Home Forums Main Forum Oops… how to fix an error

  • Author
    Posts
  • Curve Advisor
    Keymaster
    Post count: 612
    #1656 |

    My family used to own a high end construction company in New York City. We did a number of projects that appeared in magazines (Interior Design, Architectural Digest, House Beautiful, etc). One thing that my father used to say is etched in my mind… One sign of a good craftsman is in how well they deal with problems. This is also true in trading. Not everything is going to work out as planned: you buy too many of something, or the markets keep moving against a position you like, or something completely changes and you need to get out asap. So what to do when you are left holding something you didn’t want, or you no longer want?

    The most obvious thing would be to just unwind… even if it means having to cross a bid-offer spread. But many times, this will not be highest expected value alternative. Assuming you don’t NEED TO “completely undo” something (like your manager gave you the tap on the shoulder to reduce), there may be a better way to “lessen” any negative impact, or even turn your situation into a positive one.

    I thought I would go through some examples of things you could consider if you ever find yourself in an uncomfortable position.

  • Curve Advisor
    Keymaster
    Post count: 612

    Before we get to fixing the “error,” we should have a discussion of the importance of PLANNING and PREVENTION. One way to “fix” an error is to not make the error in the first place. Make attempts to reduce your chance of error. Before you do ANY trade, you need to think ahead. “Be Prepared.” I always tell my son, “Don’t be a Level 1 thinker.” I made a slightly different reference to multi-level thinking in another thread (http://www.curveadvisor.com/forums/topic/game-theory-and-trading/). But basically, don’t just sit there and say, “hey, this looks cheap… let me buy some.” Think a step or two ahead and:
    * Think about how best to execute, including how to minimize transactions errors.
    * Think about the ways in which an error could occur. A monkey could be trained to buy something, have it go up and and sell afterwards. Put more thought into the other paths your trade could take.
    * Have a contingency plan in case things don’t move in the expected direction.
    * Could there be any issues with exiting the trade? Are there scenarios where you may want to keep a piece of the trade? Are there any logistics that need to be worked out?

    There’s a lot more to doing the trade than the initial order. The above may seem like a lot of overkill. it may be in the beginning, but after a while, it can literally take seconds to consider all of the above.

  • Curve Advisor
    Keymaster
    Post count: 612

    EXAMPLE 1: You decide you like buying EDZ8-Z9 spread. It could be for various reasons: (1) fundamentally fits your view, (2) it looks low relative to the curve, (3) the chart looks good, etc. And you could be looking to make anywhere from 0.5 to 10bps on it, depending on how you like to trade. However, it “goes bad”. What do you do?

    1) Determine why it went bad. Why did it not perform as you had expected? The reason we do this is to learn for next time. And also to see…
    2) Do you still like it?
    a) If “no,” you have two choices:
    (i) Just get rid of it. This is the simplest thing for something you don’t like.
    (ii) Can you convert it to something you like NOW? For example, it may be that now, you like being short EDZ9-Z0 spread because of recent markets developments. Then you can sell EDZ8-Z0 spread to convert your old position into the new one. You saved yourself 1 leg of brokerage, but pennies add up.
    b) If “yes,” you have a number of choices:
    (i) You can hold (or even add). However, this is an “error” thread, so we’ll assume this is not a good alternative.
    (ii) Can you do something that still has the same idea, but with limited downside?

    [to be continued]

  • Curve Advisor
    Keymaster
    Post count: 612

    We have narrowed the above flow “diagram” to one alternative: what to do with a trade you like but need to cap the downside on the trade.
    The ladder of trade possibilities (in increasing order of contracts) goes like this:
    1) single contract (1 contract)
    2) spread (2 contracts)
    3) fly (~4 contracts)
    4) double fly (~8 contracts)
    5) quadruple fly (~16 contracts)
    The general idea is this… if you need to reduce risk, you move down; if you want to increase risk, you move up. In our case, if you think there is value in the trade but need to cover some risk, take a look further down the list. We bought a spread (Z8-Z9). Ask yourself if there is another spread you think it is particularly cheap against. Instead of selling the spread you bought, sell that other rich spread instead.. You may still be able to pick up some of the value in the spread there is, if it outperforms vs the other spread you sold.

    What you are left with is long one spread and short another… in other words a fly or condor. For example, you can sell Z9-Z0 spread to convert your position to a long Z8-Z9-Z0 fly. I am of course assuming your cost structure allows this. But you can basically sell any other spread… you can sell H9-H0 (to convert to a condor), or 2x U9-H0 (to convert to a uneven fly), etc. You can also move two steps down and convert that year spread into a double fly, by selling 2x H9-U9 (thereby converting to a Z8-H9-M9 3mo fly vs M9-U9-Z9 3mo fly), or by selling 1/2 as many M8-M0 spread (thereby converting to a M7-Z7-M8 fly vs M8-Z8-M9 fly). The goal is basically to convert your higher-risk position into a lower risk position that still has some of your rich/cheap view.

  • Curve Advisor
    Keymaster
    Post count: 612

    VIEWER MAIL: I am short U84D [EDU8 1yr double fly] which is just exploding because Z9 has been the brunt of the selling. I was thinking that it could come back once the outright selling stops. But, now that we’ve paused, I’m questioning whether the relationship between U84 [EDU8 1yr fly] and U94 [EDU9 1yr fly] will change that much. U84 seems to be holding, but U94 is falling now that we’ve paused. So it’s making this a bit painful.

    Do you have any thoughts on how to get out and/or what to do.

    CA REPLY: The 2 year high (I think looking at 2 yr historicals currently are reasonable) on U84D is 24.4. So try and get into the habit of looking back at the historicals before doing a trade. I know I say historicals can be crap, but sometimes it’s helpful. And with generic contracts on Bloomberg, you can create custom CIXes to see what the history looks like, and using the generics is close enough to constant maturity data for something like this. I was looking at the U84D and while this is much higher than what we are used to, it can get much higher. You can see that there is a slope bias to this trade [the blue triangles are two year data and the grey circles are five year data]. As we keep steepening, you could lose more money.

    The easiest fix (assuming you like the other reasons for the trade), is to add a slope hedge. EDU8-U9 calendar spread (the front part of U84) rolls up very well. You probably don’t want to be too short this portion of the trade. So consider adding a slope hedge of 45% EDU8-U9 spread. The second chart looks more balanced, and allows you to profit in many more scenarios (assuming history is any kind of guide).

    Of course, if you no longer like the trade, stopping out is always an option.

    Attachments:
    You must be logged in to view attached files.

You must be logged in to reply to this topic.