NexusFi: Find Your Edge


Home Menu

 





Spread / Pairs Trading - the allure and the reality


Discussion in Traders Hideout

Updated
      Top Posters
    1. looks_one josh with 21 posts (35 thanks)
    2. looks_two kevinkdog with 15 posts (23 thanks)
    3. looks_3 TheDude with 11 posts (22 thanks)
    4. looks_4 datahogg with 11 posts (2 thanks)
      Best Posters
    1. looks_one tigertrader with 7 thanks per post
    2. looks_two kkfx with 4.6 thanks per post
    3. looks_3 josh with 1.7 thanks per post
    4. looks_4 kevinkdog with 1.5 thanks per post
    1. trending_up 85,895 views
    2. thumb_up 200 thanks given
    3. group 64 followers
    1. forum 120 posts
    2. attach_file 11 attachments




 
Search this Thread

Spread / Pairs Trading - the allure and the reality

  #1 (permalink)
 
josh's Avatar
 josh 
Georgia, US
Legendary Market Wizard
 
Experience: None
Platform: SC
Broker: Denali+Rithmic
Trading: ES, NQ, YM
Posts: 6,216 since Jan 2011
Thanks Given: 6,752
Thanks Received: 18,136

To be market neutral and still make money--perhaps it's the holy grail of trading. To make money without having to worry about market direction...? Sign me up!

If only it were so simple. As I have examined spreading, on the one hand it is so promising; yet on the other hand, it seems to me (a newbie on the subject) that it is really not much different from trading outright direction. Why? Because the spread trade is trading a relationship, one which varies; we still must decide on a direction, just not the direction of a product--rather, the direction of the spread, whether it will widen or narrow.

[IMG]https://cdn.bmcharts.com/screencast/65CTZqFqM9.png[/IMG]

A typical pairs strategy says to find the mean of the ratio over some period of time, and fade a move away from that mean back to the mean. In the following example is a 100 week mean of the ratio (using the HLC value of each bar) with two standard deviations on each side:

[IMG]https://cdn.bmcharts.com/screencast/BU1Z0NcW.png[/IMG]

As you can see, such a strategy of buying the lower band and selling the upper band would have had some limited success, but some places would have resulted in a large drawdown or loss. Such a strategy depends on the ratio hovering relatively closely around the mean, and since 2002 the NDX has been steadily gaining strength vs the S&P 500, and as such has remained largely below the mean. In other words, it has been trending down. And as you can see, a mean reversion strategy would have failed miserably in 2000 and 2001, as the tech bubble burst and the Nasdaq fell out of bed. Buying it, with a view to a reversion to a mean, would have been disastrous.

So we have a major detail to fill in with this strategy, as is the case with any mean reversion strategy--which mean do we plan to revert to? I used a 100 week value; but why not a 20 week or a 30 day or a 200 day?

Even ignoring this detail, the more fundamental problem is this--we are still trading the direction of the spread. The direction was heavily up in the first two years on the chart. After that it has basically been down. A strategy which had played a move away from a mean in either case would have worked well, particularly in 2000-02 and in 2009-12. But if you did not know better, you might actually think that this chart was the chart of a stock or some other instrument, wouldn't you? So to make money trading pairs, you still have to get the direction right. Why? Because one instrument can just keep rising or falling in relation to the other--it does not have to revert to a mean. Similarly, just as with a stock or future or currency, the ratio can trend--but at some point it will stop. So, you cannot always employ a strategy which goes with the direction of the spread.

In short, you still have to trade this like it's a stock or other financial instrument. So why not just trade outright?

While I see the perceived benefits of spreading, when it's all said and done, I see no real practical reason to do this on a discretionary basis. An arb-style approach with HFT abilities will work for sure, but that market is all sewn up as far as I am concerned. So spread / pairs traders, I welcome your feedback--I would love to find out that I'm missing something simple and that I'm completely wrong, so please prove me wrong here so I can explore this avenue of trading further.

.

Started this thread Reply With Quote

Can you help answer these questions
from other members on NexusFi?
NT7 Indicator Script Troubleshooting - Camarilla Pivots
NinjaTrader
Deepmoney LLM
Elite Quantitative GenAI/LLM
The space time continuum and the dynamics of a financial …
Emini and Emicro Index
Are there any eval firms that allow you to sink to your …
Traders Hideout
New Micros: Ultra 10-Year & Ultra T-Bond -- Live Now
Treasury Notes and Bonds
 
Best Threads (Most Thanked)
in the last 7 days on NexusFi
Get funded firms 2023/2024 - Any recommendations or word …
61 thanks
Funded Trader platforms
38 thanks
NexusFi site changelog and issues/problem reporting
27 thanks
GFIs1 1 DAX trade per day journal
19 thanks
The Program
18 thanks
  #3 (permalink)
 
Big Mike's Avatar
 Big Mike 
Manta, Ecuador
Site Administrator
Developer
Swing Trader
 
Experience: Advanced
Platform: Custom solution
Broker: IBKR
Trading: Stocks & Futures
Frequency: Every few days
Duration: Weeks
Posts: 50,397 since Jun 2009
Thanks Given: 33,173
Thanks Received: 101,537


Josh happy to see you start this thread. I have lined up a seasoned pro for an advanced spreading webinar when I return from vacation. More details later.

Sent from my LG Optimus G Pro

We're here to help: just ask the community or contact our Help Desk

Quick Links: Change your Username or Register as a Vendor
Searching for trading reviews? Review this list
Lifetime Elite Membership: Sign-up for only $149 USD
Exclusive money saving offers from our Site Sponsors: Browse Offers
Report problems with the site: Using the NexusFi changelog thread
Follow me on Twitter Visit my NexusFi Trade Journal Reply With Quote
  #4 (permalink)
 
mattz's Avatar
 mattz   is a Vendor
 
Posts: 2,493 since Sep 2010
Thanks Given: 2,440
Thanks Received: 3,789

@josh. nice post.

I don't trade spread and for the most part I focus on the direction, but this area always been of interest to me.
I have seen that people who utilize this strategy know quite a lot about the major stocks that influence(have larger weighted average) on a particular index. Further, they also observe due to their additional fundamental knowledge when there is a flow from tech (NQ) to industrial (YM) to blue-chip (ES) and vice versa.
So I think the correlations goes beyond just a chart analysis.

I like to observe especially during this bull markets how far the corrections go of one index versus the other.
From the recent corrections we had, the YM has corrected about 800 points from its top, versus the NQ that held pretty solid in my opinion due to the strength of the techs. This for example, tells me that now the NQ is a lot more solid and a long way to go before we see a major corrections on it. While we have encountered unchartered territory in ES and YM with new highs, we are nowhere near on the NQ which is still trying to catch up. It could be that the spread between all should be also done on a long term basis and give us better perspective where things are going.

In general, I agree with you that finding a correlations daily is a hard task.

Trading futures and options involves substantial risk of loss and is not suitable for all investors. Past performance is not necessarily indicative of future results. You may lose more than your initial investment. All posts are opinions and do not claim to be facts. Please conduct your own due diligence. Use only Risk capital when trading Futures.
1 800 771 6748 local 561 367 8686 email [email protected]
Reply With Quote
Thanked by:
  #5 (permalink)
 vegasfoster 
las vegas
 
Experience: Intermediate
Platform: Sierra Chart
Broker: Velocity/IB
Trading: 6E
Posts: 1,145 since Feb 2010
Thanks Given: 304
Thanks Received: 844

From a day trading perspective, I've never figured anything meaningful that would justify the costs. Question I have is does anyone do that or is it purely a swing thing? I dunno.

Reply With Quote
Thanked by:
  #6 (permalink)
 
Big Mike's Avatar
 Big Mike 
Manta, Ecuador
Site Administrator
Developer
Swing Trader
 
Experience: Advanced
Platform: Custom solution
Broker: IBKR
Trading: Stocks & Futures
Frequency: Every few days
Duration: Weeks
Posts: 50,397 since Jun 2009
Thanks Given: 33,173
Thanks Received: 101,537


vegasfoster View Post
From a day trading perspective, I've never figured anything meaningful that would justify the costs. Question I have is does anyone do that or is it purely a swing thing? I dunno.

Most of what/who I know in this area -- it's all about speed. That is why there are hot topics like X_Trader's Algo Design Lab (ADL) with server-side scripting, or CQG's Spreader (also server side), and now CTS T4 has an Autospreader (soon to be server-side). So the focus is all about speed and moving as much volume as quickly as you can.

I am aware of some others that take the "swing" approach to spreading but I admit my understanding of that starts to break down. Clearly there is an argument of operational overhead and efficiency just like directional trading, if your costs are 30-40% of your gross because you are going for 1-2 ticks, then swinging can dramatically improve it. But in my head I always picture spreading in terms of high frequency arb, where an edge is largely eliminated over time.

I've seen some really "odd" spreads, like not just calendar spreads but weird mixtures of energy and softs and the sorts. I can understand spreading similar indices like @mattz was hinting at, but I am not sure about the complexity behind spreading more complex vehicles, especially with multiple legs. Luckily that is an area of expertise in the trader I will be bringing on for the webinar.

But this is absolutely not an area I know enough about. That is why I wanted to bring on a seasoned pro trader to talk about it for the group. I have a lot to learn in this area. I can't even name any spreaders on the forum, I am not sure who is actively working this angle.... Maybe with @josh's thread they'll come out of the woodwork and clear some things up?

Mike

We're here to help: just ask the community or contact our Help Desk

Quick Links: Change your Username or Register as a Vendor
Searching for trading reviews? Review this list
Lifetime Elite Membership: Sign-up for only $149 USD
Exclusive money saving offers from our Site Sponsors: Browse Offers
Report problems with the site: Using the NexusFi changelog thread
Follow me on Twitter Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #7 (permalink)
 
Nicolas11's Avatar
 Nicolas11 
near Paris, France
 
Experience: Beginner
Platform: -
Trading: -
Posts: 1,071 since Aug 2011
Thanks Given: 2,232
Thanks Received: 1,769

Hi,

I think that the idea is indeed to fade deviations of the spread compared to its mean.

But only if the spread is built between 2 instruments which are highly correlated (actually: cointegrated), that is to say that the spread is strongly and quickly mean-reverting. There are statistical tests to check it (Augmented Dickey-Fuller, etc.)

I am not sure that there is an issue with the direction of the spread. For instance, let's suppose that:
spread = 2 * price-of-instrument-A minus price-of-instrument-B
If the spread is above its mean, we could sell 2 units of A and buy 1 unit of B.
The other way would not be logical, I think.

Nicolas

Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #8 (permalink)
 kevinkdog   is a Vendor
 
Posts: 3,646 since Jul 2012
Thanks Given: 1,890
Thanks Received: 7,338

I have been swing trading calendar spreads for a number of years. I trade most of the ags, softs, energies, etc.

Yes, you do have to get direction right for the trade to work.

You can forget about arbing spreads - people much smarter and faster have this locked down.

You have to be really careful when looking at any prices unless 1) they are settlement prices or 2) they are exchange traded spread prices. You can't just throw up a candlestick chart of Nov Beans minus July Beans and expect to get something meaningful. This is because last trade price times will be different. You could accurately chart the exchange quoted spread on Nov/July Beans.

Many times spreads tend to trend more smoothly than the outrights.

Ag spreads many times follow seasonal patterns. Summer 2012 was an exception to that, as I found out.

Many pros in Treasuries are spreading. I know a mid size hedge fund that always and only spreads.

Margins are greatly reduced for spreads, so you can probably trade more pairs, or greater size, than an outright account.

You can easily lose money. "How I Lost 1 Million Dollars..." is a great book on a Soybean spreader who got crushed (no pun intended).

You will pay double commissions for each trade.

Also, any kind of backtesting is a TON harder with spreads, even if you have access to the exchange spread feed (Tradestation, for example, in their infinite wisdom, does not offer the exchange spread feed in their data). You can't really do continuous contracts. I've written giant custom macro spreadsheets to perform some of this backtesting. It is a pain.


One thing to think about: spread trading might have opportunity because so few people take the time to look at it, and learn about it.

Follow me on Twitter Reply With Quote
  #9 (permalink)
 
josh's Avatar
 josh 
Georgia, US
Legendary Market Wizard
 
Experience: None
Platform: SC
Broker: Denali+Rithmic
Trading: ES, NQ, YM
Posts: 6,216 since Jan 2011
Thanks Given: 6,752
Thanks Received: 18,136

Thanks for the replies so far. I took a simulated trade near the close yesterday to help me work through the process. For the current values of ES and NQ, to get a near-identical notional value, I traded 3 ES, versus 4 NQ. This is because my entry price of short 3 ES @ 1690 yields 3*50*1690 = $253,500 and 4 NQ @ 3160.81 yields 4*20*3160.81 = $252,864. So, this yields the best possible market neutral position as best I can determine.

Here is a shot of the positions as of a few minutes ago:
[IMG]https://cdn.bmcharts.com/screencast/PO8Ox9S6.png[/IMG]

And here is ES and the spread below (60 minute chart). You can see that since I took the trade late yesterday, the spread has widened just a tiny bit, hence the trade is in profit a few hundred dollars, with the NQ strength overtaking the ES strength. My idea on this trade was for a continuation of a widening spread (smaller value). As you can see, the Nasdaq 100 has been much stronger than the S&P throughout August, and only since 9/10 has that relationship changed. So this is not a mean reversion trade, this is playing for continuation of NQ/ES strength.


Started this thread Reply With Quote
Thanked by:
  #10 (permalink)
 
josh's Avatar
 josh 
Georgia, US
Legendary Market Wizard
 
Experience: None
Platform: SC
Broker: Denali+Rithmic
Trading: ES, NQ, YM
Posts: 6,216 since Jan 2011
Thanks Given: 6,752
Thanks Received: 18,136



Nicolas11 View Post
But only if the spread is built between 2 instruments which are highly correlated (actually: cointegrated), that is to say that the spread is strongly and quickly mean-reverting. There are statistical tests to check it (Augmented Dickey-Fuller, etc.)

Thank you for the post Nicolas.

In short, I am a newbie at this. The sample trade I have taken and the examples presented are really probably the most basic examples possible. The math involved in doing proper trading this way and doing tests for cointegration requires far more than anything I have mentioned, and even understand at this point. I don't have matlab either, which I probably would need.

Started this thread Reply With Quote




Last Updated on April 14, 2023


© 2024 NexusFi™, s.a., All Rights Reserved.
Av Ricardo J. Alfaro, Century Tower, Panama City, Panama, Ph: +507 833-9432 (Panama and Intl), +1 888-312-3001 (USA and Canada)
All information is for educational use only and is not investment advice. There is a substantial risk of loss in trading commodity futures, stocks, options and foreign exchange products. Past performance is not indicative of future results.
About Us - Contact Us - Site Rules, Acceptable Use, and Terms and Conditions - Privacy Policy - Downloads - Top
no new posts