Hi all I've searched the Web near and far but could not find an answer to my question. Could a kind soul please help out? Consider the following situation: I want to buy an exchange traded spread, e.g. the NOB, and hold for several days. Let's say the NOB is unchanged for the day. I pull up a DOM, buy and put a stop 10 ticks below and a target 20 ticks higher (both GTC). The NOB rallies and closes 10 ticks higher for the day. So far, so good. Overnight, the NOB resets. Upon the open, where are my stop and my target? Would they still be 10 ticks and 20 ticks respectively away? Or would my stop be 20 ticks (difference between the original stop and yesterday's gains) and my target 10 ticks (original target minus yesterday's gains) away? Does it depend on the broker, the platform? How do TT or CQG handle this? Many thanks in advance for any help.
I'll try to bump this one time and if there is no answer, I'll let the thread die. Does anyone have an answer? Please excuse my insolence.
Can't answer your question, never heard of exchange traded spreads, NOBs or DOMs. But if you have a moment to enlighten my ignorance, why are you wanting to trade these things in preference to other instruments?
That's an excellent question tomorton. I was backtesting a strategy on different instruments and to my surprise found out that it works better on (some) spreads. Subsequently I started doing research on how exactly spreads work, discovered exchange traded spreads and am now trying to figure out how orders are being handled on those spreads. If you are interested in learning more about spreads, the CME has a ton of educational material. I have also found this book by Stephen Aikin very helpful for STIR trading.
I have been experimenting with treasury spreads as well including the CME implied ICS spreads. It is difficult to get straight answers as to the behavior of the spreads in terms of execution in the live market. I use CQG and they have admitted that the simulator does not work properly in terms of P&L. However, if you read the CME literature on the treasury ICS spreads you will see that GTC orders are not supported. So your stops and target would no longer be active when the spread resets at close. You would have to replace them each day the trade is open. Ned
Very interesting Ned, thank you for reporting your findings. Since posting my question I have executed several live spread orders on Interactive Brokers (or rather: Interactive Brokers has executed the orders for me) with bracket orders attached. Not from the DOM but from charts. All positions were held overnight. I never had to replace any of the bracket orders. That being said, my backtests show relevant differences between ICS and synthetic spreads. I am not exactly sure why that is, but it might be because of the way the exchange handles the execution, resulting in different behavior in terms of levels, slope and curvature (of the modeled relationship of the individual contracts). I shall have to investigate further.
That's because IB is holding your order on their side when the platform sees you're trying to use a GTC order with these specific ICSes. Also if you're using a stop and not a stop limit it'll be on IB in all cases as CME doesn't support stop market orders natively (only stop protect or stop limit). When you mention testing these ICS against synthetics, what are you doing? Taking the per day delta from some point in time and reconstructing what the absolute price would be then comparing it to the synthetic? I wouldn't be surprised if there are small differences as the ICS should be tighter hence it'll throw off the chart intraday. You'd have to reset the delta at the start of each session based on the absolute price of the synthetic in order to have something realistic. If you don't do that every session your ICS-reconstructed-absolute chart will drift over time. These ICSes are purely execution instruments. They're not really meant to be charted or have orders placed on them long term. Most people would chart the synthetics based on *yields*, then use the ICS to get in or out. cc @Martinghoul @s0mmi
Thank you, I did not know that. Using the delta in price would be one way, yes. It gets worse if you add volume. I will focus my attention on the synthetic spreads. It does add another layer of complexity to backtesting since there are several methods on how spreads are executed.
If you use resting stops use a stop limit with a wider range such that'll become a marketable limit order if hit. The key takeaway here is that we're getting your stop on CME and not IB. I do this for all my stuff. Around 10-20 ticks or so of slippage allowed depending on instrument. Usually it only ends up being 1. IMO you don't want to do it via the synthetics. The ratios change and that's a complete hair shirt to calculate over time. Plus rollover and all that. You want to chart the yields against each other which means you need the data specifically for yields. You then use this data to determine entry, exit, whatever. Exchange supported stops are dead in the water for all of this because there is no instrument usable unless you're trading purely intraday.