries to strike a compromise between the first two.

But I agree, it makes icbit unsuitable for hedging, and more risky for arbitrage (but then the arbitrage gain is much larger than elsewhere, but that does of course not help the hedgers)

You can see the procedure described on the web page under Margin Calls (worst case scenario).  I don't think it ever happened from I started to use ICBIT last summer until the bitcoin price ran amok this spring.  But since then volatility has been crazy, and it appears to happen regularly, despite the reduced leverage.

[xeverse]

But I agree, it makes icbit unsuitable for hedging

Then without hedger's money inflow it's not gonna be a derivative market per se but sadly just an another bucket shop instead(
This subtle detail only means to me that an independent marketmaker, who must guarantee that
any position can be liquidated anytime at a reasonable market price, is a necessity,
as far as i concerned about emergence of a true bitcoin derivative marketplace..
At the same time a mmaker can probably go to the spot market any time to hedge their own risks.

To get other insights more information about internal exchange risk management is needed.
Are there any books or research papers about the general derivative exchange architecture available?
Or any other sources of information related to the subject at all? Been looking for such..

[Fireball]

To get other insights more information about internal exchange risk management is needed.
Are there any books or research papers about the general derivative exchange architecture available?
Or any other sources of information related to the subject at all? Been looking for such..

arxiv.org :-)

Then without hedger's money inflow it's not gonna be a derivative market per se but sadly just an another bucket shop instead(

Daer xeverse! Bucket shop information, so you know what you are talking about next time...

"Typically the criminal law definition refers to an operation in which the customer is sold what is supposed to be a derivative interest in a security or commodity future, but there is no transaction made on any exchange. The transaction goes 'in the bucket' and is never executed. Without an actual underlying transaction, the customer is betting against the bucket shop operator, not participating in the market. Alternatively, the bucket shop operator "literally 'plays the bank,' as in a gambling house, against the customer." [ref]

In ICBIT, all transactions are going to the futures or exchange market, they never go in the bucket. Everyone can check this by placing an order, which appears in an orderbook accessible to all traders. All trades happening are broadcasted in real time to all clients. So, it is an open market, with or without hedgers, arbitragers or whoever else money.

That's what makes it different from any forex-alike desk, where openness is an issue.

[picobit]

I think the limitations on the daily varation is a good idea, as it reduces volatility and reduces the sensitivity to "spikes" in the spot price.  The setup with 10% variation per clearing, and two clearings per day also seems like a reasonable compromise.  

However, I think the inability to place asks above the current maximum and bids below the current minimum is a bad thing, and should be allowed within an additional 10% or so.  I have three reasons.

1) If the spot market moves a lot, the price at ICBIT may hit the end of the allowed range.  In that case it becomes impossible to build an opposing side of the order book, and when the clearing happens the price moves rapidly as the relevant side of the order book is almost empty, except for a few ancient order left behind.

2) When the price thus moves suddenly, margin calls may happen.  But since the relevant side of the order book is empty, forces sell/buy cannot happen, and the "worst case scenario" will be invoked.  If we could pre-fill the relevant part of the order book before the clearing, we could act as counterparts to the forced liquidations.  The person being margin called would get a better price, and less people would see forced liquidations.  This is where it becomes important that there is still a limitation on the orders we can place, for example of another 10%.  That prevents me from placing buy orders at absurd prices like 0.1$ and hoping to profit unreasonably from someone to getting margin called.

3) Some of us are not able to sit in front of the browser 24/7, and may want to place orders relevant for the next trading session which are outside the range of the current one.

EDIT: Accidentally placing orders outside the trading range would be annoying, and it is nice that the interface prevents it.  So and additional step should be required to place such an order.  I suggest a check-box on the trading interface "Allow orders outside the trading range".  Even when checked, ask orders should only be allowed above the range, and buy orders below, and only within 10% as described above.

[xeverse]

Hmm well fire, basically yes i agree it's not the most accurate definition in this case.
But still i think it conveys the general idea. Though not sure how to express all of this now.

This meaning is that the money are never going to leave this market, because there would be no connection
to other marketplaces. It's like closed system, in a sense the same  as these crappy bucket shops..
In order to get such interconnectivity going one would have to open positions in both markets simultaneously.
Naturally one of these positions would be a hedging position and another one would be a trading one.
But they are depend on one another, and if there's a risk of a hedging pos being closed anytime
such would incur the losses on the other side as well so to speak. Such is not acceptable for arb.

And the part which is about betting against the house... Hmm.. Nothing can prevent you from doing so anyway.
All exchange stats are not public. And considering that most of the time most of the people in the market are wrong.
You can just take the opposite side in any other marketplace anytime..

Yeah transparency is always the issue.

In a way even these ala fx desks are ok to some extent, precisely until they break their own contract terms..

But having a risk of a position which can be closed any moment doesn't fit in any contract known to me..