"Scaling out" is inferior behavior

Discussion in 'Strategy Building' started by Buy1Sell2, Oct 18, 2006.

Do you scale out of positions?

  1. I always scale out

    113 vote(s)
    14.1%
  2. I scale out most of the time

    228 vote(s)
    28.5%
  3. Most of the time, I do not scale out

    189 vote(s)
    23.6%
  4. I never scale out

    270 vote(s)
    33.8%
  1. romik

    romik

    I wasn't saying that, crossed wires again.
     
    #451     Oct 27, 2006
  2. Buy1Sell2

    Buy1Sell2

    ok
     
    #452     Oct 27, 2006
  3. romik

    romik

    EDITED "all out" VERSION

    50 % winning percentage 4 ES Contracts 20 trades 6 pt target 2 pt loss (3:1)

    1st example without scaling out

    10 winners 6X(4 Contracts) = $240 pts ($12000)
    10 losers 2X(4Contracts) = $80 pts (-$4000)

    Net profit +$8000 before commissions

    50 % winning percentage 4 ES Contracts 20 trades 2 pt target 2 pt loss

    2nd example with scaling out half at 2 pt

    5 winners 2X(4 Contracts) =40 pts ($2000)
    2 breakeven X(4 Contracts) =0($0)
    1 winner 3X(4 Contracts)=12pts ($600)
    1 winner 4X(4 Contracts)=16pts ($800)
    1 winner 10X(4 Contracts)=40pts ($2000)
    10 losers 2X(4 Contracts) = -80 pts (-$4000)

    Net profit before commissions=+$1400
     
    #453     Oct 27, 2006
  4. Buy1Sell2

    Buy1Sell2

    This isn't quite what I was looking for.
     
    #454     Oct 27, 2006
  5. romik

    romik

    EDITED "all out" and "scaled out" VERSION

    50 % winning percentage 4 ES Contracts 20 trades 6 pt target 2 pt loss (3:1)

    1st example without scaling out

    10 winners 6X(4 Contracts) = $240 pts ($12000)
    10 losers 2X(4Contracts) = $80 pts (-$4000)

    Net profit +$8000 before commissions

    50 % winning percentage 4 ES Contracts 20 trades 6 pt target 2 pt loss (3:1)

    2nd example with scaling out half at 6 pt

    5 winners 6X(4 Contracts) =120 pts ($6000)
    2 breakeven X(4 Contracts) =0($0)
    1 winner 7X(4 Contracts)=28pts ($1400)
    1 winner 8X(4 Contracts)=32pts ($1600)
    1 winner 10X(4 Contracts)=40pts ($2000)
    10 losers 2X(4 Contracts) = -80 pts (-$4000)

    Net profit before commissions=+$7000

    In this example a reward/risk is identical 3:1, BUT (and that BUT is pretty important here IMHO) what happens in the scaled out example is a constantly changing average for the remaining 50% position, sometimes the end result will beat "all out", sometimes it will equal and sometimes it will be less. Inferior or not is determined by the average intraday range, which affects both strategies.
     
    #455     Oct 27, 2006
  6. Buy1Sell2

    Buy1Sell2

    But not in the long haul over time.
     
    #456     Oct 27, 2006
  7. What an epic struggle this thread has turned out to be.

    I just dont know whether to scale out of it or just close out.
    Maybe I will look back in 6 months time.
     
    #457     Oct 27, 2006
  8. romik

    romik

    I think the ones that do not want to read discussions here can easily unsubscribe from the thread. There is an easy decision for you. So far it's been an interesting read-up.
     
    #458     Oct 27, 2006
  9. That is a thought romik, but why dont I just jump in for what it is worth.

    I used to scale out half my position and bring my stop up into profit.
    It proved to be slightly more profitable, bearing in mind my average trade was 6 points, but the gain ( 1 point or 20%) was not worth the effort by comparison to lot size and so I now just trade to a set target.
     
    #459     Oct 27, 2006
  10. In very short, here is how I think of it.

    Out of any given month, 18 trades:

    a) 6 will get stopped out for -2 pts (or less actually, if you use a 4pt stop, but for the sake of the analogy, we'll use -2pts).
    Total: 6x(2)=(12) pts

    b) 6 will go for a range of 3 to 6 pts, (after which the market will begin to turn, so you will lose 2pts before you exit).
    Total: 6x1 to 4=6 to 24 pts (average of +15pts)

    c) 6 will go for a range of anywhere from 8 to 18 pts (very doable on reversal with the dominant trend / range / and extended range days).
    Total: 6x6 to 18=36 to 108 pts (average of +72pts net). I would in addition say that you would have a hi-lo range of probably 60 to 84 pts for option C.

    d) at least two days will have no range and just putter around, count them as scratch ... so while there may be 20 tradeable days in a given month, I'm only using 18 for the exercise.

    So for any given month, using a methodology which holds the trend while looking for an extended range day (which is where this methodology has its knock-out punch) will have the following results:

    a) netted -12 pts
    b) netted +15 pts (effectively nullifying your losses)
    c) netted a rang of 60 to 84 pts (with an average of 72 pts)
    ***
    ... and that's just for the ES, now apply the principles to other markets and instead of just staying with the intra-day method extended it out to holding overnight positions... and you can see how a technique which holds for the longer term has an automatic positive expectancy, and is most likely the optimal method for generating real wealth via trading over the long term.

    Regards All,

    Jimmy

    p.s. made edits to make the math (logic) clearer.
     
    #460     Oct 27, 2006