Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
I'm not sure what's going on. I've charted the NG October18 1 month fly, 3 different ways, and the charts look VERY different. It could just as easily apply to CL spreads or anything else, so I'm missing something very fundamental with the basics. It all came about while I was wanting continuous charts for NG and CL flys to look at the scatterplots etc. As mentioned I've attached the NG october18 1 month fly charted 3 different ways, and showing the autospreader configuration:
1. I've charted the exchange traded fly,
2. Charted the fly constructed out of 2x exchange traded calendars,
3. Charted the fly using individual contracts.
I'll give my guess. 2 and 3 are incorrect, because when they are created with the software, unless the times of the trades line up perfectly, the results will be off.
So, let's say you have a 60 minute bar. And let's say the last trade for fly #1 was one minute ago, and the last trade for fly #2 was 20 minutes ago. The 60 minute bar created will use those datapoints (they were the last traded price), which did not happen at the same time. Hence, they are incorrect.
Does that make sense?
(You can really see this impact when you create the fly out of individual contracts. NO WAY the fly varies as much as the bars show.)
With regards to chart #1 are you sure you have the same instrument? In chart #2 you have Sep-Oct and Oct-Nov so a Sep-Oct-Nov fly. In Chart #3 you have Sep -2 Oct + Nov which again is Sep-Oct-Nov fly but in chart #1 I think you have the Oct 1 mth fly which is Oct-Nov-Dec not Sep-Oct-Nov.
I would also add that the NG Butterflies do not trade as much as the CL ones because NG is so much more seasonal in nature. So your chart #1, other than I think being the wrong instrument, has the problem that it really doesnt trade at all. When they do trade they normally will trade within season. So Summer is Apr-Oct, which means the JKM, KMN, MNQ, NQU, QUV flys will all trade. Winter is Nov-Mar so the popular flies there are VXZ, VFH (3 by 2 unbalanced, trades OTC or with autospreader) and FGH. You will very rarely see cross-season flys trade, specifically GHJ, HJK and UVX because of the significantly different volatility of the spreads.
My suggestion would be to do what you have above but do it for the CL J-K-M fly and see if they look similar.
1. Chart 1 was found in the TT search bar when I typed in: NG oct18 1mo butterfly... You are right of course. I thought of the fly as being the body instead of the first wing.
2. ...In that case Chart 2 was a Sep 1month fly.
3. ...And Chart 3 was a Sep 1month fly (see autospreader below)
In that case 2 and 3 are identical, but chart very differently because of how often they trade (Thanks @kevinkdog ).
Forgetting about NG, and focusing on any spread charts ... So how should I construct robust continuous synthetic spread charts (not adjusted), and use the results to trade the equivalent exchange traded flys or calendars? The synthetic and exchange traded charts are different, even just a little bit.
Thankyou for the info about NG as well, and I'll go back testing the CL flys as suggested.
Trading: Primarily Energy but also a little Equities, Fixed Income, Metals and Crypto.
Frequency: Many times daily
Duration: Never
Posts: 5,049 since Dec 2013
Thanks Given: 4,384
Thanks Received: 10,206
I'm not sure I understand the question. I know people think of outright contracts and backward roll adjust them - which while making sense for automated trading, in reality human traders look at continous non-adjusted charts - but I'm not sure it makes sense for spreads. Maybe if your looking at the front month spread you can do it but how does it work for later month spreads. Consider a Dec-Dec spread. This will behave very differently in say January 10 months before expiry than it will in November, one month before expiry. Now you could plot say the Mth3-Mth15 spread, which would keep a more constant maturity but the problem you have then is that if month 15 is Jun or Dec, you will have a lot more liquidity than the other 10 months of the year, which will probably distort your chart.
What I do, is plot the actual spread and adjust so that different spreads or years hare plotted with the same days to expiration. Here's an example of something I did a few months back in an NG thread, discussing how Mar-Apr behaves. The X-axis is days to expiration.
1. I want to plot any synthetic continuous adjusted spread chart, and trade the associated exchange traded instrument, eg trade the exchange traded ICS Treasury spreads, etc. I just got disillusioned when the synthetic spreads for the same ideas as shown above chart differently, so not sure where to go from here.
2. ALSO I have been interested in value trades such as yours and Joseph Choi from Curve trader. I got stuck about a year ago on Joseph Choi's website, trying to build accurate synthetic adjusted continuous fly charts out of EOD outright data which he sort of shows how to do, and use data from Quandl. Your approach is simmilar, except you are using non-adjusted data. I think you're synthetically constructing the fly charts out of the exchange traded calendar data. The Curve Trade's actual trading was done with exchange traded flys and calendars. I'm trying to duplicate the charts of what both of you are doing. When I get the charts sorted, then it's easy to get the scatter plots etc sorted. I understand that you are using Morning Star and data sources which may be outside of my budget, but I thought I could use the Quandl option for very long term charts, or construct synthetic charts in TT
So the spread goes up near $4. At $10 per $0.001, I guess that would mean you need to be able to handle a worst case $40K drawdown / contract (plus margin) if you were going to fade this chart back to zero. Just one idea which is out of my reach.
I'm thinking about your last post and going through your other thread
Trading: Primarily Energy but also a little Equities, Fixed Income, Metals and Crypto.
Frequency: Many times daily
Duration: Never
Posts: 5,049 since Dec 2013
Thanks Given: 4,384
Thanks Received: 10,206
Like Choi's stuff and have even tried dabbling in Eurodollars. So far I've been successful in terms of win rate, but in terms of raw dollars it's negligible. I would need to have such large positions for the PnL to be meaningful, and I'm not prepared to do that at my current level of understanding of the market. I believe (but could be wrong) that what Choi is doing, while similar is for a different purpose. When the prompt Eurodollar contract is 20 days from expiry, it's telling you what the yield curve expectation is for a 90 day loan starting in 20 days time. When it's 10 days from expiry, it's telling you what the yield curve expectation is for a loan starting in 10 days time. While similar these are actually different (20 thru 110 vs 10 thru 100). Hence I believe what he is doing, is removing/stripping the yield effect of the additional days so that everything is perfectly aligned all the time. ie, the prompt contract is always showing what the yield is for days 0 thru 90, no matter how many days from expiry we are. I tried replicating his methodology as well but could never get modelling the year-end turn to work. As such the eurodollar tradinig I have done involves doubleflys of the same expiration month, ie Z0-Z1-Z2-Z3. Your right I do use Morningstar data to do this, it's just a very easy way to get a lot of futures data into a spreadsheet. Lining everything up and building the charts it's actually a manual process in excel though and not a feature of the Morningstar data.
The March-April spread in Natural Gas is the preeminent example of a market that can stay irrational longer than many traders can stay solvent! To see other ways of how I look at this type of spread data, take a look at the charts in this post and the one following it.
As shown above the Exchange Traded CL JKM 18 fly is very similar to the Synthetic fly constructed out of 2 calendars using TT autospreader. The fly constructed out of j,k,m outright contracts seems a little similar to the calendar fly using TT daily charts, but there are differences. I'm a little worried about the differences.
The following was charted using a free spread chart website using outright prices. The shape seems similar to the calendar fly. Both TT charts and the free charts use outrights to construct the fly, but the free charts seem more accurate or closer in shape to the exchange traded fly:
Some questions:
1. What's wrong with the TT daily outright fly spread charts? Is CQG a better option?
2. Is it safe to model these flys and calendars using charts constructed out of outright settlement prices?
3. Is this free spread charting a viable option to simplify a portion of the analysis? If it is a viable option it allows me do do this:
