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

[steve417674]

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.

I think this is the major cause of the problem.  It seems that the exchange has basically handcuffed itself with the very rule that is supposed to protect them in the first place.

Since there is no recourse against individuals that go into technical default, and no background checks to determine if a trader is financially sound to use margin/derivatives, it comes down to how to self-regulate ourselves better than the CFTC could.  We have to throw out some of the rules that normal clearinghouses use and use more creativity to protect both traders and the exchange.

There are really four variables that can be modified here:

• Leverage ratio - the quickest most direct fix.  The downside here is that I think most of us would actually like more gearing, rather than less.  Going any lower than 2.5x would probably divert a lot of volume away.  It can and should go higher, once the exchange is more stable, but now's probably not the time for it.  However, once it's higher, it opens up the possibility of having a dynamically adjusted level based on market conditions.
• Trading range - uncertain impact.  With a lower range, the market locks quickly and could remain locked across multiple trading sessions, thus further stifling trading.  With a higher range, the greater the chance positions get wiped away when someone can't get to their computer and refill their account.
• Margin call condition - effective and somewhat unobtrusive.  Increasing the percentage decreases the probability of liquidations, but at the expense of traders having to refill more frequently.
• Frequency of clearing - in practice, quite impactful, but too obtrusive.  With a higher frequency (say 3x day), you can run into situations where two trading sessions pass while someone's asleep.  And higher than that you brush up against the natural delay in confirming BTC transactions.

   

The way I see it is, the best option is changing the point at which maintenance margin has to be refilled.  Something like: issue a call for maintenance replenishment at 50%, giving the trader one trading session (12 hrs) to top back up to 100%.  If the call was not obliged, then force liquidate.  Within the one trading session, if the market continues to move quickly, margin call at < 25%.  So you basically wind up with an added buffer where traders can top up margin without giving them the implicit of option of running away, and without having to offset their trades in a locked market.

In addition, I would like to see the trading range move to something that's algorithmically adjusted.  Rather than a fixed number, it should be something that adjusts relative to market conditions and the current risk present in the exchange's book.  For example, if the Average True Range over the past n hours in the spot market is > x BTC, or if y% of trades in the book have < z% maintenance margin, decrease the range by a function of these amounts.  And vice versa.  The trick here is that you want to be protective, but also not drying up trading completely.  Which leads me to...

To picobit's point about pre-filling, I'm having a hard time determining why this would be a bad idea.  Obviously you can't have market orders being crossed outside the range, but at least there is the possibility to get a more realistic quote during liquidation.  It seems that "pre-filling" already occurs indirectly by not clearing the book after range adjustments, and having old out-of-range trades sitting there.  Just looking at the current market depth for BUM3, you have ~75% of the asks outside the range vs. ~5% of the bids outside the range.  I think you either have to allow limit orders outside the range at all times, or none of the time.

There is also a fifth variable - the length of the settlement period.  But, it would probably take a long sampling period to determine what works best.

Just my two cents.  I've only been trading for 2 weeks and only once has this scenario come into play while I've been on the site.  I'm sure it's happened many times before, but unfortunately I missed the discussion in the past.  And I'm sure these topics have been debated plenty of times as well, so my apologies in advanced.

[1714042851]

1714042851

[1714042851]

1714042851

[1714042851]

1714042851

[boomerlu]

Guys, I have a proposal for "Clearing Membership" status. I will post it when I get home. If it's accepted, then we will all be voting on the ideas you guys have posted.

The essential idea is that Clearing Members post capital to guarantee liquidation at close to market price. In exchange for taking this risk, they get voting rights to set settlement and clearing policy and voting rights for emergency situations.

Changes I see needing to be made immediately though...
1) Forced liquidation regardless of win vs loss
2) Forced liquidation even in the case of small contract size (fractional contracts would help here)
3) If a trader is margin called (at 75% initial margin), and IF the current liquidity in the market would allow for liquidation with (<25%) of initial margin remaining after liquidation, then liquidation must occur immediately. That is, if there's not enough liquidity in the market for a liquidation with a safety margin, liquidation should happen immediately.
4) More transparency (pseudocode for example) on the exact liquidation process.

1 + 2 would help, again it's better to have everybody lose 10% of their position, than the winners lose 20% of theirs.

3 is a better liquidation rule, allows for some time cushion to get funds but keeps things a bit safer.

4 is to help avoiding nasty surprises like two Fridays ago. If we see something wrong we can point it out and have it fixed.

[picobit]

I really think you should consider allowing bids to be placed below the current trading range, and asks above.  That will allow us to build an "opposing side" of the order book in case of large price changes.  That will both dampen the swings, and provide a market when someone is margin called.

And +1 to the pseudocode suggestion

[Super T]

i agree

[boomerlu]

Here's a sketch for clearing membership rules. All numbers are of course subject to change.

A member must keep some amount of BTC reserves at the exchange to help with forced liquidation scenarios. In exchange for those reserves, they get votes (proportional to their designated clearing reserves, but capped so that no person dominates the discussion say 2000 USD worth) in determining policy in the following areas
- forced liquidations
- margin
- margin call
- clearing and settlement
With an appropriate 1-2 week lag to allow for implementation.

In addition, when a major event in any of these categories occurs (usually manipulation or massive insolvency) all members are contacted and have 1 hour to get to a member's chat and come to a decision.

Members above 100 USD must have their emails on file, above 1000 USD must have their phone numbers (or skype, whatever is appropriate) on file so that they can be called to make such decisions. Any member is allowed to file their emails and phone numbers, but those above thresholds MUST. Decisions can only be made if at least 50% of the votes are cast and are decided by 2/3 majority. Alex or some other neutral third party from the exchange itself would be the executor of the decision, and the decision maker if voting fails to meet the standards set.

Any group of members with total reserves above 1000 USD may initiate a vote to amend the rules. The members with lower $$$ but good ideas have to find a willing sponsor or create a coalition to initiate (there's nothing preventing discussion in the chat). These rules changes are voted on bi-weekly with the same rules for emergency decisions, except there is no tiebreaker if the vote fails.

Reserves are updated biweekly and thus, once capital is designated as "membership reserved" it is locked in for 2 weeks.

If a member's reserves are used in an orderly liquidation, that member is entitled to a rebate on his trades until that amount is made up. The risk would be that they suffer such heavy immediate losses that they cannot trade anymore, so the rebate is gone.

[picobit]

I am a bit sceptical (to say the least) about this proposal.  It seems that little is gained by members to offset the economic risk they are taking, unless they are able to change the rules in such a way to favorize themselves - which would of course chase everybody else away.

[boomerlu]

I am a bit sceptical (to say the least) about this proposal.  It seems that little is gained by members to offset the economic risk they are taking, unless they are able to change the rules in such a way to favorize themselves - which would of course chase everybody else away.

There are two primary reasons to be a member, besides attempting to manipulate
1) This is the main point - there ALREADY is economic risk, but it's hidden in the corner cases of a clearing/settlement policy that's not visible to us (it's been 3 weeks since the last problem and still nothing), and not controlled by us. I've already lost money (as have others) on this so don't talk to me about not offsetting economic risk.
2) Members would get first look at emergency liquidations if there's not enough market liquidity - a chance to buy or sell at a deep discount.

There are a few reasons that there should not be manipulation
1) There's no bias between long and short sides.
2) Members are by definition first in the line of fire for counterparty liquidation (after first look, they must make up the shortfall). This cannot be changed.

[Fireball]

I am a bit sceptical (to say the least) about this proposal.  It seems that little is gained by members to offset the economic risk they are taking, unless they are able to change the rules in such a way to favorize themselves - which would of course chase everybody else away.

There are two primary reasons to be a member, besides attempting to manipulate
1) This is the main point - there ALREADY is economic risk, but it's hidden in the corner cases of a clearing/settlement policy that's not visible to us (it's been 3 weeks since the last problem and still nothing), and not controlled by us. I've already lost money (as have others) on this so don't talk to me about not offsetting economic risk.
2) Members would get first look at emergency liquidations if there's not enough market liquidity - a chance to buy or sell at a deep discount.

There are a few reasons that there should not be manipulation
1) There's no bias between long and short sides.
2) Members are by definition first in the line of fire for counterparty liquidation (after first look, they must make up the shortfall). This cannot be changed.

Picobit outlined absolutely same concerns which I have myself.

First of all, the problem should be solved entirely and fairly. That means timely margincalls, and forced emergency liquidation in corner cases, which is fair to every participant according to his position.
Second, if someone is willing to pose as a risk-taker, thus bringing all risk to himself/herself, than that someone would want some benefits.

There is not much to offer regarding the voting rights over rules. I would be strictly against the fact that rules would be defined by a small elite group, as this would make it unfair. So far, ICBIT sticks to the rules developed by existing stock exchanges (and they will be formalized and published pretty soon).

As for boomerlu's suggestion about being first in line for margin calls liquidation - yes, that makes sense as a more-or-less safe benefit.

[boomerlu]

I am a bit sceptical (to say the least) about this proposal.  It seems that little is gained by members to offset the economic risk they are taking, unless they are able to change the rules in such a way to favorize themselves - which would of course chase everybody else away.

There are two primary reasons to be a member, besides attempting to manipulate
1) This is the main point - there ALREADY is economic risk, but it's hidden in the corner cases of a clearing/settlement policy that's not visible to us (it's been 3 weeks since the last problem and still nothing), and not controlled by us. I've already lost money (as have others) on this so don't talk to me about not offsetting economic risk.
2) Members would get first look at emergency liquidations if there's not enough market liquidity - a chance to buy or sell at a deep discount.

There are a few reasons that there should not be manipulation
1) There's no bias between long and short sides.
2) Members are by definition first in the line of fire for counterparty liquidation (after first look, they must make up the shortfall). This cannot be changed.

Picobit outlined absolutely same concerns which I have myself.

First of all, the problem should be solved entirely and fairly. That means timely margincalls, and forced emergency liquidation in corner cases, which is fair to every participant according to his position.
Second, if someone is willing to pose as a risk-taker, thus bringing all risk to himself/herself, than that someone would want some benefits.

There is not much to offer regarding the voting rights over rules. I would be strictly against the fact that rules would be defined by a small elite group, as this would make it unfair. So far, ICBIT sticks to the rules developed by existing stock exchanges (and they will be formalized and published pretty soon).

As for boomerlu's suggestion about being first in line for margin calls liquidation - yes, that makes sense as a more-or-less safe benefit.

Regarding the "small elite" criticism...
1) Currently you and the others running the exchange control the rules. You are receptive to opinions, but ultimately you have the final say.
2) On top of that, only a few of us actually post or chat (I remember you had about 6k traders, I have maybe only seen about 100 names active in the chats or forums).
3) It can be made cheap to be a clearing member, and like I said there can be a cap to the amount of capital contribution. Or you can make the voting rights go up as the square root of capital contributed (diminishing returns). There are any number of essentially mathematical ways you can get around this.
4) It's not necessarily bad to make the group smaller than the full set of traders. Some of us have more experience than others (and I'm not even saying myself). Remember there can be no perfect political system. Somebody's interests always have to win out over others in the end. At least here the winners actually have to cover losses.
5) Edit: Also, again, it seems a lot of traders don't even chat or post. Maybe they're shy; handing them the option to easily lock up some of their extra BTC here (a form, I devote X BTC to covering liquidation shortfall in exchange for voting rights) and then a simple poll for proposals could INCREASE their participation in the process.

[Fireball]

I introduced a few improvements to the trading API based on alpet's suggestions. It's only additional information channels and additional fields in existing data structures, so backward compatibility fully remains.

Now user_order data structure is enhanced with  a new field called "fills" which contains trade information if that incoming update resulted from a trade or trades. And the API page spec page got improved too, adding previously undocumented data structures.

TBD: Also small addition to the web interface is that now when your order is filled, you will see a notification mesage box in the top right corner.

[boomerlu]

Now you have user_fills information channel which will deliver you messages when your active order is partially or fully filled.

This is excellent and essential. Avoids us having to poll the API to get fills. Thanks.

[boomerlu]

Calendar spreads would be nice for the upcoming roll on Friday. K-M and K-U contracts (long the front, short the back, back px pegged to front px).

[Fireball]

Work on API continues: This time futures "ticker" message (denoted as private:ticker in the api spec) got improved:
- It's update frequency increased 10x (max. 100ms refresh interval now instead of 1 second previously)
- A whole bunch of new useful fields is added

Also, fills data structure in user_order update messages now contains information about current balance of the futures (or current base/quoted currencies for the exchange market). This simplifies the code for trading bot writers. (Note that you still get the usual balances update message, but you don't have to wait for it anymore, as most traders want to put algo logic in a fill event handler).

An update was posted to the twitter channel, regarding the timestamps in the system. Alas, I missed some places which use imprecise timestamps. The idea is to unify all of timestamps to use microseconds. Please make sure you update your code (you can easily add a check now what kind of timestamp you receive: unix time stamp or time stamp in microseconds, to be ready for the switch).

[Fireball]

2105 contracts of BUK3 settled today at $130.9759 (MtGox VWAP), total trading volume $168770.

The low trading fee experiment is rather successful, it brought enough of liquidity into the contract, and BUK3 will be substituted by the new contract called BUN3, launched today or tomorrow.

Happy trading!

[robottwo]

I'm not neccessarily a fan of the membership idea. But, there were a few things Boomerlu pointed out which are a big deal. This hasn't come up recently because the market has been reasonably stable for the past month, but if it isn't resolved then traders will have big problems as soon as it moves again.

1) Forced liquidation regardless of win vs loss
2) Forced liquidation even in the case of small contract size

To put more color on it:

[robottwo]

I'm not a fan of having memberships. I don't think there's any real benefit. The problem is not that the exchange needs additional capital to cover counterparty risk. The problem is that the exchange is not properly mitigating the counterparty risk. This hasn't been an issue for the past month because the market has been fairly stable. However, as soon as bitcoins move again (in either direction) this will become an issue again.

There are two points BoomerLu made which I think should be addressed ASAP:

paraphrase:

1) Forced liquidation regardless of win vs loss
2) Forced, immediate liquidation even in the case of small contract size (fractional contracts would help here)

The below puts a bit more color on these points, but it boils down to the two items above.

To put more color on it, it's not the exchange's role to decide who has "positive returns" or "negative returns" -- especially when deciding who to liquidate against. That's because more advanced traders will have more complicated positions on against multiple instruments. For a simple example, if I'm trading BUM3 against BUN3, then looking at the PNL of only one contract doesn't tell the whole story. Similarly, I may be involved in a trade using multiple exchanges, so icbit doesn't even have access to all the information it needs to determine my PNL.

The exchange's role is to maintain predictability of behavior. So, if someone has breached margin, *and* there are no orders to fill against in the market, then the only fair thing to do is liquidate evenly across all the existing open interest on the other side.

And, since that implies that the orderbook is empty, the execution price should be the worst-case scenario -- which is the maximum price movement for the session, on the empty side of the orderbook.

I think the above is clear and uncontraversially fair in terms of how to liqiudate positions. In the past, however, it wasn't achievable because positions weren't liquidated in a timely manor. At times, there has been lag of 4+ sessions before a margin call occurred. The fallout is that prices moved so far in that time, that the member in bad standing didn't have enough capital to close out the position, and instead margin was taken from members in good standing.

That's unacceptable, but it is avoidable. The policy should be this:
   1. Warning letters are sent when margin falls below 100% of initial margin. Members can post additional margin to be safe.
   2. Positions are liquidated as soon as margin falls below 75% of initial margin. This is the current policy, but it turns out that the liquidation doesn't really occur.


We've spoken about this quite a bit over the past month or so, both in this forum, int he icbit chat, and in offline conversations. Fireball, can you please summarize the current status with respect to the above, and what the plans are going forward?
   

[Fireball]

New futures contract on S&P500 index value just launched! It's ticker is ESU3 (full name S&P500-9.13, expiration month is September)

It differs from the currently trading ESM3 in size: the new one is 100 smaller, which allows for more liqudity and lower initial margin values. It's fully linear contract, so 1 tick change in the price results in 0.0001 change in BTC.

Happy trading!

[Fireball]

Two news, good and bad.

Bad news is that AurumXchange announced ceasing of VouchX operations. Thankfully, they offer a full refund, so no money are stolen. However, it's not possible to withdraw them automatically at the moment. We will work with Aurum to determine the best return strategy for our users.

Good news is that the ticker update message now includes Open/High/Low fields, which provide session-wide statistics for Futures market (for Exchange market they are transmitted in experimental mode, do not rely on their value yet).

[boomerlu]

Fireball, I've been asking for a while, can you roll out a Dec contract (BUZ3) soon?

[boomerlu]

I'm not a fan of having memberships. I don't think there's any real benefit. The problem is not that the exchange needs additional capital to cover counterparty risk. The problem is that the exchange is not properly mitigating the counterparty risk. This hasn't been an issue for the past month because the market has been fairly stable. However, as soon as bitcoins move again (in either direction) this will become an issue again.   

I agree that memberships are not the solution to THIS problem.

But what about the next one? Can you foresee all possible future problems? Do you think admin can? Do you think he will pick the right solution every time (the fairest with the least loopholes), when he can't lose any money (besides future value)? Do you think it's better for him to aggregate our opinions himself, or do you think it's better that we have a voting mechanism? Whose interests do you think admin will have in mind if he is making emergency decisions?

Do you remember wanting some type of repayment as a matter of "principle"? Why was that?

That is the real crux of having membership, not a specific clearing problem. People aren't careful when there's nothing on the line, or when there's no possibility of downside (only reduced upside), it's psychology. Just check out play-money poker tables.