Systems which have the biggest drawdowns

Discussion in 'Strategy Development' started by bungrider, Dec 6, 2003.

  1. Ladies,

    I am hoping we can start a thread about pretty spectacular systems which have consistent, large, drawdowns.

    I don't know anything about backtesting, so I'm going to guess the specs:

    -consistently profitable except for the drawdowns
    -drawdowns occur in <= 5% of trades made
    -each drawdown costs <= 1% of capital
    -return >15% per year on capital not including losses from drawdowns (obviously if 15% included the drawdowns it would work and nobody would post anything here) - the drawdowns make the account unprofitable

    Preferably it would be a former money maker (like the turtle system, if its legit) that nowadays kinda seems to work but blows up alot.

    GG, you may only post to this thread using one alias. If you post with >1 alias, it's going to start a crusade where I try to get everyone else on ET to pester the mods to boot your 46 aliases off ET.
  2. ...your serious post having received no serious responses, I am obliged to offer a frivolous one: BUY WEDNESDAYs! The drawdown is so big that SELL WEDNESDAYs works great, at least until randomness ends the run. You may laugh your ass off, but before you do, look at BT results for buying/selling specific days. It works until it doesn't, which is about all that you can ask of bitch market.
  3. MR.NBBO


    Really great idea....I wish I had something to contribute, but all my stuff is damn near zero drawdown. But seriously, someone please post some phat drawdown info on a good system!
  4. Any system that is implictly short a straddle will have occasional massive drawdowns admist steady income.

    Trend followers are long straddles, so they have occasional jackpots admist steady bleeds.
  5. LOL

  6. ...speaking philosophically, I am of the opinion that any system with a large drawdown is poorly designed. When I run into a case like that in testing, I attempt to split the system into two systems, one which retains the strengths of the original system, and one which exploits in the opposite direction its weaknesses. Of course you could do that by simply adding a reverse to the original system, but IMO that is inefficient.
  7. i can't disagree with you there...i started this thread because i wanted to think about ways to hedge such a system (which is profitable in >95% of trades and then blows up) and thus make it profitable...

    regardless, your post goes to the heart of this thread, so thanks!
  8. My wife does not trade..

    Michael B.
  9. of the things I do to try to eliminate the blowup problem is to have systems which cascade. This probably isn't possible for the daily time frame, but it can work in the morning. I have an early entry system which is usually right and is bread-and-butter, but watch out when it's wrong. I have four other systems which don't make as much and kick in as volatility increases. Their use is that usually they support the early entry, but when they don't, it's a powerful signal that the early entry was dead wrong.

    Also, I have had some hideously humiliating experiences with systems which quit working spectacularly, or maybe never really "worked" at all, so I use exceedance of the maximum historical drawdown or the maximum number of losing days as signals to cry uncle.
  10. Any volitility breakout system will have big drawdowns. Some systems more so than others depending upon the system's specific parameters.

    A good example is the Turtle System. It's a 100% winner if you can follow the rules. However, the rules are VERY hard to follow for most market participants. One must be sufficiently capitalized for that systems AND one must be able to handle the inevitably huge drawdowns that occur with it.

    This why breakout systems work much better with markets that exhibit a high degree "trendiness" (not a word, I know) like Bonds (and all interest rate markets) and Currencies. And this is why these systems usually suck with Share Index markets like the ES, NQ, DAX etc, etc.

    Dr Zhivodka
    #10     Dec 7, 2003