Topic: ICBIT Derivatives Market (USD/BTC futures trading) - LIVE - page 12. (Read 97699 times)

Hey now that's more like it!  Thanks for giving this a go, I'm very interested to see how this works out.  I hope and expect you'll pick up more speculative volume.

Wow the trading range is wide too.

I think you mean 25th May not March above.

"Challenge accepted"!

We just introduced a new, very experimental and short (one month) futures contract BUK3 (BTC/USD-05.13).

Features:
- 10 times lower trading fee! (subject to your usual volume discounts too)
- Expires 25th of May (so it trades roughly one month)
- 0.0001 BTC clearing fee per 1 contract (to fight buy-and-hold-till-expiration strategy). It's taken during each scheduled clearing (twice a day) if you have this futures in your balance tab by the moment clearing happens!

Let's see what this experiment shows. The first trading session has enhanced price limits and increased margin requirements. The first clearing will be free (no clearing fees will be deducted).

P.S. Buying and holding previous contract would cost 0.003+0.003 BTC = 0.006 BTC. Buying and holding BUK3 for 30 days will be 0.0001*62 = 0.0062 BTC, so it's nearly the same for b&h strategy.

The structure is pretty reasonable. 2 clearings is ample time to prepare, and adding them isn't killer.

I'm not sold on those indicators of dysfunctional market though.

That's best achieved by deciding not to charge 6 times Gox's already steep charges.

I wanted to speak up to voice support of the changes Fireball has proposed.

To grow and mature this market it's essential that there be market makers. It means more liquidity, tighter spreads, and overall less expensive futures for everyone. Market makers will not participate if there is ever a time they cannot trade as this makes it impossible to continuously manage risk.  Unfortunately this is currently is the case due to the way clearings and limits are structured.

Clearings that happen automatically as a limit is hit gets us closer to this.  Continuous real-time clearing would be even better, but we can come back to that later.   

I urge all members to support these changes.

Another set of updates are deployed.
The most important visible change is that Time&Sales widget (modeled after Clark Moody's Time&Sales table) is now live. It doesn't include history pre-loading yet, but it shows you all trades of the selected futures or currency in real time.

Tomorrow - more improvements coming (as the empty Time&Sales box after page reload doesn't look very nice).

IMHO:

1. the chart is to high and makes the whole section to high and dominating the screen to much. New order section doesn't even reach the bottom of its place. Please make the chart smaller.

2. I would be pleased to see the Time&Sales section next to Order book and the chart far right.

3. Buy/Sell buttons would be better placed just below Total/Initial amount fields - and the hint and trading fee info should be placed below those buttons. ATM the buttons are pushed down at the very bottom in spite of being quite important.

Some UI changes are necessary to further improve the functionality of ICBIT. I just deployed the first bunch of changes, so please update your pages.

Time & Sales widget is not functional yet (would be in the next update), and New Order widget on the Futures tab has some space reserved for enhanced functionality.

I hope you like those. And I would be glad to listen to your feedback - what's good, what's bad, how to make it better. Don't forget to check Exchange tab too.



FIXED: Bug (IE/Firefox only): In Exchange tab, the BUY/SELL switch always remains green instead of red.

Even more, I think we're completely in the wrong category if we discuss minutes.
Let's see it this way: we have two clearings per day. Under any normal circumstances this should be enough. So IMHO the task is to detect the situation, when two clearings per day aren't enough.

Thus I'd propose this rule: if there is the situation of a dysfunctional market (i.e. only one side, ask or bid, and extended up to the trading range), and if this situation pertains for 2 consecutive clearing periods (or at least 90% of the time), then insert two additional clearings

I think the fact that the futures market moves and reacts slower than spot is a feature, not a bug.

I would think that 15 or 30 minutes is way too short, both to prevent market manipulation and to prevent rapid fluctuations.  Would there be anything lost if that time is set to an hour, with an automatic announcement half an hour before?

Thanks that clarifies a lot.
Probably we were all way too much focussed on the movement of the spot rate

In fact, that's exactly what I described. If the price is "pushing" into one of the limits (which means that one side of the orderbook is empty AND another one contains offers at the best possible price within the trading range), and this situation is not intermittent (hence the N minutes rule), then it's obvious that the traders want range extension (usually, exchanges raise margin requirements after that), and an additional clearing should happen.

N may be considerably long, like, 15 or 30 minutes to prevent manipulation possibility.

on a second thought....

why not using the shape of the futures market to a greater extent?

The rule would be: whenever the market becomes totally loopsided, i.e one side of the orderbook is empty within the trading range, and the other side is stacked with orders up to the very limit of the range.


When this condition is met, then the market needs help, e.g. one additional clearing per day.
But if this condition is not met, then there is no necessity for any intervention.

The whole point of futures is that they represent a bet on a future situation.
Thus, if the BTC spot crashes to $30 right now, of what relevance is this for a future due next september? If market participants still feel fine trading (both buying and selling) them at $150, then everything should be fine. Only if everyone wants to get out of their positions, but can't due to the trading range, only then an intervention seems adequate.

I too share the concerns regarding additional clearings.
What exactly to you want to achieve with those?


As long as the limits are not moved, all positions are covered (not entirely true, but roughly speaking). The market might go dead, but no harm is done. But whenever the limits are moved, the probability of forced liquidations is increased. Thus, why should we strive at increasing the speed of possible adjustments to the trading limits?

We have now two clearings per day (correct?).
Even in case of a crash of the spot market, this gives a moderate adjustment rate.

If there is really the need for additional adjustments, then these should be triggered by the spot rate's movement solely (never by the futures). And there should be a relation to the absolute distance between the trading range and the current spot rate.

In the past, we had several times a stable market situation, where the spot rate was outside the futures trading range, due to contango or backwardation. Only when the spot moves significantly and stays at the new position for more than one trading session, plus when additionally the futures market is dysfunctional (e.g. only bids or only asks, stacked to the limits) -- only then a rule for additional clearings seems indicated.

This makes no sense as proposed.  Are there range limits or not?  Pick one.

Given the thin nature of the market I can understand a desire to prevent a certain sort of manipulation via range limits.   If the desire is to
retain limits, then perhaps the limits (as well as margin requirements) should be based on historical volatility of spot.  Unfortunately the
options market is now thin with extremely wide spreads.

More sessions ending at previously scheduled times is a better choice (and allows for scheduled market down time).

If/when there is sufficient maker, why have limits at all?

In any event, note that a lower bound on the probability of an arbitrageur missing a margin call in any given session of a fast market is given by
http://latex.codecogs.com/gif.latex?\sum_{c=0}^{R-1}e^{-0.1N}(0.1N)^{c}/c!
where R = required number of confirmations for a deposit (R>0).   Please choose R and N carefully to avoid dysfunction.


I hope this is published for trader review.   Something fixed with a varying (as often as per session) table of margin factors would be excellent.

I think these are all good ideas - except for the first point.  

On one hand, it was a problem during the rally (and probably also during the crash) that ICBIT had difficulties following the movements of the market - but two daily clearances must have helped.  On the other hand, during the many "mini-crashes" leading up to the big crash, the market moved far down for an hour or so, and then moved back up again.  In these cases, the lack of clearing gave considerable stability to ICBIT compared to the CFD that 1broker was having at the time, where most long positions were liquidated in those situations.  

Also, I would fear that market manipulation would be easier for big players: they can almost trigger a clearing "on demand".  Open a huge position at the end of the trading range during a "sleepy" period, and force a lot of people out of the market.

Fireball,

It's great to be getting the opportunity to discuss this stuff, my feedback below.

1. Additional clearing happens automatically if the price hits the limit for N minutes (N may be 5, 10 or more minutes - to be discussed).

There are two scenarios in which limits are typically locked:
1) During a spot price reversal, spot moves away from futures range and the range is trending TOWARDS the spot but is not able to move quickly enough.
2) During a bull/bear run, where the range is moving AWAY from the spot and speculation results in contango/backwardation reaching range thresholds.

Is the objective to solve one or both of these issues?

As for the proposed solution, it is not clear what "Price" is being referred to here, do you mean Mt Gox spot price, ICBIT "last" price, or some kind of VWAP?   

If basing on Mt Gox prices, this might be effective in times of trend reversal, but will not prevent limit locks during contango/backwardation stretching.

If basing on ICBIT prices, I would suggest using modified VWAP rather than last price to reduce the scope for manipulation, and I would also suggest any trigger times are measured in hours rather than minutes, again to reduce the scope for abuse.

Whilst limit-locked ranges are frustrating at times, the range ceilings/floors do provide a buffer against intraday volatility which is both useful and relevant in a futures market.  Nobody will benefit from changes which make 100% daily range movement possible if the primary outcome of this is increased counterparty default and contract liquidation.

2. Margin calls are executed automatically against the market according to the existing 75% rule.

Does this not currently happen?  I agree this should happen, if a holders margin is exhausted contracts should be auto-offered to market and either sold/bought at the best available price, or placed as an offer at the minimum level which will guarantee 0 counterparty loss.  If no-one picks up the distressed contracts, the corresponding contracts should be liquidated at some specified point (e.g. the subsequent clearing session).

3. Margin requirements (both initial and maintenance) would be different for long and short positions. The formula for margin requirements calculation will include contango/backwardation into account, so that if some traders drive the price away from spot price, they are guaranteed to have enough money reserved in case contango is reduced.

This needs to be thought through very carefully.  A blanket rule to automatically raise margin requirements because of increased backwardation/contango could have the effect of moving innocent contract holders into negative margin and possible liquidation, even if they are on the gaining side of a contract (!).  This could potentially be mitigated if such a margin adjustment were programmed to occur at clearing after unrealised P/L has been transferred to balance (and then used to refil margin), but I suspect that is not what is being proposed.

4. Daily settlement price is going to be calculated as the last hour average

This is far better than using last trade price. However, if there has been no trade in the past hour the "last second single trade manipulation" would appear to remain possible.  To prevent this could you ensure the last hour average somehow apportions appropriate "weight" to periods in which no trades have occurred (e.g. because ranges are limit locked).

5. Contract specifications will be united into one (with the only difference of settlement dates), so that e.g. BUM3, BUU3 would be the same spec, with different settlement dates. The contract specification may be amended, and a list of all amendments must be maintained along with the recent edition of the contract spec. Amendments may come into effect at least in the next trading session after announcement, but usually for big changes the waiting interval after announcement may be a few days.

Cool.

Proposed changes to the futures trading aimed to improve the trading process:

• Additional clearing happens automatically if the price hits the limit for N minutes (N may be 5, 10 or more minutes - to be discussed).
• Margin calls are executed automatically against the market according to the existing 75% rule.
• Margin requirements (both initial and maintenance) would be different for long and short positions. The formula for margin requirements calculation will include contango/backwardation into account, so that if some traders drive the price away from spot price, they are guaranteed to have enough money reserved in case contango is reduced.
• Daily settlement price is going to be calculated as the last hour average
• Contract specifications will be united into one (with the only difference of settlement dates), so that e.g. BUM3, BUU3 would be the same spec, with different settlement dates. The contract specification may be amended, and a list of all amendments must be maintained along with the recent edition of the contract spec. Amendments may come into effect at least in the next trading session after announcement, but usually for big changes the waiting interval after announcement may be a few days.

Thanks to alpet and potential market makers who proposed one or more of these changes. Your help is really appreciated!

Please let me know your thoughts about this.

Wow, what a coincidence!

I tend to agree.  The first round of Gold futures were essentially gold futures.  There was a contango in Gold at the same time as there was a contango in BTC versus USD.  The second round were BTC futures traded in the currency "gold", a bit more complicated to figure out, but essentially just the same as BTCUSD.  There was constantly a backwardation corresponding pretty much to the contango in BTC/USD.  Gold futures probably only make sense in times where the BTC is pretty stable, and even then there would be a tendency to treat them as another BTC futures.