Example Trading System

Discussion in 'Strategy Building' started by sidinuk, Jan 17, 2005.

  1. the system does not trade on Thursdays. Its backtesting showed that this was the most unprofitable day of the week.

    Dennis
     
    #111     Feb 11, 2005
  2. What's different about Thursdays vs. any other day of the week? I think this is curve fitting plain and simple.
     
    #112     Feb 11, 2005
  3. If you take the trouble to read the original paper, you'll see that that's what the author also suspects.
     
    #113     Feb 11, 2005
  4. mogul

    mogul

    curvefitting or not, who cares? people get so bent out of shape when it comes to that word

    the only problem with excluding thursdays is if thursday's become one of the main money-making days.

    the core strategy is profitable with or without thursdays, so if excluding thursdays gives it a boost, exclude them until you see reason not too!
     
    #114     Feb 11, 2005
  5. My testing shows that the Thursday factor statistically is significant at the 90% confidence limit, so I think it is not curve fitting.

    DS
     
    #115     Feb 11, 2005
  6. Anyone,


    Can you speculate as to why Thursdays fundamentally do not work with this system?

    Michael B.

    P.S. The reason I ask, are there conditions such as highs and lows before Thursdays that could be significant to help filter whether to take the trade or not, also ocher days of the week? There could be a additional probability study as to the strength of the signal based on previous days range....distance of high or low from the open on the next day...etc.....etc....... ?



     
    #116     Feb 11, 2005
  7. I believe this is a sound way to do, that is, as you say, until it stops working. But I would use both systems, one without the Thursdays excluded and the other one as in the original work.

    The main problem I see with this system as applied to YM right now is not the Thursday problem. The problem is that the daily range in YM has recently shrunk considerably even compared to ES so this system may not work as well as it did in the past until YM regains its previous levels of volatility.
     
    #117     Feb 11, 2005
  8. Electron,

    How about the Russel?

    The figures looked Robust at Sidinuk's website...But is ER also less volatile as the YM is?

    Michael B.


     
    #118     Feb 11, 2005
  9. I follow mostly YM and ES these days, can't say what ER2 dynamics is these days. Was great in the past, perhaps it still is.
     
    #119     Feb 11, 2005
  10. dbs119

    dbs119

    ElectricSavant,

    Fundamentally the market might go sideways on thursday,simply because we almost always have some nivce reports kicking in at 8:30 AM on friday, so nobody wants to place a bet?
    Anyone else have any other suggestion :))
     
    #120     Feb 11, 2005