what the hell happened!?!?

Discussion in 'Strategy Development' started by feng456, Dec 23, 2011.

  1. feng456


    I've been doing an automated strategy that has been profitable for 2 years and all of a sudden this month it has been catastrophic in losses. I did 4 years of backtesting with nothing that suggests this would happen. What the f*ck do I even do now? My account is destroyed. I'm not sure if I should just quit or attempt a new approach.

    I followed my strategy to the letter and the backtesting was done in a precise and logical manner. How does a system all of a sudden do that?
  2. Bob111


  3. dom993


    It happened to me a few times, too.

    To give myself some objectivity in assessing whether a drawdown is "reasonable" or catastrophic, I run MonteCarlo simulations using the trade distribution out of backtesting, take for "MADD" (Maximum Allowable DrawDown) the mean + N*std-dev of the MonteCarlo sim's max drawdown. I usually run the MonteCarlo sim for a number of trades corresponding to 1-year worth of trading, and typically use N=5 to define the MADD. If a system's drawdown gets to the MADD, I consider it dead and stop trading it.

    If you wish so, provide the backtesting trade distribution, an approximate number of trades for 1-year worth of trading, and I will run the MonteCarlo sim for you.
  4. You did only 4 years of backtesting, but maybe you should've done 14? Yep thats right, strategies need to work on the ancient data as well, and good ones do. On the other hand, some of the top systems have very bad months where they look broken, and one month later they start working again. This december i've seen more automated strats (futures) fail than other months historically. Not mine, but other people's. Maybe the market is really weird for some of the bots right now. Who knows. There's many ways to judge the robustness though. You can look at where the trades are falling within the distribution of the backtest and forward test and if the mean center of gravity starts going weird, or stddev is way off, it's a different problem and a warning.

    Switch it off and paper trade it for 20 days, see what it does.
  5. You silly thing OP. You brazenly assumed that past performance was a guarantee of future results. Not to mention you had no Plan B for what happens if your system stopped working.

    I like how the narrator yells "what the fuck" in English.
  6. It takes a whole lot of work to get to the bottom of what makes price move. You are omitting something from your thinking that is obviously an essential component vector of price moves...

    I, and probably about anybody that designs systems, have seen strategies that were quite good at times but were "fragile". I had one based on volume bars that could backtest great on a days worth of data but if I changed the start time of the price data it would not work well at all!! That is very fragile!! Obviously it was not a strategy that was worth anything but it demonstrated very clearly to me just how much success can be simply built in by accident.
  7. wrbtrader


    Although I myself do not use automation trading nor mechanical system (I'm discretionary trader)...I do have a handful of personal friends that are automation traders. Most of them have said their systems do not perform well (performs poorly) in December most of the past 10 years.

    Further, most of them have stated they do not trade at all in December because of what their backtest have shown them. Simply, December is one big vacation month away from the markets for them.
  8. western


    Rather than be upset, you should be extremely thankful your system worked as long as it did.

    The market is constantly evolving, your strategies need to as well.
  9. Bob111


    pretty good suggestion and yes, many times this does helps. just take a vacation OR go back to paper trading.to see,how it goes.
    or reduce the size by a half or so.
  10. feng456


    I have reduced size because it's gotten to the point where if I don't I will get a margin call. I've been through bad drawdowns before but nothing even close to this. I am currently down 35 points and normally I am down at most 25-30 (which happened once in all my data). Historically I have never been down this much in all the backtesting I've done.

    Also historically speaking, December has not been a weak month. In fact, I do not have weak months that repeat annually. It's fairly random.

    I am very disciplined as a result of learning from my past mistakes about the alternative but it is really hard to follow your rules when you lose almost every single trade...when you 80-20 win rate turns upside down and suddenly u lose 80 and win only 20.

    I designed my leverage under the assumption that the worst drawdown in the backtested 4 years (prior to now) was the worst it could get so now since it's much worse, I am getting wiped out.
    #10     Dec 23, 2011