"Scaling out" is inferior behavior
- Thread starter Buy1Sell2
- Start date
All well and good, but some people don't trade like that. Let take Mr. Quant as a hypothetical example: He has built a model that assesses the expectancy of various indices on a daily basis. Two trading days ago, to his great delight, the model assessed the S&P 500 to have a positive expectancy of +1.44% for the coming week, with an expected winning percentage of 68.7%, for the coming week. So, he placed a leveraged bullish bet with 200% of the funds allocated to that market. Today, toward the end of trading, his calculated gain of 3.4% on his position gave him a warm happy feeling inside. Checking the model, he saw an expected winning percentage of 54.1% and a one week positive expectancy of +0.37%.
What should Mr. Quant do?
In order to receive any partial credit, you must show your work.
I could take some time to break this down into points etc. however, there is information missing: How much loss are we willing to take on each trade? What is the winning percentage for where we stand now, which is at 1.7% positive expectancy? Once we know that, then your exercise will only help us determine which of the three targets that we should shoot for (we have already exceeded the 1.44% positive expectancy, so I assume that Mr Quant has done some homework and found that 1.7% is a better target).
In any event , your exercise is geared toward finding an optimal target only and has nothing to do with my premise which is this: Once a trade is initiated, then the full position should be allowed to run to the target and it is inferior over the long haul to scale out before the trade runs to maturity. --Most of my trading is with trailing stops, so trade maturity is when my trailing stop is taken out. ie I rarely use targets although I wouild tell you that very short term traders can easily identify targets.
--I appreciate your efforts here, but it has fallen well short of the mark. --
Which would you rather be ? A 75% trader who hits 2-4 points consistently multiple times intraday? Or a 30% swing trader who nets 20-30 points but must wait long periods for each trade to mature ? If you have a profit target there is nothing wrong with waiting for it to get hit but it really depends on your current win rate. I have noticed in your ES trades you often take quite a bit of heat, such as the past one that went 40 points against you before reversing. It is obvious that this ES trading makes up such a small % of your portfolio that it allows you to use such wide stops.
There are trends ... there aren't
Scaling in works ... it doesn't
Scaling out works ... it doesn't
We predict .. we don't
Is 5 months and 120 pages on one of these arguments evidence of obsession or boredom?
------------------------
The things people believe in are usually just what they instinctively feel is right; the justifications and arguments are the least important part of the belief.
That's why you can win the argument, prove them wrong, and still they believe what they did in the first place. You've attacked the wrong thing.
So what do you do? Agree to disagree. Or fight. - C. Zakalwe.
Scaling in works ... it doesn't
Scaling out works ... it doesn't
We predict .. we don't
Is 5 months and 120 pages on one of these arguments evidence of obsession or boredom?
------------------------
The things people believe in are usually just what they instinctively feel is right; the justifications and arguments are the least important part of the belief.
That's why you can win the argument, prove them wrong, and still they believe what they did in the first place. You've attacked the wrong thing.
So what do you do? Agree to disagree. Or fight. - C. Zakalwe.
I have just tuned into this board and think its really good. I am intrigued by this debate.
I think b1s1 has a point...but it also depends on how your system is quantified.
For example...I feel that my exit rules get me out quite near the top of a trend.
So if someone has an exit strategy that gets you out usually near the top of the trend then it makes sense to not scale out.
On the other hand if someone has an exit plan that gets them out half way up the trend then i think its time to use scaling out and leave some contracts on the table and then relax the exit rules.
however I would stress that this needs to be quantified.
for example in backtesting I make a note of all my entry price and exit price according to my systems rules. I then measure the percentage increase or decrease in price between the entry and exit. I do this for a security over as many years as possible, some currencies I will do this for nearly 20 years of data.
Now if I add up all these percentages between my entry and exit and take an average then I know what the average move is between my entry and exit.
i then create a frequency table in excel.
1-3% 13
4-6% 14
7-10% 13
10-15% 10
15-20% 9
20-30% 7
This only an example. I also mark the peak of the trend or low point in a down trend and I calculate the percentage between my exit and the actual peak of the trend. So if I have an exit strategy that 4 times out of 10 gets me out within 10% of the peak then it doesnt make sense to scale out. It all depends on how good your exit strategy.
I think another poster said that you will know the answer to scale out ir not with study of your system.
if you quantify using stats about your system exits within 30% or 40% from the peak of the trend then it would be a good idea to look at your exit strategy and try to change it to catch more of the move.
As I said it all depends. mine is still a work in progress but if my system gets me out 1 out of 10 times within 10% of the peak then it would be wise to scale out some at an earlier point.
However my system gets me out with 10% of the peak of the trend 3 or 4 times out of 10 then its better not to scale out.
Scaling in...? Hmm,still not sure. again I think it depends on your system.
Looking at your trade success rate should offer some clues...
if you have a success rate of less than 50% winners to losers then scaling in would bea good thing. If your success is more than 50% but you only catch small parts of the move then I think its better to get fully in early.
If your system catches agreat portion of the move but has a rate of success less than 50 percent then I think its better to scale in.
Foe example winning only 4 out of 10 times means that you have 6 losers.
Assume we risk 2%in total in any one trade
Lets say that we go all the way in from the beginning. That means we lose 12% of our bank at some point in those 10 trades.
However if we scale in starting at 0.5% and the trade turns into a loser we have only lost 0.5%. This could be a saving of say
6*1.5%=9%
A 9% saving of our bank in every 10 trades.
But what if these other 4 times we win that we get 80 to 90% of the move?
Or on the other hand we get only 30-70% of the move?
these things have to weighted and tested.
is our exit strategy off setting the 9% saving we made by scaling in?Not really if we dont catch a good portion of the move...
I think if our exit gets us out at a high point then it is better to scale in because the part of the move we lose at the start will be made up with a good exit. My heads starting to fry must go...
I think b1s1 has a point...but it also depends on how your system is quantified.
For example...I feel that my exit rules get me out quite near the top of a trend.
So if someone has an exit strategy that gets you out usually near the top of the trend then it makes sense to not scale out.
On the other hand if someone has an exit plan that gets them out half way up the trend then i think its time to use scaling out and leave some contracts on the table and then relax the exit rules.
however I would stress that this needs to be quantified.
for example in backtesting I make a note of all my entry price and exit price according to my systems rules. I then measure the percentage increase or decrease in price between the entry and exit. I do this for a security over as many years as possible, some currencies I will do this for nearly 20 years of data.
Now if I add up all these percentages between my entry and exit and take an average then I know what the average move is between my entry and exit.
i then create a frequency table in excel.
1-3% 13
4-6% 14
7-10% 13
10-15% 10
15-20% 9
20-30% 7
This only an example. I also mark the peak of the trend or low point in a down trend and I calculate the percentage between my exit and the actual peak of the trend. So if I have an exit strategy that 4 times out of 10 gets me out within 10% of the peak then it doesnt make sense to scale out. It all depends on how good your exit strategy.
I think another poster said that you will know the answer to scale out ir not with study of your system.
if you quantify using stats about your system exits within 30% or 40% from the peak of the trend then it would be a good idea to look at your exit strategy and try to change it to catch more of the move.
As I said it all depends. mine is still a work in progress but if my system gets me out 1 out of 10 times within 10% of the peak then it would be wise to scale out some at an earlier point.
However my system gets me out with 10% of the peak of the trend 3 or 4 times out of 10 then its better not to scale out.
Scaling in...? Hmm,still not sure. again I think it depends on your system.
Looking at your trade success rate should offer some clues...
if you have a success rate of less than 50% winners to losers then scaling in would bea good thing. If your success is more than 50% but you only catch small parts of the move then I think its better to get fully in early.
If your system catches agreat portion of the move but has a rate of success less than 50 percent then I think its better to scale in.
Foe example winning only 4 out of 10 times means that you have 6 losers.
Assume we risk 2%in total in any one trade
Lets say that we go all the way in from the beginning. That means we lose 12% of our bank at some point in those 10 trades.
However if we scale in starting at 0.5% and the trade turns into a loser we have only lost 0.5%. This could be a saving of say
6*1.5%=9%
A 9% saving of our bank in every 10 trades.
But what if these other 4 times we win that we get 80 to 90% of the move?
Or on the other hand we get only 30-70% of the move?
these things have to weighted and tested.
is our exit strategy off setting the 9% saving we made by scaling in?Not really if we dont catch a good portion of the move...
I think if our exit gets us out at a high point then it is better to scale in because the part of the move we lose at the start will be made up with a good exit. My heads starting to fry must go...
Er, no. You should probably try reading the "Mr. Quant" post again. There is no such thing as a "target" in his trading methodology. In fact, he had achieved the original expectancy (and then some) and still had a forward-looking positive expectancy remaining.
Please put your thinking cap on this time...
I think so too. I usually sell all at once (to save money on commission costs), but I'll buy into a position over time, because of all the things Cramer has said, I agree with him on that issue of "dont buy all at once". And on the issue of scaling out....why's it bad to take some money off the table?.... I'm almost positive position traders do this to have some kind of weekly/ monthly income?
cm
Quote from version77:
Since the word "inferior" is used in the thread title, I assume that
someone with a superiority complex wrote it and is trying his best
to remain superior with his hypothesis...
You have pretty much nailed it, I am afraid.
If and when the author of this thread ever begins trading with real money, this thread will cease to exist.
In reality, this thread is an accurate representation of ET in general, in that there are some very good comments lost amongst a sea of drivel.