How about reducing drawdowns using stop loss limits?

Home » Topics » FAQ » How about reducing drawdowns using stop loss limits?
How about reducing drawdowns using stop loss limits?2017-03-03T15:27:36+00:00
  • Author
    Posts
  • Frank Grossmann
    Moderator
    Post count: 167
    I studied all possible ways of stop loss limits, and the conclusion is, that whatever you do, it lowers the return of the strategies. The simple reason is, that you have a lot of up and down spikes during each month. Most of them can be considered as pure noise. They have nothing to do with the general trend of these ETFs, however they are strong enough to trigger the stop loss limits.
    But once you sold, what can you do then? Reinvest in the ETF which is best at this moment?
    If you do this, most of the time you realize a loss by selling low and you buyback high. This is the typical “sell low, buy high” strategy which many inexperienced investors are using.
    The use of stop loss limits may be good for trading highly volatile single stocks, but it is the worst thing to use, if you trade ETFs which follow longer lasting economic cycles.
    You could probably think, that another way to reduce drawdowns would be, to use shorter ranking periods. You could use for example one week instead of one month, or even do a new ranking every day.
    However again, the problem is, that the shorter these periods become, the more you trade pure noise up and down spikes. If you want to trade the longer economic cycles, you need to use a period length which filters out these noise spikes.
    Using backtests, you can optimize this period length.
    However, you are right, that such intermediate losses are not a nice thing. The only way, I think you can do to reduce such volatility is, by adding some hedge to a strategy. A hedge is normally a negatively correlated asset. The best way to reduce such noise drawdowns is to hold a percentage of the investment in long duration Treasuries.
    At the moment I am working on such rotation strategies which do not switch 100% from one asset to the other, but rather adapt the ratios of stock ETFs and Treasury ETFs to the current market situation.
  • STEPHEN
    Member
    Post count: 22

    Frank,

    I hear you about “it (stop-loss) lowers the return of the strategies.” But with TLT and EDV up so much since Jan 2, it was difficult for me to sit there and watch my profits evaporating on Tuesday (2/3) and Wednesday (2/4), so I exited close to the lows of today (2/4). And right on schedule, they reversed and closed up considerably – presumably on oil and the latest Greece news this late this afternoon.

    At what point does “noise” become a trend change? In 3 days TLT retraced more than 43% after its 20 trading-day rally.

  • Frank Grossmann
    Moderator
    Post count: 167

    There is no problem from time to time to take your own decisions. The current allocation is always only a recommendation. You can well deviate from this recommendation. The strategy for example does not know anything of upcoming Fed meetings and does also not know really what is high and what is a low price of an ETF.

  • jaradgiese
    Participant
    Post count: 2

    The problem with saying treasuries are overvalued is what occurred before this week. They kept rising relentlessly despite the market calling for them to fall nearly all of last year and this year as well. Unfortunately this was not a great week for treasuries, or bonds anywhere in the world for that matter, with PCY down along with the rest of the bond complex. Will they continue to breakdown? Will treasuries rally on some unknown stress next week? Impossible to say. It does hurt when it is your hedge that is crashing down. Is an early summer rate increase now priced into bonds? I don’t know, so maybe better to stick to the plan and exit when enough evidence is there to suggest an allocation shift.

    • STEPHEN
      Member
      Post count: 22

      “exit when enough evidence is there to suggest an allocation shift.”

      I exited ALL my positions on Feb 4 basically because TLT and EDV closed below their EMA21 and retraced substantially since Jan 2 when I bought them. I agree with Frank that “it (stop-loss) lowers the return of the strategies,” but at some point enough is enough, and I think there must be an exit plan. This “crash” was brutal.

      Or when in doubt, get out.

      I read this study of US stocks – “If you place a trailing stop closer than 36% below the current high price, you’ll lose money. The closer the stop is to the current high, the more money you lose (from 36% to 8%, anyway).”

      P.S. Everything I read supports the idea that stop-loss exits hurt performance, so I think Frank is correct.

  • Vangelis
    Moderator
    Post count: 157

    Stephen,

    We encourage everyone to layer their own knowledge, risk profile, judgments on top of our strategies. Forecasting next day or two price moves is very tricky. Thanks for the post.

  • Carlos
    Participant
    Post count: 1

    Hello,

    What about this June post from Gary Antonacci (and 2 papers) regarding that stop loss is good:

    http://www.dualmomentum.net/2015/06/momentum-and-stop-losses.html

    Regards,

    Carlos

  • Alexander Horn
    Keymaster
    Post count: 389

    Hola Carlos,

    thanks for the reference, had not seen Gary’s post before. The difference in guidance lies in the fact that Gary’s strategies use cash as crash protection, e.g. whenever the absolute momentum of the risky asset falls below the treshold, he switches to cash. Adding a stop-loss in this strategies may (!!) make sense, as cash itsself has no return/momentum.

    In the case of our strategies we use long terms bonds as cash protection, under the (proven) assumption, that in the event of any market turmoil long term bonds will receive a major part of the money flow fleeing out of risky markets – therefore showing a positive performance as safe heaven.

    In this case, a stop-loss into cash would be counter productive. We’ve run many tests with any kind of short and long term stop-loss options, and always find inferior performance and wip-saw effects from getting stopped out at the worst possible moment.

    In addition to above, “hard wired” stop-loss orders bear the risk of getting stopped out in a mini-crash as we had several in the recent years. As above 50% of volume is coming from HFT algos in most markets these days, these “beasts” may overreact within seconds, pulling pricing down below your stop-loss level, and then recover within seconds. You would end up loosing money before you even notice..

    Hope this clearifies our position.

    All the best,
    Alex

You must be logged in to reply to this topic.