The Long Short - Deep Fundamental Pair Trading

Discussion in 'Journals' started by comfy_slippers, May 2, 2010.

  1. Each month i'll post 20 Long & 20 Short trades. Stocks will be evenly weighted & there will be no bias as to which sector they trade. The portfolio will be rolled over at the end of each month where 40 new stocks will replace the existing porfolio. I will discuss stops/position sizing/risk management later in the thread.

    My development of as a trader wouldn't have been possible if it weren't for the generosity of many and I guess I feel it is only right to return the favor.

    The journal is here for myself, and readers, to network, swap ideas & generate constructive debate.

    So, with formalities out of the way, grab a coffee, your comfy slippers & enjoy the thread!
     
  2. I dabbled with traditional methods of Pair trading with mixed success... I broke even over 18 months... After commissions, spreads & margin interest, I guess you could call it a small win.

    The common theme was many winners being wiped out by one or two large losses... Sometimes 3-4 months of gains were wiped out with one trade....It happened time and time again.

    Over time I realised by running with a fixed limit, but no stop loss, you create a negative skew.... It also became impossible to think in terms of risk vs reward when you don't have a pre-defined stop because your loss is potentially unlimited. Any seasoned trader will stress the importance of fully appreciating Risk vs Reward and without this measure I began to lose confidence.

    The market then began its education course on the pitfalls in relying on backtested results. I fell victim to Survivorship Bias numerous times as I kept experiencing "anomalies" that conveniently never showed up in m backtested results.

    I was still a strong believer in the advantages of running a neutral portfolio, but I decided a more logical approach must exist.
     
  3. These trades were opened on Friday:

    20 Long stocks
    BRNC – 4.99
    CEP – 3.535
    CLFC – 6.69
    CNO – 6.08
    ENER – 7.33
    FCAL – 2.995
    FRME – 9.138
    HAWK – 16.72
    HMPR – 2.335
    HTCH – 6.25
    IMN – 11.285
    LACO – 2.27
    MIR – 11.75
    NGAS – 1.586
    NUHC – 3.65
    RRI – 4.08
    SNBC – 5.58
    SUPR – 3.71
    TWMC – 2.437
    UWBK – 1.747

    20 Short Stocks
    AGAM – 16.93
    AIQ – 5.54
    ALSK – 8.47
    ARB – 8.93
    BLC – 4.99
    CKEC – 17.42
    CNSL – 18.757
    DLX – 21.75
    ENZN – 10.83
    EVC – 3.232
    FGP – 23.00
    FRZ – 4.295
    JBT – 18.47
    KOP – 29.05
    NKTR -14.39
    OMN – 7.953
    OSTK – 19.29
    RCNI – 14.72
    SONC – 12.04
    STXS – 4.97

    All excecuted through IB.
     
  4. I see 40 stocks right there, $40 to initiate the trade to go in , $40 mininum to get out.

    commissions must cost a lot trading like this?
    Whats the cost of commissions % compared to your net profits
     
  5. Claudius

    Claudius

    Having exactly the same experience as you. After four months of pair trading, I have an 80% win rate, 18% small losses and 2% huge losses that blew my slowly accumulated profits out of the water overnight.
    My account is more mean reverting than the pairs.


    Good luck with the new strategy.
    However I don't see how it's going to be anymore successful than the old one. Is stock selection solely based on fundamentals? One month is a pretty short time frame for fundamentals.
     
  6. i'm on the unbundled comission structure, and most months able to generate a rebate back from the exchange for adding liquidity..


    Some of the stocks are not overely liquid and paying away the spread in the long term can erode the edge. By keeping your market impact small the volatilty they offer makes the hassle worthwhile
     
  7. Hi Claudius, Good to hear from you.

    To answer your questions i'll answer in two parts.

    (1) Because I am not trying to replicate back tested performance, I don't have to stick to a rigid rule-set. Your backtested results probably didn't include stops so neither do you. By not employing a stop, and having a fixed profit target, you will always cut short the black swans that work in your favour, but hold the ones for an unlimited amount of time if they work against you. No matter what the cost. How on earth is that going to even out in the long run?

    Sounds a bit silly doesn't it? Never to worry I did the same for almost 2 years and was oblivious to such simple logic.

    Having a stop in place WILL lower your win rate but you'll benefit from a positive skew. Our brains aren't wired to feel good taking frequent small losses so we just need to pony up and accept we aren't always going to get it right.

    Please note: This is just my opinion and what has worked for me. I do know of one or two traders who successfully trade without a hard stop and that's fine, whatever works for the individual. But in my world it makes sense that if things aren't working out just cut it and run. Look for other opportunities and redeploy your capital elsewhere instead of trying to max you winrate % and "happy feelings"

    Tomorrow I'll take some time to post the second part in regards to why I believe certain fundamental measures are better for predicting reversion. I am absolutely stuffed, just spent the weekend surfing up in Byron Bay, Northen NSW......Great part of the world...Get up there readers and don't forget to pack the Slippers!
     
  8. I’m of the belief that reversion to underlying value of a company is stronger than most classes of mean reversion.

    <b><i>Measuring the value of any company is subjective in itself and I guess this is where the edge lies.</b></i>

    I eliminate analysis that makes forward, unknown assumptions. My staple ratios are Price to Cash, Price to Book, Debt to Equity & I pay close attention to the Enterprise Value. There are a few other bits and pieces that I won’t disclose I hope you understand this.

    Instead of thinking "A month to realize fundamental edge is too short a timeframe" I think "I want my money in the 20 most attractive opportunities available"...there is enough movement in the small caps to warrant a fresh set of stocks each month.

    Whether the gap between price and value has closed or not i'm not concerned, just as long as my money is parked in the most attractive opportunities available at that time. If the edge exists then then just let it play out.

    I am also being paid to roll the positions over each month via liquidity rebates so that also adds to the bottom line.
     
  9. Claudius

    Claudius

    I would agree with a lot of what you say here. Pair trading can feel a bit like writing OTM options, a fixed and pre-known upside and an open ended and unknown downside. The exact opposite of what a good system should be.
    But on the other hand, it is a system based on sound logic that gives clear entry and exit signals.

    As you pointed out, survivorship bias in back-testing is a real worry with both selecting pairs and working out the optimal signals. Until recently I had been selecting pairs with a bias towards past profitability. The result was that I had a lot of micro cap stocks with low volume and high volatility. Great historical returns on paper, but exactly the kind of stocks that can suddenly gap 30% and destroy a month of gains.

    I recently weeded out the micro caps and started to focus on the historic consistency of the relationship. But yes, it still assumes that past patterns will hold into the future.

    I've previously looked into using stop loses, and I'm still not convinced. Mean reversion will always supply an exit signal at some stage, and I was using a pretty short MAV that caught up with the ratio pretty rapidly. For every dollar saved by a stop loss, I probably would have lost a similar amount by not allowing a trade to run its course.
    On top of that, my two biggest losses have been opening gaps that no stop loss would have saved.

    How does your fundamentals system gets around the problem of chasing losers ever downwards, and cutting winners short? If a stock on your long list goes down, surely that makes it more attractive as a long, the cheaper it gets, and vice versa for the short list.
    At the end of the month, all your losers will still be there, and your winners will have dropped off.
     


  10. Yeah same payoff diagram. the only difference being you are paid a risk premium with the oppies and get nothing long/shorting the stock.


    Yeah but you are still creating a negative skew. No large, multiday move will ever be in your favour. These large moves just so happen to be the ones that have the largest impact on your account so why not have them on your side rather than against? You will sacrifise a few small winners but i'm fine with that.

    You can also visualise this as an edge in the market. Most investors find it physically hard to take small losses and would preffer to hold onto losers. By cutting the losses you straight away set yourself ahead of the pack and it actually feels good knowing many wouldn't have the balls to cut and run.




    The winners are never closed untill the end of the month. So I will let them run for the full 4 weeks...The losers are cut whenever my stop is hit and I replace them with another short to remain neutral for the rest of the month. Each month starts fresh.

    Yes theoretically it looks more attractive but the downfall of holding on means I lose the positive skew.

    Does that make sense?

    Think of it in terms of trying your best to create a positive skew....the only way is to have a hard loss on your losers and let the winners run. It may contradict your system somewhat but you doing this should ensure you are net positive from large events over the long term.
     
    #10     May 5, 2010