Scaling or legging out of a winning trade - the empirical evidence using the NEW "Equity Curve" worksheet in the m3 modeler 0 comments
Nov 24, 2009 1:24 PM
There is a great debate in the trading community as to whether one should close out an entire profitable position at a single pre-determined profit objective or, whether to close out part of your position at the same pre-determined point objective booking a portion of the profits, then move your stop loss to break even thereby leaving you in the trade in the event that the move becomes a large or runaway winner. If you stop and think about the dynamics of these 2 options, the only way a trader can participate in a large move is to either consistently trade for large point objectives or, to apply the scaling or "legging" methodology on which the m3 - Money Management Modeler was built upon. For example, if you exclusively trade the SP500 E-mini contract and you use a single point objective when closing out your entire position, then how will you ever participate in a large runaway move? An argument can be made for the mechanical trader whose decisions as to when to buy and sell are generated by a computer algorithm, but if there is a pause in momentum of a particular trade, even the most "patient" system will eventually trigger a sell signal (or buy if you are short). Trading a "patient" system also has its serious drawbacks. Most systems that go for the large winning points have very poor win/loss ratios, meaning most signals to enter a trade result is losses but developers defend this by pointing to the size of the average win when they are correct. Think of a "home run" hitter in baseball - usually they have poor on base percentages but when they connect - they connect. Other systems take the exact opposite approach where they have great win/loss ratios but the average win size is minimal. Brokerage firms love these types of traders because of the number of transactions that ultimately generate commissions fees. The scaling or "legging" approach is the best of both worlds and with sophisticated trading platforms that are readily available today to even the most novice of traders, there has never been a better time to employ this methodology.
In my original m3 platform on the "EquityCurve" worksheet, I allow the user to "check off" which equity curve he or she wishes to see
- you can have as many as 6 equity curves populated at one time - its completely flexible. If you are un- familiar with the tool, as I mentioned above, there are countless blog postings and web tutorials that I have archived for viewing at your convenience. I felt that the cumulative equity curve on legs 2 and 3 needed improving. If a trader has reached their profit target 1, then using this methodology, one would close out a portion of the position and then adjust their stop loss to break even. In my programming, if a trader reaches profit target 1 and then adjusts their stop to break even, the trader is then in a no lose or risk free trade. With a risk free trade, if the position did NOTreach the profit target 2 (and there is no guarantee that it would) then with the stop adjustment in place there is zero risk. When you look at the equity curve you were constantly seeing a 45 degree angle upwards given the way the programming was written. A win ,for example, would be recorded as lets say 3 points and a losing trade would be denoted with a "0" because of the no risk stop adjustment. While this is completely accurate, it is also misleading. given the equity curve was solely based on the wins with the "non-wins" which shows no draw down because of the adjusted stop. It can still be used as long as the user acknowledges that this is a leg 2 ONLY equity curve where there is NOloss reflected in this equity curve. A loss can only be recognized within the 1st leg that is the only time risk is present. Once again - if you have reached your profit target 1 - you adjust your stop to break even removing all risk from the trade while leaving yourself still in the trade looking for additional profit.
I have addressed this challenge with new programming. In the pic below you will see 2 equity curves.
The 1st seen in a burgundy/red colour is the "SingleExit" equity curve and the white is the "RandomEquity" curve. This new and improved Monte Carlo simulation is superior in every metric. The "SingleExit" equity curve is self explanatory. This is for traders who feel that exiting at a single, pre-defined profit objective is the superior way to go. The results speak for themselves. The "RandomEquity" curve now incorporates lossesin conjunctionwith wins and the wins can be any of the 3 possible outcomes - profit target 1, profit target 2 and profit target 3. Not only can the random outcome be any of the 3 winning profit objectives or losses but the winning outcomes have now been weighted to reflect the probability of reaching any of the winning profit targets. So, a loss is still denoted with a -1 but a win could be any of your 3 profit point objectives and most importantly when a win occurs the probability of which leg will be reached reflects the user inputs. Given some of the complexities in this explanation I have attached a 5 minute web tutorial highlighting this new and very important dimension to the m3. This is the empirical evidence proving that to scale out of a winning position incrementally with stop loss adjustments is the superior way to go over a large enough sample size.
One final feature that has been added to this latest platform is one you wont see but the capabilities of this new functionality can not be stressed enough. As before the tapping of the F9 on your keyboard engage the RAND or RANDbetween function which is basically a Monte Carlo simulation. Now, with the "DataTables" function added into the mix, a SINGLE tap of the F9 button automatically runs 1000 iterations of your settings. This saves you the user massive amounts of time knowing that result has been crunched repeatedly to produce the final results. You still have the ability to repeatedly use the F9 key but just remember that each result is as if you tapped the F9 1000 times. Here is the screen shot below where I have highlighted the formula.
Instablogs are blogs which are instantly set up and networked within the Seeking Alpha
community. Instablog posts are not selected, edited or screened by Seeking Alpha editors,
in contrast to contributors' articles.
Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.
Scaling or legging out of a winning trade - the empirical evidence using the NEW "Equity Curve" worksheet in the m3 modeler 0 comments
There is a great debate in the trading community as to whether one should close out an entire profitable position at a single pre-determined profit objective or, whether to close out part of your position at the same pre-determined point objective booking a portion of the profits, then move your stop loss to break even thereby leaving you in the trade in the event that the move becomes a large or runaway winner. If you stop and think about the dynamics of these 2 options, the only way a trader can participate in a large move is to either consistently trade for large point objectives or, to apply the scaling or "legging" methodology on which the m3 - Money Management Modeler was built upon. For example, if you exclusively trade the SP500 E-mini contract and you use a single point objective when closing out your entire position, then how will you ever participate in a large runaway move? An argument can be made for the mechanical trader whose decisions as to when to buy and sell are generated by a computer algorithm, but if there is a pause in momentum of a particular trade, even the most "patient" system will eventually trigger a sell signal (or buy if you are short). Trading a "patient" system also has its serious drawbacks. Most systems that go for the large winning points have very poor win/loss ratios, meaning most signals to enter a trade result is losses but developers defend this by pointing to the size of the average win when they are correct. Think of a "home run" hitter in baseball - usually they have poor on base percentages but when they connect - they connect. Other systems take the exact opposite approach where they have great win/loss ratios but the average win size is minimal. Brokerage firms love these types of traders because of the number of transactions that ultimately generate commissions fees. The scaling or "legging" approach is the best of both worlds and with sophisticated trading platforms that are readily available today to even the most novice of traders, there has never been a better time to employ this methodology.
In my original m3 platform on the "EquityCurve" worksheet, I allow the user to "check off" which equity curve he or she wishes to see
m3 - Equity Curve Check Boxes
http://www.screencast.com/t/NzY0NGUwY2 (Click to enlarge)
- you can have as many as 6 equity curves populated at one time - its completely flexible. If you are un- familiar with the tool, as I mentioned above, there are countless blog postings and web tutorials that I have archived for viewing at your convenience. I felt that the cumulative equity curve on legs 2 and 3 needed improving. If a trader has reached their profit target 1, then using this methodology, one would close out a portion of the position and then adjust their stop loss to break even. In my programming, if a trader reaches profit target 1 and then adjusts their stop to break even, the trader is then in a no lose or risk free trade. With a risk free trade, if the position did NOTreach the profit target 2 (and there is no guarantee that it would) then with the stop adjustment in place there is zero risk. When you look at the equity curve you were constantly seeing a 45 degree angle upwards given the way the programming was written. A win ,for example, would be recorded as lets say 3 points and a losing trade would be denoted with a "0" because of the no risk stop adjustment. While this is completely accurate, it is also misleading. given the equity curve was solely based on the wins with the "non-wins" which shows no draw down because of the adjusted stop. It can still be used as long as the user acknowledges that this is a leg 2 ONLY equity curve where there is NOloss reflected in this equity curve. A loss can only be recognized within the 1st leg that is the only time risk is present. Once again - if you have reached your profit target 1 - you adjust your stop to break even removing all risk from the trade while leaving yourself still in the trade looking for additional profit.
I have addressed this challenge with new programming. In the pic below you will see 2 equity curves.
Random vs. Single Exit
http://www.screencast.com/t/YjFkYzg0OTgt (Click to enlarge)
The 1st seen in a burgundy/red colour is the "SingleExit" equity curve and the white is the "RandomEquity" curve. This new and improved Monte Carlo simulation is superior in every metric. The "SingleExit" equity curve is self explanatory. This is for traders who feel that exiting at a single, pre-defined profit objective is the superior way to go. The results speak for themselves. The "RandomEquity" curve now incorporates lossesin conjunctionwith wins and the wins can be any of the 3 possible outcomes - profit target 1, profit target 2 and profit target 3. Not only can the random outcome be any of the 3 winning profit objectives or losses but the winning outcomes have now been weighted to reflect the probability of reaching any of the winning profit targets. So, a loss is still denoted with a -1 but a win could be any of your 3 profit point objectives and most importantly when a win occurs the probability of which leg will be reached reflects the user inputs. Given some of the complexities in this explanation I have attached a 5 minute web tutorial highlighting this new and very important dimension to the m3. This is the empirical evidence proving that to scale out of a winning position incrementally with stop loss adjustments is the superior way to go over a large enough sample size.
One final feature that has been added to this latest platform is one you wont see but the capabilities of this new functionality can not be stressed enough. As before the tapping of the F9 on your keyboard engage the RAND or RANDbetween function which is basically a Monte Carlo simulation. Now, with the "DataTables" function added into the mix, a SINGLE tap of the F9 button automatically runs 1000 iterations of your settings. This saves you the user massive amounts of time knowing that result has been crunched repeatedly to produce the final results. You still have the ability to repeatedly use the F9 key but just remember that each result is as if you tapped the F9 1000 times. Here is the screen shot below where I have highlighted the formula.
Data Tables
http://www.screencast.com/t/NGZjMWUwN2E (Click to enlarge)
Enjoy the web tutorial and hopefully this will change the way you approach your trading endeavours going forward.
http://www.screencast.com/t/YmQ2NWIxMGIt (Click to play)
Fulcrum Shift Trading
Instablogs are blogs which are instantly set up and networked within the Seeking Alpha community. Instablog posts are not selected, edited or screened by Seeking Alpha editors, in contrast to contributors' articles.
Share this Instablog
Latest Followers
Posts by Themes