Spreads

General discussions about futures.
Dave S.
Senior Member
Senior Member
Posts: 38
Joined: Tue Apr 15, 2003 4:51 pm

Spreads

Post by Dave S. » Thu Apr 17, 2003 4:07 pm

Is there any reason one couldn't trade futures spread positions in the same manner as outright futures positions?

My experience with spreads (i.e. long May wheat/short Sep wheat) has been that they are less volatile, the margins are much lower, and the overall risk is reduced. There are undoubtedly situations where this is not true, but in general, this has been my experience.

My method for evaluating these positions has not been what I'm learning about with longer-term trend following, so if anyone has some longer-term experience with this, I'd love to hear about it!

Thanks for any responses!

Dave

damian
Roundtable Knight
Roundtable Knight
Posts: 815
Joined: Tue Apr 15, 2003 8:43 pm
Location: dusseldorf

Post by damian » Thu Apr 17, 2003 6:53 pm

dave, from the perspective of your desk, you can trade a calendar spread just like it was a commodity (with double bro and an extra expiry to watch out for etc). From the perspective of charting/TA/models, you can analyse the spread in the same manner as an outright.

From the perspective of the actual market (ie the exchange), spreads are often quoted and traded as a product of their own and they work a spread order as a spread order, not as two orders, one to buy and one to sell.

I used to enjoy trading spreads. I found that I would often enjoy an underlying spread position that make or lose money with less volatility. Winners were in general a slower, smoother trade.

Not all is smooth sailing, as you say, a spread can cause just as much $ damage as an outright.

Also, never turn a losing outright into a spread. Never turn a losing spread into a winning leg outright.

Chuck B
Roundtable Knight
Roundtable Knight
Posts: 469
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B » Sun Apr 20, 2003 4:15 pm

Dave,

Spread trading on the LIFFE is the way it will be everywhere some day I would hope. The LIFFE engine is designed to excel in trading complicated spreads and combinations in the STIR markets (Short Term Interest Rate futures). The Euribor spread market is unreal -- this is the deepest and most liquid spread market. In a front end like Trading Technologies, you can trade the spreads effortlessly and even show an implied + outright combined order book for a given spread. For example, trading the Sept03-Mar04 spread in the Euribor right now.

Calendar spreads in the STIRs have just recently had huge trends in both directions, and the reward to volatility risk ratios have been outstanding.

The CME has implemented Globex spread trading in ED, but as is typical the CME execution costs are insane, and you have the underlying pit junk. Therefore, I think right now the best spread trading market from a liquidity point of view is the Euribor followed by the spreads on the Eurex financials (i.e. Bund-Bobl). With TT you can also easily spread trade the Euribor versus the Schatz even though they are on two different exchanges -- this is a very popular and heavily traded spread (and probably one of the reasons that the Euro 2yr note sometimes trades close to 1 million lots a day).

Chuck

Dave S.
Senior Member
Senior Member
Posts: 38
Joined: Tue Apr 15, 2003 4:51 pm

More info

Post by Dave S. » Sun Apr 20, 2003 8:37 pm

Hi Chuck,

Thanks for your reply!

Can you tell me more about Trade Technologies, and how this works? I'd also like to know where and how I can start monitoring these spreads.

Thansk for your help!

Dave

Chuck B
Roundtable Knight
Roundtable Knight
Posts: 469
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B » Mon Apr 21, 2003 6:49 am

Dave,

TT is what I've found so far to be the best "front end" software out there. Along with many competitors, TT really got going in 1998-9 when the LIFFE decided that was the way to go (having ISV's -- Independent Software Vendors to have traders connect to the up and coming LIFFE Connect system). Eurex was on that bandwagon with their API too as the original Eurex screen was a bit "limited". I first saw an ISV program (http://www.easyscreen.com) in Sept 1998 when I was visiting a friend at the LIFFE. It was incredible to see what was coming in trading, and that experience helped shape my thoughts of how the future would evolve. I think Easyscreen has done ok since then, but TT and also PATS (http://www.patsystems.com) I believe are the biggest front-end's now.

TT's website is http://www.tradingtechnologies.com. The main program is called X-trader, and you can see some detail of it on the site. X-trader Pro has something called Auto-Spreader which allows you to automatically trade spreads based on your rules. When it can fill the legs of the spread on your rules, it's lightning fast in execution. (as an aside, if you watch the order book in the mini-Dow (YM) versus the e-mini SP, you can see that the majority of the decent size bids and offers are electronically managed spread orders versus the e-mini SP).

As far as data goes, that you have to create in most cases. Trying to create a OHLC daily bar for a spread presents one difficulty since the OHL can occur at different times for the legs. In CQG, I can create a daily chart by using intraday data. If you have CQG, type: CONSOLIDATE(QEAU3-QEAU4,15),D to get a daily bar chart built from 15min data for the Sept03-04 Euribor spread for example. For just close only spread data, you can use something like TradeStation to plot the spread as an indicator.

Chuck

Moodaeng
Senior Member
Senior Member
Posts: 34
Joined: Thu Apr 17, 2003 4:23 am

Post by Moodaeng » Tue Apr 22, 2003 11:27 am

Chuck,
Do you apply a trend following system to Euribor spreads?
Do you exploit the mean reverting nature of spreads of spreads (for example, long ERM3, short ERM4, short ERZ3, long ERZ4)?
Thanks in advance.

Chuck B
Roundtable Knight
Roundtable Knight
Posts: 469
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B » Tue Apr 22, 2003 11:40 am

So far, I have not done any extensive Euribor spread trading. I think there is something to be had there however. Interest rate curve spreads likely aren't mean reverting. Look at the 1 year ED or Euribor spread on a conituation basis since May 2000 :shock: .

Moodaeng
Senior Member
Senior Member
Posts: 34
Joined: Thu Apr 17, 2003 4:23 am

Post by Moodaeng » Tue Apr 22, 2003 12:20 pm

Chuck
Interest rate curve spreads likely aren't mean reverting
I agree. That's why I was wondering if trend following systems worked on them. My testing has been mixed, so I was wondering if anyone had obtained clearer results.

On the other hand, 4 leg trades like
-Long EDM3
-Short EDM4
-Short EDZ3
-Long EDZ4
are mean reverting (at least, by looking at a long term chart)

Thx for your quick response.

Bollinger
Senior Member
Senior Member
Posts: 40
Joined: Wed Apr 16, 2003 2:42 pm
Location: New York City

Spread Leg Ratio

Post by Bollinger » Sun May 11, 2003 6:54 pm

Adding spreads certainly sounds like a great way to add extra markets to your longterm trading portfolio. For starters there are various intermarket interest rate spreads -- NOB (10Yr Note over Bond), MOB (Muni over Bond), FOT (5Yr Note over 10Yr), etc --, crack spreads and crush spreads. Also Intramarket spreads between different delivery months with STIR contracts and other markets that trade further out delivery months, but these are less easy to examine historically.

The question I have regarding intermarket spreads is how one determines how many contracts to put on each leg of the spread, given that you have different underlying assets with different volatility levels. With something like treasuries, where each contract has equal point value and size it would seem to make sense to use a 1:1 ratio. However, traders apparently use a "dv01" ratio to determine the relative values of 1 basis point (.01%) moves in the yield for the two interest rates and then put on a number of contracts on each leg that correspond to this ratio.

I'm sure this is something that comes quite naturally to one who trades bond spreads regularly, but I admit being confused by it. Is there anyone with experience here who can elaborate on this?

Best regards,
Neal Stevens

Kiwi
Roundtable Knight
Roundtable Knight
Posts: 513
Joined: Wed Apr 16, 2003 1:18 am
Location: Nowhere near

Post by Kiwi » Sun May 11, 2003 9:17 pm

Neal,

What is a div01 spread ratio please?

I have used ratio's developed to give the best trends on recent and long term data. So you adjust the spread and measure your trend/system results for each new ratio until you get a "not to curve fitted" result.

John

Bollinger
Senior Member
Senior Member
Posts: 40
Joined: Wed Apr 16, 2003 2:42 pm
Location: New York City

Post by Bollinger » Mon May 12, 2003 6:52 pm

John,

I'm trying to learn more about this and want to make sure I understand it thoroughly before posting further on it or (of course) trading interest rate spreads.

As I understand it, DV01 is a measure of interest rate sensitivity for bonds. It is essentially the price move in a bond for a 1 basis point (.01%) change in the bond yield. Apparently it is used when determining how many contracts to use to hedge an interest rate exposure or to put on a spread position. There's a little on this on the CBOT website.

If it sounds like I'm ignorant on this it's because I am. I'm having a conversation with a bond guy at my FCM tomorrow and hopefully that'll clear things up a bit. If I have a eureka experience I'll report back. In the meantime, if anyone understands this further, I'd love to be showered with knowledge. The next thing I want to sort out after this is whether a 1:1 ratio or something different is appropriate with other spreads -- crush, crack, intramarket STIR spreads, etc.

Best,
Neal

Moodaeng
Senior Member
Senior Member
Posts: 34
Joined: Thu Apr 17, 2003 4:23 am

Post by Moodaeng » Tue May 13, 2003 3:57 am

You are right about the definition of DV01. Suppose you want to put a "curve" trade where you buy a 5 year bond (with a dvo1 of 4.25) and sell a 10 year bond (with a dvo1 of 8.5), then you will have to buy twice as much of the 5 year bond than you will sell of the 10 year bond.

In real life :

- the curve rarely moves in a parallel fashion, so your historical hedge ratio is frequently not the dvo1 ratio. Some people do a Principal Component Analysis of the yield curve to find the "best" historical hedge.

- when you buy a bond future, you do not know which bond will be delivered to you. To determine the "dov1 of the contract", you have to use a model that takes into account all the deliverable bonds...and it gets pretty complicated.

rs

Post by rs » Tue May 13, 2003 4:01 am

Hi,

I used to trade cash corporate bonds for several institutions so I can give you my perspective on this.

You are correct in your definiton of a DV01. It is basically the sensitivity of a bond's price to a change in yield. Thirty year bonds have a higher dv01 than 10 year notes and 10 year notes have a higher dv01 than 5 year notes etc...

So, effectively the volatility is higher on a 30 year bond than a 10 year note. As you have stated this volatility must be equalized when trading a spread ASSUMING that the trader does NOT want any direct directional exposure to interest rate movements.

By trading a bond or interest rate spread which is dv01 matched you are attempting to neutralize your interest rate exposure and gain exposure to movements in the yield curve. So if you re long 10 year notes and short 5 year notes and are dv01 matched then you are hoping that 10 year notes will outperform 5 year notes.

I hope this helps, if you have any more questions I would be happy to try and answer them for you.

rs

Bollinger
Senior Member
Senior Member
Posts: 40
Joined: Wed Apr 16, 2003 2:42 pm
Location: New York City

Post by Bollinger » Tue May 13, 2003 11:05 pm

Moodaeng and RS:

Thanks much for the comments. I'm still a little confused, this is obviously one of those things that will click at some point and all make sense, but that hasn't happened yet.

I think what I'm hearing is that if you want to minimize the directional aspect of the trade and isolate the spread you need to adjust the ratio of the number of contracts for the spread legs based on the relative volatility of the two contracts. That is, if we're doing a NOB spread of 10 Yr Notes over Bonds, and we see that the Bond contract has 20% more volatility than the Note contract, we might put on 20% fewer bond contracts than Note contracts.

But this confuses me -- after all, when examining the spread itself we're simply talking about the difference in price between the two contracts, which it seems to me incorporates the volatility of the two contracts already. That is, if we see the Bond is trading at 114 and the 10Yr is trading at 116 the spread is 2. Perhaps we've followed this spread down from 4 and we believe it has begun trending down. It seems to me this analysis already incorporates the volatility of the two contracts, which, after all have the same tick value and contract size. Surely I'm missing something here since spread trades are done by adjusting for interest rate sensitivity / volatility, but it hasn't sunk in yet.

Regarding DV01, what is the practical way to discover it? Compute it oneself? Look at a Bloomberg terminal or other independent news source? When talking to my FCM today they already knew the ratios for different bond spreads and said the guys on the floor at the CBOT did too, but they didn't seem to know how the actual ratios were determined.

Once again, the comments on this are much appreciated.

Best regards,
Neal

Bollinger
Senior Member
Senior Member
Posts: 40
Joined: Wed Apr 16, 2003 2:42 pm
Location: New York City

Post by Bollinger » Tue May 13, 2003 11:28 pm

Correct me if I'm wrong, but it seems to me we're talking about two different things. One is the price spread between the futures contracts, and the other is the price spread between the underlying bonds. The dv01 computation would seem to make sense when using futures contracts to trade the price spread between the underlying bonds (ie, using the latter of the two to make trading decisions but using futures to effect that trade). BUT if one is actually trading the price spread between the futures contracts (the former of the two) then it seems to me one should use a 1:1 ratio.

rs

Post by rs » Wed May 14, 2003 3:17 pm

I think the point is that as long as you know what you are trading it doesn't matter.

There is no reason why you have to be dv01 matched unless you want to do a straight curve trade.

If you are charting the simple subtraction of two contracts then that is what you are trading and that is fine too. The difference being that you will have directional exposure to interest rates. This is not necessarily a bad thing at all, I am just making clear the difference between dv01 matching and simple spread trading.

rs

Moodaeng
Senior Member
Senior Member
Posts: 34
Joined: Thu Apr 17, 2003 4:23 am

Post by Moodaeng » Fri May 16, 2003 5:23 am

Bollinger.

Suppose you want to buy a 5 year bond (dvo1=D5) versus a 10 year bond(dvo1=D10). The 2 most popular methods to structure the trade are the following :

- Dvo1 method : buy (1) 5year bond/ sell (D5/D10) 10 year bond. Roughly said, your PL will not move if each bond yield moves by the same small amount.

- Historical method. On X axis, plot 5 year bond price. On Y axis, plot 10 year bond price. Chart the best fit line. Your trade ratio is the line's slope.

The dvo1 of a bond is rather straightforward to calculate, but the dvo1 of a bond future is much more complicated because a bond future is a complex derivative of a basket of bonds. (If you want, I can recommend a book on the subject). If you don't want to put too much time in the subject, I suggest you use the historical method directly on futures prices.

Kiwi
Roundtable Knight
Roundtable Knight
Posts: 513
Joined: Wed Apr 16, 2003 1:18 am
Location: Nowhere near

Post by Kiwi » Fri May 16, 2003 7:01 am

I'd be interested in the name of the book :)

John

Moodaeng
Senior Member
Senior Member
Posts: 34
Joined: Thu Apr 17, 2003 4:23 am

Post by Moodaeng » Fri May 16, 2003 9:38 am


Allen
Contributor
Contributor
Posts: 2
Joined: Tue Jun 03, 2003 9:47 am

Index Spreads

Post by Allen » Tue Jun 03, 2003 10:02 am

Hi there,

I just beginning my education on spreads in general and have been looking into intra-day calendar spreads in DJ($10) and the big S&P.

Does anyone have experience there or is it a pipe dream considering professional scalping and execution? I would like to look at some 60 min. charts but my provider can't do them. But there seems to be a tradeable range to be exploited--I just question if it can be done efficiently.

Any insight would be appreciated.

With Respect,

Allen

Post Reply