why scale out

BlindLemonBoosh · Dec 5, 2008

Quote from Thunderdog:

As for "averaging," which is adding to a position that is going against you, that's like driving faster when you're lost. If the market is telling you that your timing is off, then this is hardly the time to increase your exposure.

You're arguing against the premise of Regression To the Mean. I'd say this is a losing argument over the long haul. However, RTM was a very dangerous game to play the past couple of months. I'd be willing to bet that history will show quite a few RTM traders blowing up in the 8/08-11/08 period... trampled under the hooves of the trend followers.

BlindLemonBoosh · Dec 5, 2008

Quote from MGJ:

When I apply "scaling out" techniques to a wide variety of fully mechanical trading systems, my backtesting results generally agree with the above conclusion expressed by member stevegee58: scaling out typically reduces the volatility ("jaggedy-ness") of the equity curve, and increases the various gain-to-pain ratios that measure system desirability. Sharpe Ratio, Lake Ratio, MAR Ratio, Kestner K-Ratio, Ulcer Index, Sortino Ratio, Return Retracement Ratio, are all generally better when scaling out is enabled than when scaling out is disabled.

http://www.elitetrader.com/vb/showthread.php?s=&postid=1640765&highlight=scaling#post1640765

Great post. One that will surely be ignored by those hypnotized by the false idol 'Absolute Return'.

fluttrader · Dec 5, 2008

Quote from MGJ:

When I apply "scaling out" techniques to a wide variety of fully mechanical trading systems, my backtesting results generally agree with the above conclusion expressed by member stevegee58: scaling out typically reduces the volatility ("jaggedy-ness") of the equity curve, and increases the various gain-to-pain ratios that measure system desirability. Sharpe Ratio, Lake Ratio, MAR Ratio, Kestner K-Ratio, Ulcer Index, Sortino Ratio, Return Retracement Ratio, are all generally better when scaling out is enabled than when scaling out is disabled.

In my own case, it has worked reasonably well in real life (actual trading) too.

Additionally, my experience is that the selfsame analysis of "scaling out" can be applied to "scaling in" as well, producing similar test results and similar real trade experience: scaling in appears to improve gain-to-pain measures of equity curves, both in backtesting and in real life trading.

(FWIW, my testing and trading employ mech systems applied to portfolios of >50 futures markets. The systems often have >25 simultaneous positions in different instruments.)

http://www.elitetrader.com/vb/showthread.php?s=&postid=1640765&highlight=scaling#post1640765

Great post.

Thunderdog · Dec 5, 2008

Quote from BlindLemonBoosh:

You're arguing against the premise of Regression To the Mean. I'd say this is a losing argument over the long haul...

Most really big mistakes started out as small mistakes that were not corrected in time.

operator · Dec 5, 2008

i'm either all in or all out. scaling in or out is to complicated for me...

NoDoji · Dec 5, 2008

Quote from Thunderdog:

Most really big mistakes started out as small mistakes that were not corrected in time.

That is the absolute truth and the market pounded that lesson into my head the hard way. I am far more careful now and adhere to my mental stops diligently. I realized that most of my profitable trades go in my favor immediately and I can set a stop to b/e or better right away. I had many trades where I averaged in when they went against me, but only in extremely overbought/oversold conditions (such as at intraday highs/lows AND 52-week highs/lows) and they were all profitable. However, I don't even do that any more because it's simply too risky IMHO. I have a different philosophy now: If the trade moves against me and I stop out with a very limited loss, I can re-enter the trade at a better point and this tactic limits my risk immensely.

BlindLemonBoosh · Dec 6, 2008

Quote from Thunderdog:

Most really big mistakes started out as small mistakes that were not corrected in time.

Agreed. Please note that I did NOT suggest forgoing 'risk management'.

rock34748 · Dec 6, 2008

Quote from stevegee58:

Hmmm. "Scaling in" is definitely something I don't do, at least based on the normal definition. I see it as adding to a losing position.

Now, I do like reverse scaling in. i.e. adding to *winning* positions.

I agree with this 100%. Although, for a few specific 5 to 10 day swing trades in the last 3 months I must admit I purposely violated this rule. Went back to the old add more to the loser waiting for the bear rally. Worked well in most cases and overall, only because research continued to point the right time and direction.

Looking at each individually would be scary from a psych point of view so not to be done often and not to be done on HOPE. When I listen to that "hope feeling" the market does what it does, which translates to beating my "hope" to a pulp.

good one · Dec 6, 2008

Quote from stevegee58:

Hmmm. "Scaling in" is definitely something I don't do, at least based on the normal definition. I see it as adding to a losing position.

Now, I do like reverse scaling in. i.e. adding to *winning* positions.

I scale in and out.
Acrary has some good post in the archives if you dig for them with ideas about scaling in.
Ultimately its just a way to have smaller losses and more capital on with big wins and longer runs on winners.
The only way its "inferiour behaviour" is if you have a crystal ball and can buy the low and sell the high of a move..

fluttrader · Dec 6, 2008

Quote from jdeezero05:

I scale in and out.
Acrary has some good post in the archives if you dig for them with ideas about scaling in.
Ultimately its just a way to have smaller losses and more capital on with big wins and longer runs on winners.
The only way its "inferiour behaviour" is if you have a crystal ball and can buy the low and sell the high of a move..

How much difference in results does it make in your trading ?