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

[zebedee]

There are a couple more differences from the old contract which I would like to point out:
a. The fee in contract specification is not fixed, so it may be changed in future, if it becomes necessary.
b. There is a settlement fee for these 3 new contracts, which is the same as a trading fee. This fee exists on most futures exchanges and it's needed to stimulate liquidity on the market, which is good for the market.

Yikes!  So even J3 is going to cost a full point to trade in-and-out, whether you out or not!

 I think I'll leave it alone, thanks.  I wouldn't call that "stimulating liquidity", more like "killing it".

[Fireball]

As many of you are already aware, we launched three new contracts:

• BTC/USD-4.13

- the short one, expiring in April, with fees lowered to 0.003 BTC per trade

• BTC/USD-6.13

- the medium one, expiring in June, with even lower fees: 0.002 BTC per trade

• BTC/USD-9.13

- the longest one, expiring in September, with the lowest fees: 0.001 BTC per trade, which is 5 times less than the currently trading BTC/USD-3.13 contract
[/list]

There are a couple more differences from the old contract which I would like to point out:
a. The fee in contract specification is not fixed, so it may be changed in future, if it becomes necessary.

Yikes!  So even J3 is going to cost a full point to trade in-and-out, whether you out or not!

 I think I'll leave it alone, thanks.  I wouldn't call that "stimulating liquidity", more like "killing it".

The magnitude of USD/BTC rate changes allows traders to make large enough profits (without revealing any confidential data, profits from 100 to 300 BTC are rather common).
Eventually, either USD/BTC rate or our fees would drop

[zebedee]

The magnitude of USD/BTC rate changes allows traders to make large enough profits (without revealing any confidential data, profits from 100 to 300 BTC are rather common).
Eventually, either USD/BTC rate or our fees would drop

You said you introduced the rule to "stimulate liquidity".  It seems you are concerned liquidity is not good enough.

If I were you, I'd be lowering your fee, so tightening the bid-offer and encouraging trading by making it cheaper and easier to get in and out.

Instead the lesson from "insufficient liquidity" you seem to have drawn is "raise fees by making it twice as expensive as it already was" for people who simply want to take a long position.

I really don't think this will work out well for you.  The nature of the contract spec, with its inverse quote and USD denomination, but the egregious taking of fees in BTC, has probably obscured the true cost to participants.  But when they look at their P&L I don't think it will escape their attention for long.  As I said, I won't be trading these until the inverse-square relationship of price to bid-offer spread is fixed, or fees are significantly reduced.

[Fireball]

It still would be nice for there to be shown the history of daily clearing prices.  (And also the ability to access the charts and order book without first authenticating, incidentally.)

Indeed I heard many similar requests, however I want to make some tradeoff by making the frontpage way more informative and "live" than the currently existing one, however imposing some limitations (slower update rate, for example) not to overload the server with many unauthenticated and thus uncontrolled connections.

This is already in progress, so should be deployed relatively soon.

How far are you along this path?

I would also appreciate unauthorized access to the order book. You see, I'm considering creating a website that actively tracks backwardation in bitcoin. So in order to track that I need to know the asks for the futures contracts, and the bid on BTC on Mt. Gox. Mt. Gox already has public order books, so your site is the only missing piece of the puzzle.

There were higher priority stuff, but now it may be perfect time to do this, because I will be doing a publicly available futures dashboard anyway.

[Fireball]

Instead the lesson from "insufficient liquidity" you seem to have drawn is "raise fees by making it twice as expensive as it already was" for people who simply want to take a long position.

I appreciate your comments and critique, but I don't follow your logic in this specific message.
The fees are 5 times lower for BUU3 than for BUH3 already now, and additional discounts will be available pretty soon for all contracts.

[zebedee]

Instead the lesson from "insufficient liquidity" you seem to have drawn is "raise fees by making it twice as expensive as it already was" for people who simply want to take a long position.

I appreciate your comments and critique, but I don't follow your logic in this specific message.
The fees are 5 times lower for BUU3 than for BUH3 already now, and additional discounts will be available pretty soon for all contracts.

It's simple - you doubled the cost to trade-and-hold.  Instead of encouraging people to trade, they simply won't open a position unless they're willing to bear the doubled cost.

I was talking about BUJ3 - still 0.003 BTC per contract.  At current prices that's 1 full point it has to move in your favour before you make any money on a new position.  But if BTC goes to 50, that becomes 1.50 full points it needs to move in your favour before you make any money on a new position.  At 60 USD per BTC it's about 2.2 points before profit!

BUU3 is much more reasonable *at the moment*, though still more expensive than BUH3 when it came out.  But that would be quickly wiped out by BTC price rises as ticks-to-profit goes up with the square of the price.  BUH3 price has almost quadrupled, so let's say that equates to a 14 times increase in ticks-to-profit.  Buy you only reduced trading fees by a factor of 5.

[Stephen Gornick]

profitable traders positions were reduced (as if they sold them to realize the profit)

Doh! I knew that was a risk.  So had I sold those BUH3's yesterday at $40+ I would have kept my profit but because I pushed my luck I see part of my position sold at a lower thirty-something level because some anonymous trader was too short.

Doesn't this make the perfect scam opportunity then?  

Scammer opens two accounts, long and short.  
Long account is used to buy at about 5X leverage.  Each buy is matched with a sell from the short account, also at 5X leverage.
Wait for rip roaring rally or selloff.   Upon receiving first margin call notice on either the long account or the short account, close out positions on the other account that has the gains, maybe even use the account with the negative balance to provide liquidity (assuming a short can still buy to close, for example, when the account is negative).  Then the profitable account withdraws the coins (variation margin is already settled when the position is closed).  Two days later (or whatever the delay is) other customers with profitable positions see forced trades to cover for the negative account.

This is where leverage/margin accounts + anonymity don't mix.

[killerstorm]

Scammer opens two accounts, long and short.  
...
This is where leverage/margin accounts + anonymity don't mix.

Duh. I discussed this very problem with Fireball ~2 years ago, before ICBIT was implemented.

It is fundamentally impossible to implement an anonymous market for classic futures.

One either needs to introduce some kind of a counter-party risk, but it has an unfortunate effect: it is quite like a lottery in sense that market participants never know when it will bite them; and it is exploitable by scammers.

Or introduce limits on profit and loss, so contracts won't behave as futures anymore.

[zebedee]

Doesn't this make the perfect scam opportunity then?  

Scammer opens two accounts, long and short.  
Long account is used to buy at about 5X leverage.  Each buy is matched with a sell from the short account, also at 5X leverage.
Wait for rip roaring rally or selloff.   Upon receiving first margin call notice on either the long account or the short account, close out positions on the other account that has the gains, maybe even use the account with the negative balance to provide liquidity (assuming a short can still buy to close, for example, when the account is negative).  Then the profitable account withdraws the coins (variation margin is already settled when the position is closed).  Two days later (or whatever the delay is) other customers with profitable positions see forced trades to cover for the negative account.

This is where leverage/margin accounts + anonymity don't mix.

I don't fully follow what you mean.  The deal with margining is quite opaque on icbit.  I assume a day's variation margin can cover two day's maximum move.  So if an account hasn't funded an EOD margin call by the end of next day, they have enough money at hand to close them out at that 2nd days' worst price.  It means someone else loses their position, but at least they're getting out at the best possible price of the day.

Am I misunderstanding the way icbit "works"?

[zebedee]

I really think the BTC/USD contracts should be like 99% of the futures contracts in the world:

a) Denominated in some currency; USD is more natural than BTC
b) Have a constant tick size in that currency (by tick size here I mean 0.01; I realize these contracts are free-quoting and don't have ticks as such)
c) Have a trading fee that is a constant number of ticks.

The current contract specification is very confusing, and in particular makes calculating effective profit and fees very difficult.  One consequence is that market move required to cover the fees is surprisingly large, and not constant.

Is there a reason this can't be done?

[Ichthyo]

Scammer opens two accounts, long and short.  
Long account is used to buy at about 5X leverage.  Each buy is matched with a sell from the short account, also at 5X leverage.
Wait for rip roaring rally or selloff.   Upon receiving first margin call notice on either the long account or the short account, close out positions on the other account that has the gains, maybe even use the account with the negative balance to provide liquidity (assuming a short can still buy to close, for example, when the account is negative).  Then the profitable account withdraws the coins (variation margin is already settled when the position is closed).  Two days later (or whatever the delay is) other customers with profitable positions see forced trades to cover for the negative account.

This is where leverage/margin accounts + anonymity don't mix.

in practice it won't be that extreme. The scammer in your proposed scheme will loose the collateral on the account developing negative. So basically he can scalp off just from one account what he looses on another one, minus the trading fees. Moreover, if you deal with yourself, using a second account, you're not harming anyone else. Generally speaking, the only damage anyone can do is to prevent other people from receiving the full amount of gains possible due to the market movement.

Thus, in the end, I think, anonymity + leverage plus collateral work perfectly well.

[killerstorm]

I don't fully follow what you mean.  The deal with margining is quite opaque on icbit.  I assume a day's variation margin can cover two day's maximum move.

It doesn't even cover one day maximum move.

You can see price range and initial margin yourself. There is a formula for var. margin. So you can calculate maximum loss.

[killerstorm]

in practice it won't be that extreme. The scammer in your proposed scheme will loose the collateral on the account developing negative. So basically he can scalp off just from one account what he looses on another one, minus the trading fees.

Well, the idea is that profit on one account isn't limited but loss on another is. Suppose I fund each account with 1 BTC.

If market moves in one direction over several days (and it is often the case with Bitcoin), my loss is ~1 BTC, but my profit might be 2 BTC.

Effectively it is a bet on high volatility.

Moreover, if you deal with yourself, using a second account, you're not harming anyone else.

How so?

Generally speaking, the only damage anyone can do is to prevent other people from receiving the full amount of gains possible due to the market movement.

Well, Fireball is either dishonest or he doesn't understand how futures work. (LOL!) And neither do you.

Suppose I bought at 40. Then next day futures price went to 35. Then back to 40. My P&L over two days should be 0, right?

Wrong!

It turns out that when price moves from 35 to 40 my position is "profitable" and so it might be forcibly closed "at a very good price", which won't cover my previous loss.

Thus on ICBIT it is possible to suffer loss due to counterparty risk.

It is even possible to have a loss when you should have had profit.

Thus, in the end, I think, anonymity + leverage plus collateral work perfectly well.

It works only as long as collateral covers price movements.

[Fireball]

It turns out that when price moves from 35 to 40 my position is "profitable" and so it might be forcibly closed "at a very good price", which won't cover my previous loss.

Allright, you're nitpicking , but by "profitable position" I mean open position for a user whose total profit for this security over whole trading range is positive.

So if trader lost 100 BTC before, then bought $35, price went to $40 and he/she did not cover losses yet, his position will not be considered for reduction due to non-paying customer.

[Fireball]

Scammer opens two accounts, long and short.  

No, not really. The thing is that before going negative, in majority of the cases, scammer's counter-market position will be liquidated (by margin-calling).
However, in the worst case like yesterday, forced trades will occur. And still this won't bring the scammer profit, because scammer's "short" account will not reach negative state, but 0, so money would not be pumped out of nowhere.

So maximum he would get is just slightly reduced profit from "long" minus evreything he had in the "short" account.

[Fireball]

I really think the BTC/USD contracts should be like 99% of the futures contracts in the world:

a) Denominated in some currency; USD is more natural than BTC
b) Have a constant tick size in that currency (by tick size here I mean 0.01; I realize these contracts are free-quoting and don't have ticks as such)
c) Have a trading fee that is a constant number of ticks.

The current contract specification is very confusing, and in particular makes calculating effective profit and fees very difficult.  One consequence is that market move required to cover the fees is surprisingly large, and not constant.

Is there a reason this can't be done?

This was heavily discussed before the futures exchange was launched. However, it boils down to the fact that if we want to see futures quoted in $ per BTC, then the variation margin would be transferred in US Dollars. This contradicts the desire to start a real Bitcoin economy, a place where your Bitcoins are put to work, not just another site to gamble on Bitcoin rate and get dollars back. Additionally, this fact brings legal issues (running a futures exchange with US dollars transactions would require licensing, bank accounts, etc etc).

So instead, the futures was introduced which uses BTC per $. However, as you realize, it's easy to convert the price back and forth - just perform 1/x operation, and you have price which everyone is used to.
Also the example helps in understanding this choice.

I was thinking and discussing that with my friends with solid financial background, however when it comes to this configuration, this seems to be the best possible option. Nevertheless, if there are some proposals - I'm all ears.

Bytheway, as for b), the tick size in USD is 0.0001 for this contract, it's not really free floating.

[killerstorm]

Allright, you're nitpicking , but by "profitable position" I mean open position for a user whose total profit for this security over whole trading range is positive.

So if trader lost 100 BTC before, then bought $35, price went to $40 and he/she did not cover losses yet, his position will not be considered for reduction due to non-paying customer.

Oh, I thought you meant profitable within a single trading session.

Still it might be somewhat problematic for people who use futures to hedge their positions (e.g. if one sells goods for BTC but needs to pay supplier in USD), but I guess they have to live with it.

Have you thought about making futures with guaranteed profit? Maybe via insurance (higher fees) or something like that.

[zebedee]

Bytheway, as for b), the tick size in USD is 0.0001 for this contract, it's not really free floating.

Thanks for the correction.  Why not have fees as an explicit number of ticks, charged in BTC at the price traded?  Would be simpler and avoid the issues you're having.

Wondering if / when we'll ever get any offers back...

[Fireball]

Still it might be somewhat problematic for people who use futures to hedge their positions (e.g. if one sells goods for BTC but needs to pay supplier in USD), but I guess they have to live with it.

Have you thought about making futures with guaranteed profit? Maybe via insurance (higher fees) or something like that.

Yeah, this maybe problematic. As the exchange grows (it's still very very small compared to e.g. MtGox), this problem would not be so important because of bigger quantity of bids/asks.
As for insurance - this is a very interesting thing, but needs some research.

[picobit]

Scammer opens two accounts, long and short.  

No, not really. The thing is that before going negative, in majority of the cases, scammer's counter-market position will be liquidated (by margin-calling).
However, in the worst case like yesterday, forced trades will occur. And still this won't bring the scammer profit, because scammer's "short" account will not reach negative state, but 0, so money would not be pumped out of nowhere.

So maximum he would get is just slightly reduced profit from "long" minus evreything he had in the "short" account.

But that is still a gain for the "gaming" trader (I think calling him a scammer is a bit strong).

I.e. I open two positions with 10 BTC margin, one short and one long.  The market moves significantly, one way or the other.  One account gains 20 BTC, the other would have lost 20 BTC but the loss is limited to 10 BTC by available funding.  Net gain 10 BTC.  That gain has to come from someone, either ICBIT if the account was allowed to go negative, or from other traders' profit if there is worst-case closure.

By the way, something like that could soon be very relevant.  I hold a long position (mainly by shorting gold), and liquidity is low during these crazy market conditions.  If bitcoin crashes, I will of course take a loss.  And just like the bid side of the gold order book is empty now, the ask side is likely to be gone by then, making it hard to get out of the market in time.  I am counting on my loss being more-or-less limited by the amount I have on ICBIT.  One one hand, I do not want to risk more money than I already have, but on the other hand I do not want to end up owing people money and not paying.  Do the margin calls on ICBIT protect me, i.e. will the automatic liquidation prevent my account from going significantly negative?

On the flip side, whenever there is a large movement upwards, I risk to get less profit than expected because somebody with a short position is forcibly closed.  That is the rules of the game as stated on the web page, and I have no problem with that - at least not if it happens rapidly, i.e. that I loose the last day or last few days of profit. But I would be very unhappy if this happens at a much later time (for example at settlement) and that I loose weeks of profit, killing my own margin account.

Are my expectations approximately correct?