What does PF=1.075 tell us?

Discussion in 'Strategy Development' started by trd, Aug 4, 2009.

  1. trd

    trd

    Hi,
    what does a Profit Factor of 1.075 for a system tell us?
    Can one calculate from that number the annual result of a trading account?
    How important is PF ?
    What is/are the most important of such numbers for a system?
     
  2. auspiv

    auspiv

    A profit factor of 1.075 means you should start looking at creating a new system...
     
  3. pf=1.075 says that the sum of money gained divided by the sum of money given to the market is very close to 1.

    You should be concerned. Is this an indication of future performance?

    Typically people look for system pfs of 2 or more although you may go down to 1.3-1.5 in special circumstances.
     
  4. trd

    trd

    But it means for example GrossWin=$107,500 and GrossLoss=$100,000.
    Ie. making $7,500 gain, that's a profit of 7.5%.
    Why is making a profit of 7.5% bad?
     
  5. Retief

    Retief

    Because there is not much of a safety factor. The slightest slippage, bad fills, and mistakes will wipe that profit out and probably result in a loss. Paper trade your system for a month. I bet it won't be profitable with a profit factor that low.
     
  6. trd

    trd

    I'm using the averages of simulated 30 years.
    Here are the other results of this system:
    Annual PL% = 42.1699 (StdDev=6.3530, MinObserved=24.8692, MaxObserved=59.9619)
    MDD% = -1.5251 (StdDev=0.3299, MinObserved=-2.2053, MaxObserved=-0.7925)
    MaxLoss%_of_InitialAccountValue = -0.4438 (StdDev=0.3968, MinObserved=-1.8244, MaxObserved=0.0000)
     
  7. Retief

    Retief

    Yes, but what are your assumptions? If you're assuming that every buy is at the high of the bar and every sell is at the low, and stops get filled with ten points slippage, you might possibly have a profitable system.
     
  8. trd

    trd

    I've used limit orders and used 5 second bars.
    My experiments show that Profit Factor is IMO very misleading
    as it highly depends on the number of intervalls per day (ie. on the bar size):
    The bigger bar size the more is the PF, and vice versa...
    Yes, I can achieve even a PF > 2.5 when I simply make the bar size bigger,
    keeping everything else the same... :)
    My conclusion: PF belongs into the trashcan! :)
     
  9. Retief

    Retief

    This is based on actual trading or solely on your vast experience with back testing?

    Probably the reason profit factor increases with bar size is because the size of profits and losses start to completely dwarf commissions and slippage with larger bars. At smaller bar sizes, commission and slippage has a much greater effect. This is also probably why the vast majority of day traders are losers - because they're dealing with bar sizes that are too small and have very small profit factors.

    Just a theory.
     
  10. It is fairly worthless, not quite trash though.
     
    #10     Aug 4, 2009