Home Forums Main Forum Basics: How to Get Started with Curvature

  • Author
    Posts
  • Curve Advisor
    Keymaster
    Post count: 612
    #995 |

    I think I started off a little too fast and did not cover some of the basics as thoroughly as I could have. Part of this was just answering member questions, but it would probably help a number of readers to just go over what I think are the very basics, before you start curvature trading. Assuming you want to look at curvature trading (the 4 part series I am writing for the CME will highlight some of the reasons why you should want to), below are the general steps I would recommend:

    1) Get a basic book on Eurodollars. I have Galen Burghardt’s “The Eurodollar Futures and Options Handbook,” and I think it is pretty good, but it could probably use an update, as it is 11 years old. Just read chapters 1-4 for now. Looking on Amazon, Stephen Aiken has a more recent book called “STIR Futures.” I have not read this book, but looking at the table of contents, it looks like it should cover most of the very basics you should have. This will not cover the theory, but will probably give you the more practical nuts and bolts. I just want you to know what the various details of the ED contract are, and what rate it is that you are trading. You should be aware of the contract specifications. Because… BEFORE YOU TRADE ANYTHING, YOU SHOULD KNOW WHAT ALL THE DETAILS ARE!

    2) Make the spreadsheet as outlined in the “Basics: Basic Eurodollar Analytics” thread. This will get you familiar with the basic structures involved, and where they are trading.

    3) Make sure you understand when some thing gets rich or cheap, as per the “Basics: How Rich or Cheap?” thread. This will help you develop a better feel for when to get in and out, and how to scale in and scale out (see “Basics: Scaling into Trades”).

    4) Start looking at specific trades. To start, just look at “simpler” analyses of historicals (ie look at charts) – I wrote somewhere that it is not necessary to do that huge historical sheet (as described in the “Data Set Methodology” thread) – you can sometimes “make do” with what your software vendor provides you with. I will try and put up some more educational case studies in future months.

    5) Develop a P&L sheet (see “Basics: Tracking P&L” thread). There is no substitute to understanding how something trades, than to follow it closely. Use the P&L sheet to develop a better understanding of your P&L, which will also help you develop a comfort level with structures, and give you a better feel for how something trades in various market moves.

    FOR ADVANCED USERS (this is not necessary):
    6) If you have time and the Excel aptitude, make the huge historical spreadsheet as described in the “Data Set Methodology” thread.

    7) Read the more “theoretical” parts of Burghardt (or parts that you find relevant in your trading).

    [I may revise this post as I think of new things]

  • Curve Advisor
    Keymaster
    Post count: 612

    When you think you are ready to trade curvature, consider the following:

    * You can learn just as easily from paper trading. Make sure you are comfortable with the daily closing volatility, as well as the actual trading volatility. Depending on your set-up, you may also have to check about any capital impacts and risk management impacts. If you think you have a strategy/system that works, you should be back-testing and getting additional feedback.

    * Make sure you have a good way of looking at where flies are trading. Just taking the “last price” and calculating a butterfly price off of that is probably one of the worst ways you can look at butterfly prices. If you get streaming quotes to a spreadsheet, consider setting up something similar to what is described in the “Basic Eurodollar Analytics” thread. If your trading system does not provide live fly pricing, you can get 10 minute delayed quotes on flies and spreads for as little as $20 a month through the CME e-quotes page (http://www.cmegroup.com/market-data/real-time-quotes/), so you can see what the actual market and trading range is on various structures. They have a 2 week free trial.

    * Make sure your cost structure is conducive to trading a lot of flies. In general, the potential net gain on a trade should be on the order of 3x to 4+x the cost of execution (brokerage, exchange fees, etc). Many brokerage houses will give you a substantial discount if you intend on trading more volume.

    * And most importantly, be a Forum member! One of the advantages traders have at a bank or hedge fund is that they have colleagues to bounce ideas off of. Whether you are an institutional or individual trader, having an additional point of idea exchange is always good. But this is even more critical if you are an individual trader. You can ask questions you may have through the Forum. You can also ask for feedback on a trade here, before you commit any money to it.

  • ed8ge
    Participant
    Post count: 4

    Joseph, I’ve spent a few weeks reading and re-reading all your posts, excerpts and CME papers, and hugely appreciate the depth of commentary, analysis, insights, advice and guidance across the range of curve topics. I might have escaped blowing out 25% of my account equity when I placed my first (double fly) ED trades the day of the Fed meeting in March ’04 if I’d had your content to educate myself prior !

    In the examples and excerpts, profit targets are often in the realm of several basis points. As a private trader with Tier 1 costs and paying standard commissions (i.e. around US$2 per side/per contract), it wouldn’t appear to be viable. But I might not be seeing the whole picture. Is there sufficient movement in the flies in these market conditions to find prospective trades that could potentially satisfy your Risk/Reward ratio mentioned above?

  • Curve Advisor
    Keymaster
    Post count: 612

    $2 per side is very high to trade curvature, but possibly common for private traders. Larger customers can pay less than $0.30 a side. That is not to say you couldn’t find curvature trades to do – however, a round trip on a butterfly is 0.64 bps. That means you need to make at least 2.5-3 bps gross on a trade. There are trades available – especially if you want to express a view, but this would make it difficult to do some of the smaller trades (like 6 month flies). I think most brokerage houses will give you a volume discount as you trade more contracts. Since butterflies and double butterflies tend to be contract-intensive, this may allow you to negotiate a lower rate. There is definitely a chicken-and-egg issue here, as lower brokerage would lead to higher volumes, and higher volumes lead to lower brokerage.

    The other issue you may have being a private trader is that of margin/carry costs. I do not know what an individual gets charged to carry a butterfly position overnight. It’s hard to trade butterflies in a high frequency manner unless you had rock bottom brokerage. This means that you will have to hold a trade for days, weeks, or even months. So unless you had excess capital, this may be (somewhat) restrictive in terms of your ability to take risk.

    As I have time, I’ll look into this more. Most of the Forum members and CA subscribers work at banks and hedge funds, so I am not sure what the offerings are for individuals.

  • ed8ge
    Participant
    Post count: 4

    I’m with IB (not a good platform for spread trading) and have a simulation account, so I can place some trades on that to check the O/N margin impact. According to CME Group website, minimum overnight margin for, say, EDM15-M16-M17 fly is US$300.

    Presently I’m on a Reg T basis; at least portfolio margin would be better if and when I decide to allocate at least $100k to this. Actually, I’d figure on losing at least a whole basis point in costs after I factor in my sub-optimal executions. Good point about the chicken-egg scenario.

    Okay, seems the learning curve plan might be: concentrate more on the 1 year flies, keep it simple, and – as per your (4) above – use historical comparisons as the main foundation for trade ideas then ?

  • draker
    Participant
    Post count: 5

    I am going to put this out there and you chose to do what you will with it. I would recommend leasing a cme IMM seat for about 850/month. That would reduce your trading cost to approximately 75 cents per car. If you are trading a few hundred flys and double flys, it pays for itself rather quickly. If you want to drop your rates even more dramatically buy a seat. Not the most practical thing but if trading is your business you want to keep costs as low as possible and focus on trading.

  • ed8ge
    Participant
    Post count: 4

    Okay, seems the learning curve plan might be: concentrate more on the 1 year flies, keep it simple, and – as per your (4) above – use historical comparisons as the main foundation for trade ideas then ?

    Oh, I’ve now read the General Questions thread thoroughly and see that this question is answered there.

    @draker: thanks for the idea, and especially the figures. At first I balked, but as you point out the overhead could be readily recovered. I live in Asia so that maybe difficult; OTOH, CME has spent $$ out here promoting itself and incentives. Of course, I have to prove to myself I have the ability first….

  • draker
    Participant
    Post count: 5

    75 cents was a gross estimate because I am certain of IB’s commisson surcharge. Most of the fellows I know who lease are paying close to 50-60 cents per contract. Regardless, here is a cryptic but official pricing list from the CME. http://www.cmegroup.com/company/files/CME_Fee_Schedule.pdf

  • ed8ge
    Participant
    Post count: 4

    Much appreciated!

  • Garvit Omer
    Participant
    Post count: 5

    Hi, I have read almost every article and forum Q/A and just loved them.

    My question:
    1. How do you think that the curve should look like this and not the current way? Do you assume any hypothesis? Or more inclined toward fundamentals (ex: dot plot) ?
    2. As you have so much experience in EDs, I would like to ask do people use “quant” methods in ED, i.e.Do they use more mathematical approach toward curve trading like curve fitting OR more gut feel and discretionary trading?
    Thank you!
    3. I have an excel sheet were I see the curve of 3m, 6m and 12m spreads and flies till Z1. Now, I don’t able to think that curve should be more like that way and not what it currently is. That is I don’t know how should I approach it!

    Thank you.

    • Curve Advisor
      Keymaster
      Post count: 612

      Glad to hear you like the CA. Here are replies to your questions:

      1) The way the curve should look is purely dependent on the current environment. I think some people use growth and inflation models, some people like to think about the path of rates (like the dot plot), some people look at supply and demand for fixed income, and some people look at things like technicals. You will find that most systematic models of the curve will project drastically different rates from the market rates at various times. When I was saying in my Dec 18, 2016 commentary that the back of the curve should be “smoother”, I was merely suggesting that the further out you go in time (3-5 years), the more uncertainty there should be, so there should be fewer sharp spikes. However, we could always get an unusual positioning move that could cause a spike.

      2) I’m not sure what you mean by “quant methods.” There are many ways to use quantitative methods in trading interest rates – curve fitting, high frequency trading, and curve modeling to name a few. Most places seem to have a focus on high frequency trading using various proprietary algorithms. I tend to think having a discretionary curve view overlay on top of any systematic trading is best. We’ve seen a lot of models go bust recently that probably relied too much on historicals, with not enough emphasis on trading common sense.

      3) I have a thread that discusses Basic Eurodollar Analytics, and discusses some of the main structures you should look at. Take a look at that, and post any questions in that thread.

  • Curve Advisor
    Keymaster
    Post count: 612

    Viewer Mail:

    In some of your posts you talk about the non-linearity of returns and that taking a non-directional curvature approach is a good/positive thing to do – in general. Let’s say I put on any of the following non-option futures trades: a) long an ED contract b) an ED fly c) an ED condor d) a double ED fly …. etc. For each of these positions, if it moves in the direction opposite of the position I’ve put on then I lose money. I may being a bit pedantic, though I hope not, but aren’t all such trades “directional”? I’m leaving out options trade where you may sell premium and it’s where the underlying doesn’t go which is important.

    Reply:

    You can put on any structure, and it can either move up or move down. There are very few structures that can’t move against you (there are some that won’t). But the point of non-linear “directional” structures is to find a structure where the P&L profile of the structure is better than doing the outright single contract trade. As an aside, there are trades you can do that aren’t “directional” but are more rangebound, and this is a different type of trade.

    As a very basic example, let’s say in a hypothetical trading environment (not the current environment) you are bullish and want to buy ED9. If ED9 rallies 10bps you make 10 bps and if ED9 sells off 10bps you lose 10bps. A single contract is very straightforward, easy to get into and out of, and lower in transactions costs. Let’s hypothetically say that for reasons x, y, and z, you feel like on a 10bp rally ED5-ED9 calendar spread will lose 2bps, while on a 10bp selloff it will make 4 bps. Doing some multiple of the year spread is much better risk/reward than doing the outright trade.

    Of course, trade selection is important – you have to pick structures that make sense for the environment. There will be also be some variability of the structure’s price relative to the single contract. So sometimes, you may want to diversify your directional view by doing some of both the non-linear and linear trade. However, it’s not hard to see why a lot of experienced traders prefer to look for structures with non-linear payouts.

You must be logged in to reply to this topic.