To SCALE or Not to SCALE a profit?

[daniel_m]

Originally posted by profitseer
jcom, just to set the previous misunderstanding straight, I am not against pyramiding. All that I have to show for myself is the result of a pyramid. I am against portraying both the scale in and scale out as some sort of mm scheme which can compensate for a trading system that can't survive on an all in or all out basis.

But just to mention one more thing more politely this time, if you scale in, or out, one must also consider the unused capital sitting on the side when the full load is not on, and then fairly compare that to an all in all out system where the full capital is being used the whole trade.

requiring a trading system to survive on an all in or all out basis misses the point of the impact of position size. the amount of size at any point in a trade is an integral part of the profitability of any trading style/system (whether or not this is realised), not something to be considered separately. Various position sizing algorithms do return different profit levels for the same price movements of the underlying subsequant to a position being taken.

i'm not sure exactly what kind of consideration you have in mind for the unused capital. even when using an all in/all out system are you using your entire trading account every single time? most people aren't, so there isn't much to compare there. the real basis for comparison is the impact of the trades on your equity, not the amount of equity you are using (in and of itself).

at this point i'm still somewhat mired in an "if it ain't broke dont' fix it" frame of mind with the way i trade (daytrading, not quite scalping), so i'm not actually employing any creative position sizing strategies. between trading, engaging in useless debates in moronic threads like ZSG/religion/politics on ET, relocating my office, and trying to lead some semblance of a "normal" life, i haven't found too much time to further develop my ideas beyond some simple mathematical calculations, but i fully intend to go down the path of being right BIG and being wrong small.

 

[bone]

Vlad, two words: Victor Niederhoffer.

You are a ticking bomb. May not happen this week or even this year, but you will get caught. This system has an unreconcilable risk parameter unless you can explain your stop-loss system in greater detail. Hey, if it works, fine. I've been wrong plenty of times in my life. I'm usually wrong when it comes to predicting where the market is going on any timeframe further out than one hour.

Down on the floor, we used to call this "cannonballing".
Nobody wanted to clear them because of their wild equity swings.

Why not just sit down at the open, enter a tier of bids below the market and a tier of offers above the market, and go home? Does this have a positive expectancy after you get caught in that nasty trend day? It might because the market will probably range trade for two or three days before a trend day appears. What if the market trends for two or three days in a row? What happens when another September 11 or Black Monday develops - do I scale out of my loser or buy into it?

 

[vladiator]

Originally posted by nitro

Vlad,

This is a bizarre "system." I understand what you are saying - somehow it feels like there is a fundamental mathematical error in your analysis...However, given your credentials, it hardly seems likely...

Let me think about it....

BTW, is this an NYSE or NASDAQ or both system?


nitro

I guess you can call it bizarre to the extent that it's not pure TA based and may seem odd to TA users. But I wouldn't call it fundamental either. It's based on a mixture of both and the entry exit points are pretty much always a the same time (as the principles it's based on are most pronounced at those times.) That's why averaging in/out works the way it does. As for the error, I'll be glad to see your input Basically what I'm trying to say is that if you have two distributions of daily outcomes, say they are Normal, just for the sake of example (don't have to be). one has a sigma of 20 times larger then the other. If you now create a distribution of cumulative annual returns based on those daily observations, you will see how exaggerated the extremes will be in the case of the more leptokurtic one. By averaging in/out I stay closer to the mean in the daily distribution and reduce the likelihood of getting wiped out in the cumulated one.
Thanks for the feedback
V.

 

[J_Commisso]

Jem thankyou for the reply

You raise some good points and like I said before -- if the shit makes more cents in the bigger picture then obviously it makes sense, but I suspect many do it for the opposite reason. It makes sense to them in some sort of rational way, but does not make more cents in the bigger picture.

Over the years she has embedded this one simple truth into my brain -- not everything that makes sense -- will make cents, but what makes cents-----> makes sense!.

Many try in vain to bend the spoon, but they really should be bending themselves...

Oh and as for the "smoothing" the equity-curve and decreasing the size and length of drawdowns -- IT IS AN ILLUSION! I am not a mathematician, but from my direct experience with the game it will be the method producing the highest expectancy not highest win % that will give the "smoothest" curve. The illusion is set forth because they are looking at the extreme micro-scopic levels of the curve -- they just need to stretch it out slightly and see the bigger picture...

PEACE and good-trading Jem, (and Ienoyed reading what you had to say)
Commisso