|
Discussion of April Newsletter in Detail, Planning the system, Setting up Buys/Sells/Profit Target, Exits Signals etc |
May 2 , 2007
16:30 Ensign Howard: Welcome to today's training class with Ensign.... this is Howard and I will be the trainer.
16:31 Ensign Howard: This coming weekend we will be upgrading equipment in our server data center. Downtime should be minimal, and chat off-line for less than a minute. We have the new equipment installed, cabled, and just waiting to be put on-line this Saturday. The FXCM feed continues to run very well, so we think that project is finished and stable. I see we had 178 FXCM connections today.
16:34 Ensign Howard: Today's training topic will be the material published a few days ago in the April Trading Tips newsletter on designing trading systems.
16:35 Ensign Howard: Before we dive into that subject, I will bring to your attention that the 05-02 release today adds a few useful DYO | Global Flag statements to make it easy to test if a bar's price is near a value such as a support and resistance line.
http://www.ensignsupport.com/email/1612.mhtml
16:36 Ensign Howard: This trivial example shows Line A parking a price value in GV[1], and then Line B is the new statement that tests if the bar's price is near this GV value plus or minus 1.0 points.
16:37 Ensign Howard: Use the # field to set your +/- level of tolerance.
16:39 Ensign Howard: Now on a tool like Support and Resistance, each line can have its value stored directly in a GV using the tool property form. Then you can add a DYO like this one to do a test if the bar's close price is near the line of interest and put on markers or be used in an Alert.
Planning the System
16:40 Ensign Howard: The April Trading Tips newsletter has 2 articles about designing trading systems.
http://www.ensignsoftware.com/tips/tradingtips88.htm
16:41 Ensign Howard: Occasionally I get asked about back testing, so I decided to write up an example and walk the user through the thought process to create a system from scratch. I wanted to include some examples which would illustrate some of the questions I get asked. So the purpose of my design was to be an example for training purposes, yet I still wanted or hoped for a system that would be honorable at pulling some money out of the markets. However, it was not my intent to find and give away a phenomenal trading system. I have not put in the hours of research that a commercially sold system deserves to have been researched and back tested. That said, the system that will be shown today is still a decent contender, and many of you will probably find something of use in its design to factor into your way of doing things.
16:46 Ensign Howard: The Stochastic study has popularity, so I decided to try to find a system based on Stochastic. Stochastic can also be representative of using any study as a building block in the design of a trading system.
16:46 Ensign Howard: So 1st criteria was to try to reference a study, a 2nd objective was to factor in time constraints. A 3rd objective was to do some pattern recognition.
16:47 Ensign Howard: A 4th objective was to manage a trade such as put 2 contracts on initially and then conditionally lift a portion of the position at a profit objective and let the other portion of the trade ride for a run, if possible.
16:48 Ensign Howard: So, we have lots of things to consider or try to achieve in the design of the trade system.
16:48 Ensign Howard: To begin with, I referenced my earlier Trading Tips article on the Fundamental Behavior of the Stochastic study.
16:49 Ensign Howard: In an Elliott 5 wave move, the Stochastic will have an initial thrust above 50, then another thrust above 70, and then a 3rd hump for the 5th wave and this hump is divergent. See the first picture in the April newsletter showing this classic wave and its Stochastic behavior.
16:51 Ensign Howard: Knowing this, we will use the Stochastic characteristic of the thrust above 50 as the signal to enter the new trend, and then skip the first Stochastic downturn and exit on the 2nd or 3rd Stochastic downturn. Looks like a good plan on paper, and we will use back testing to confirm its worthiness of uselessness.
16:52 Ensign Howard: The general rules listed in the article apply to the trading day, such as limiting the time when signals can be taken. I did not want a signal to be taken minutes before the market closed, for example.
16:53 Ensign Howard: Also, I did not want the position carried over night, so there is a force exit of any position based on the time of day, such as 15 minutes before the session close.
16:54 Ensign Howard: And, another oft asked question is to see an example of exiting a position of some profit objective has been achieved.
16:54 Ensign Howard: The general rules outline the rules to be in the system design.
16:54 Ensign Howard: The specific rules for entering a Long trade or a Short trade are listed in the article, and basically are a statement of the signals to be derived from the Stochastic curve. Occasionally a position will be wrong, and we need some mechanism to bail out of the trade.
16:55 Ensign Howard: One method could be to have a protective stop.
16:56 Ensign Howard: The method for a bail out I chose to use was to watch for a Stochastic value that is on the wrong side of the trade. The rules list the idea initially designed for and tested.
16:57 Ensign Howard: In the article, scroll down to the Time Conditions section.
16:58 Ensign Howard: Here a DYO is used to test the bar's time stamp against some Time boundaries. The time to test is listed in the Number fields on Line A and B. The time of 7:00 is entered as 700, i.e. omit the colon.
16:59 Ensign Howard: You would edit the Number fields to change the time boundaries to match the time zone and the market hours for the symbol to use.
16:59 Ensign Howard: The example was for trading the EUR/USD which is a 24 hour market, but restrict the hours of trading to be 700 to 1500 Eastern time.
17:00 Ensign Howard: The results of the Time test are stored in GV[10] for later testing when potential signals are to be taken. If this flag is not True, the signal will not be generated.
17:01 Ensign Howard: Any questions about how a time Rule was or could be implemented?
17:01 mikey: Explain the syntax of the AND(#)
17:02 Ensign Howard: The AND[#] statement is a Boolean operation that will do the AND operator on a set of Global Variables that are in sequence.
17:02 Ensign Howard: The SIZE of the set is the number field parameter which in the example is 2.
17:03 Ensign Howard: The START of the GV set will be the Read GV reference, which in the example is a 1.
17:03 Ensign Howard: So the RESULT of this statement will be Result = GV[1] AND GV[2]
17:03 Ensign Howard: Note that Line A put its flag in [1] and Line B put its flag in [2]
17:04 Ensign Howard: So we AND both flags together to have a True condition only when both flags are True.
17:04 Ensign Howard: So C is True when Time is after 7:00 AND before 15:00.
Buy Signal
17:05 Ensign Howard: OK, moving on we want the Buy signal to be based on a thrust of the Stochastic upward and crossing 50.
17:05 Ensign Howard: I did not want a sideways market to dribble back and forth across 50 and cause signals, so I added to the Thrust concept the condition that the Stochastic crossing of 50 had to be preceded by being below 30.
17:06 Ensign Howard: Now we do not know if that below condition happened on the prior bar or happened several bars earlier. All I know is I need a mechanism to be set by the below 30 and used in the test when crossing 50
17:07 Ensign Howard: We will use a Global Variable of [250] to remember when the below 30 condition exists. This will be like cocking the trigger on a gun. Then if and when the Sto crosses 50, if the trigger was previously cocked, the signal will be fired.
17:08 Flag: We do not need to know how far back was the dip?
17:09 Ensign Howard: That is correct Flag, and therefore we need to manage this GV and reset it after it is used.
17:09 Ensign Howard: A private GV in the range of 240 to 255 needs to be used because the contents of this GV needs to be remembered for this chart between calculations. This system might be running on more than one chart, and we would not want a Sto on an ER2 chart to be cocking a signal on a YM chart. A private GV that can only be access by the chart owning the DYO is used to prevent this cross contamination potential.
17:11 Ensign Howard: Line D is the test for the Sto being below 30 and sets a Flag in GV[1]. This flag is tested by Line E to cock our 250 trigger GV.
7:12 Ensign Howard: The Line E statement reads as follows: IF the flag in GV[1] is True then Return the value from the Number field. And if the flag is False, then return the contents of the Value GV reference. The Value reference is the Read GV which is 250.
17:12 Ensign Howard: The returned value is then written to the Write GV reference and that is also set to 250.
17:13 Ensign Howard: So the net effect of this statement is to put the value 1 into GV[250] when the flag is True, and leave the contents of 250 alone when it is False.
17:14 Ensign Howard: When False, we read and rewrite the same content of 250, so it does not change when False.
17:14 Ensign Howard: Any questions about this [250] event being set when seen and the clever use of the Line E statement to set it when it is True?
17:14 Shar: I hope I didn't miss this, but what setting is your Stoch 9? Simple, exp, or Howard's"? And how does one reset that GV concerning the 50 cross? Or is it reset automatically? Is that what you explained?
17:15 Ensign Howard: Shar, the Sto settings used were Bar=9, %K Ave = 5, and Formula = Howard's. The 50 crossing is to be discussed next.
17:16 Ensign Howard: Line F is the test for the crossing of 50 going in the Upward direction. The notation on the Line F statement can be read as %K Cross Above 50, the X> means Cross Above. This is an event that will be True for a single bar.
17:17 Ensign Howard: That is different than the relationship test done on Line D which was %K <= 30 which was Less Than or Equal to 30.
17:18 Ensign Howard: Line D can be True for several bars in a row, which is fine, it just continues to put a 1 value in [250]
17:19 Ensign Howard: Now we are ready to see if we have a Buy Signal which is the AND condition of 3 flags being True simultaneously.
17:19 Ensign Howard: The 3 flags are Line F: %K X> 50 -- Our below 30 condition setting our [250] to a value of 1,7 -- and our Time Rule is True, meaning the time is between 7:00 and 15:00.
17:20 Ensign Howard: So, Line F puts the %K X> 50 flag in GV[11]
17:21 Ensign Howard: Line G copies the [250] contents down to GV 12 just so it will be in a sequence for use in the AND function.
17:21 Ensign Howard: [250] was written as a value of 1, or could have a value of a zero.
17:21 Ensign Howard: Just so happens that zero is used for a Boolean False, and ANY non-zero value will be treated as a Boolean True. So the writing of a 1 value to [250] is equivalent to putting a Boolean True flag in [250]. That little bit of knowledge is not obvious to you, so that is why I am explaining it in greater detail.
17:22 Ensign Howard: So [12] has a flag of the below 30 condition did happen earlier.
17:23 Ensign Howard: And we already have our Time Rule in GV[10]
17:23 Ensign Howard: So Line H can be the 3 condition AND test starting in GV[10]. Signal = [10] AND [11] AND [12]
17:24 Ensign Howard: Signal = Time Rule OK AND (%K X>50) AND (Below 30 happened earlier)
17:24 Ensign Howard: The location of this signal is marked in the example with a vertical red stripe. In the final design a green stripe was used.
17:25 Ensign Howard: Any questions so far about the Buy Signal?
17:26 Ensign Howard: In the example it shows a pale green for where the Below 30 is cocking the trigger, and a red stripe for the signal being fired.
17:26 mikey: That must be the ZON?
17:26 Ensign Howard: yes, the stripe uses the ZON marker
17:26 Kongfuzi: Green strip is the vertical shading?
17:26 Ensign Howard: yes kongfuzi.
17:27 Ensign Howard: This signal is stored in GV[1] and ready to be used by Line I which has the purpose of clearing our Below 30 event flag.
17:28 Ensign Howard: So Line I reads, when the Flag in GV[1] is True, put the Number 0 into GV[250], and when it is False, write [250] with its existing value. Effect of Line I is to clear our below 30 event flag in [250] so that if we dip below 50 and recross we DO NOT get another signal.
17:29 Ensign Howard: The STO would have to go below 30 to recock the taking of another Buy signal. So this is a good illustration of some techniques for pattern recognition
17:29 Ensign Howard: We have part of the pattern needing to be below 30 and then a crossing of 50 in an upward direction.
17:30 Ensign Howard: The 2 need to work together an the Below 30 happens ahead of the crossing above 50.
17:30 Ensign Howard: Any questions about this implementation?
17:30 Flag: neat
17:31 Ensign Howard: Some possible areas of research would be to see if different levels for the crossing and the Below improve or worsen the system.
17:31 Shar: Brilliant
17:31 Ensign Howard: I just eye balled the chart and it looked like good signals to enter were when STO crossed 50 but only if preceded by being below 30. That is the formation seen on the theoretical chart used to show the fundamental behavior, and it is also often seen in real-time too.
Sell Signal
17:33 Ensign Howard: Ok, lets move on, we have the design of the Buy Signal in place, and the Sell signal is a mirror in the logic, and just uses a 2nd private GV to cock the Sell trigger when the STO was above 70 before it crosses below 50 going down.
17:33 Ensign Howard: In the Buy Signal DYO, we had the buy signal flag written in GV[1], so we then added a Study Alert following this DYO to do the actual buying.
17:34 Ensign Howard: The form is shown in the article next.
17:34 Ensign Howard: The Panel A reads the flag in the GV, and when It is True, that Trading System action of Buy Reverse Long is executed with a quantity of 2 contracts.
17:33 mikey: Howard, what if one wanted to test a longer term stochastic running on another chart? Say, like a 15 minute chart?
17:35 Ensign Howard: Mikey, you can do what you seek by having the study on the 15-min chart write its values to a GV for a study Transfer. HOWEVER, you will not be able to back test very well because you cannot see the 15 min signals bar by bar being transferred. You will ONLY see the final transfer in the GV. The two charts will recalculate independently of each other. That is fine in real-time but not fine for trying to back test results. Do you understand that issue or challenge?
17:37 mikey: Yes, but it would work for visual backtesting.
17:37 Ensign Howard: No it will NOT work for visual backtesting. You will only have correct signals in real time (or doing a playback) When the 15 min chart is recalculating, it has no way of knowing its results are also needed bar by bar by a 5-min chart.
17:37 Ensign Howard: ok, we can move on again
17:38 Ensign Howard: The Study Alert is what does the trading in the system, and it also is what puts on the dark green stripe for Buying 2 contracts.
17:38 Ensign Howard: The other striping in the DYO are unchecked in the final design. They were useful, however, to see an intermediate step during the design to verify the logic as it was being created.
17:39 Ensign Howard: Any questions about the Study Alert doing the trade system buying?
17:39 Ensign Howard: I used Buy Reverse Long instead of just the Buy selection so that if the system was short when the signal was seen, it would buy enough to close out the Short position and then put on a Long position of 2
17:43 mikey: How can I have the words show up too?
17:44 Ensign Howard: Mikey, you would add test to the Message section and check the box to show an Alert message box on this Study Alert. Check the Message | Show check box
Profit and Exits Signal
17:42 Ensign Howard: Scroll down to the topic of Profit Exit in the article.
17:43 Ensign Howard: Line A of the Profit Exit DYO puts the Position Size in GV[1], and the Profit in Points in GV[2] for use in testing in this DYO.
17:45 Ensign Howard: Line C of the Profit Exit DYO tests to see that the Position Size is 2, meaning Long 2 contracts, which is the case following a Buy Signal. This test result is stored in GV[3].
17:46 Ensign Howard: Line D tests the Profit in Point against the exit condition of being greater than or equal to the Number field which has a 0.0010 entry.
17:46 Ensign Howard: For my EUR/USD chart is 10 pips. In the final design I found that a profit exit of 0.0015 for 15 pips was more profitable.
http://www.ensignsoftware.com/tips/tradingtips88.htm
17:47 Ensign Howard: For the EnsignStoES template for use with 5-min ES charts, the Profit Exit objective on Line D was set to 4.75 points.
17:48 Ensign Howard: Now we can test for the 2 conditions in [3] and [4] being True simultaneously using Line E
17:49 Ensign Howard: Line E means Result = (Position Size = 2 Long) AND (Profit above 0.0010 points per contract)
17:49 Ensign Howard: When this condition for a profit exit is True the chart is marked with a light Green Stripe and the result is saved in GV[10]
17:50 Ensign Howard: Line F tests for a Position size of negative 2, which means Short 2 contracts.
17:50 Ensign Howard: Line G is the test for a profit exit when Short, and this flag is stored in GV[11] and marked on the chart with the salmon colored line.
17:52 Ensign Howard: A study Alert then Tests GV[10] flag, and when True it does the trade system action of Selling 1 contract. This will reduce the Position Size to 1 and then the DYO signal for a profit exit cannot be triggered again.
17:53 Ensign Howard: A 2nd DYO was used to watch GV[11] and Buy 1 contract when Short.
17:53 Ensign Howard: Any questions about the implementation for the Profit exit DYO and its 2 companion study alerts?
17:54 Ensign Howard: Again, if you want to see an alert message in real-time, check the Show box in the Message section on the Study Alert form.
Stochastic Exit Section
17:54 Ensign Howard: Scroll down to the Stochastic Exit section which we will discuss next.
17:55 Ensign Howard: This DYO is another pattern watching DYO, where we need to count the stochastic tops and when we get the 2nd or 3rd top, we have an exit signal.
17:55 Ensign Howard: Line A is reading the DYO Buy Signal flag and putting the value in GV[1].
17:56 Ensign Howard: Line B is testing the Signal to resent a Top Counter in a private GV [240] to zero
17:56 Ensign Howard: So when we Buy 2 contracts our Top counter is reset to zero.
17:57 Ensign Howard: Line C and D will also reset the Top Counter to zero when we exit all positions and are Neutral.
17:58 Ensign Howard: Line E and F do a formation test of the STO %K line being ascending on the prior bar and descending on the current bar.
17:58 Ensign Howard: This combination is our indication of a top in the Stochastic curve.
17:59 Ensign Howard: Note the use of the Bar Offset of -1 on the Line E...so the %K Rising Test is on the prior bar.
17:59 Ensign Howard: Line E and F both being True mean we have an Up Down combination on the %K curve.
18:00 Ensign Howard: That Top test is True on Line G and is used to then INCREMENT our Top Counter in [240] by Line H.
18:01 Ensign Howard: Line H would read. If the pattern is a Top, then read the value in GV[240] and add 1 to the value and resave in [240]. If false, leave the contents of [240] alone.
18:01 Ensign Howard: Line I is then our test to see if the Top counter has arrived at the number 2, meaning we are at the 2nd top pattern.
18:02 Ensign Howard: This is our condition to liquidate the Long trade, which might be 2 contracts or it might be 1 contract. We do not know if this exit condition was preceded by a Profit Exit which removed 1 trade.
18:03 Ensign Howard: In the final design, I found it more profitable to wait for the 3rd top to exit, so the count test on Line I has a Number field of 3 instead of 2. And this seems like a logical change since in our Fundamental Behavior image, we would be exiting at the end of wave 5 instead of wave 3
18:04 Ensign Howard: Any questions about this important DYO for counting top formations in the Stochastic pattern? A similar DYO counts stochastic bottoms and increments a different counter in [241]
18:06 Ensign Howard: Since the action to take is to go Neutral, and this action is taken regardless of being long or short, a single Study Alert can be used to accomplish the task by watching the Top Exit condition in [10] and the Bottom Exit condition in [11]
18:07 Ensign Howard: And finally, we have a Study Alert that watches the Time to exit any position if any at 4 pm, which is an hour after we permitted the system to initiate any trades. So the system cannot put on NEW trades between 15:00 and 16:00. It can, however, exit trades at profit exits or sto top and bottom signals. If neither happen then the 4 pm exit will force going neutral so no position is held overnight.
18:08 Ensign Howard: OK, that is the system.
18:09 Ensign Howard: It is of moderate complexity and shows some clever combinations of signal requirements, and 3 different types of exits.
Wrong Way Exit
18:09 Ensign Howard: Then we have a 4th type of exit which I call the wrong way exit, meaning we got into a trade that did not work out, the market is moving is the opposite direction.
18:10 Ensign Howard: Example we sent Long with a Cross above 50, failed to get the profit exit, failed to get the triple Stochastic top, and now the market is moving down and the Stochastic responded by going lower, such as below 40 or 30.
18:11 Ensign Howard: I initially thought below 40 would be a good exit, but found in back testing that for a Long, we can often hit 40 is sideways movement noise, and that more winning trades were kept on if the trigger to exit on a wrong way long was a 30 level instead of 40.
18:12 Ensign Howard: So in the final design, the wrong way long tests for Stochastic being below 30 and a wrong way short tests for Stochastic being above 70.
18:13 Ensign Howard: Ok, lets look at some examples
18:13 Ensign Howard: Here is today's first trade in EUR/USD 5-min chart using the system.
18:14 Ensign Howard: Dk Green is the Buy Signal after 7:30, position is Long 2 contracts.18:15 Ensign Howard: Light Green stripe is the Profit Exit at 0.0015 pips
18:15 Ensign Howard: Cyan Stripe is the 3rd Sto Top exit
18:15 Ensign Howard: And you can easily see the Elliott 5-wave pattern in this example
18:16 Ensign Howard: We got on board nicely early on, got a profit in the 3rd wave, and let the other contract run through the 5th wave. This trade looks pretty ideal.
18:16 Ensign Howard: You can see the 3 humps in the Sto %K line plotted in blue in the sub-window
18:17 Ensign Howard: Now for 2nd trade of the day (today) in EUR/USD.
18:18 Ensign Howard: The market being above 70 cocked the trigger for a Short sale, and the Red stripe shows where it went short 2 contracts when Sto crossed below 50.
18:19 Ensign Howard: Unfortunately, it chopped a bit but did not trigger a profit objective of 0.0015 on a CLOSE. It is making the profit objective on the Low of the bar in the next swing down
18:19 @Buffy2: Howard when this template is on a chart other than minute, the exit signals do not seem to given. Case of optimizing for that time frame?
18:20 Ensign Howard: Probably Buffy. Particularly need to have Stochastic rhythm match the cycles in the chart you use. Count the number of bars in a cycle, and try a Stochastic Bar count that is half that count. Then backtest using higher or lower Sto bar count from this initial setting to see which works best.
18:22 Ensign Howard: On our profit exit, I guess I should change the Profit In Points test to do my own test between entry price of the Close and the Bar's Lows when Long and the Bar's Highs when Short.
18:23 Ensign Howard: I see in the example, the profit of 0.0015 was achieved, but not marked because the evaluation was done using the bar's close instead of its low.
18:23 Ensign Howard: We would have or could have had a profit exit at around 10:10
18:23 Ensign Howard: Then the market went higher, against the short position, and you see the Rule for a Wrong Way short exiting the trade at the Orange stripe.
18:23 Flag: Hence my fight for Open only option
18:24 Ensign Howard: Open Only would not change any part of this discussion about a profit exit being triggered by the Low when short.
18:26 Ensign Howard: This last example shows a Short at 13:25 that dribbled sideways and was exited on a 3rd bottom in the Sto pattern at 15:20.
18:26 Ensign Howard: No new trades were taken after 15:00.
18:27 Ensign Howard: Today was slightly ahead $100, but yesterday was a wonderful day being ahead $1300. This chart shows the trades for yesterday, which were Short, then Long with a small loss.
18:28 Ensign Howard: Then profit exit long and Neutral at cyan stripe.
18:28 Ensign Howard: Then Short, and short profit exit and then Reverse long and out with small loss on the Long.
18:29 Ensign Howard: Then Short, then reverse Long, and that Long exited with a profit later in the day.
18:29 Ensign Howard: So the system did not overtrade yesterday, but it did have 3 very nice winning trades and they more than paid for the 2 small losers.
How to Measure Cycles
18:30 mikey: How do you measure cycles to pick the stochastic? Which swings do you choose?
18:30 Ensign Howard: Estimate the number of bars in cycle bottom to bottom and use half for your STO bar parameter.
18:30 Shar: Does Mikey's q have to do with how to figure number for AB/ER and YM?
18:32 Ensign Howard: Here I put a Cycle tool on the chart and measured from bottom at 13:15 to bottom at 11:40 and the count is 18 bars and that is very agreeable with the Sto using a count of 9 for Bar parameter. So the example substantiates the directive I have given
Close only
18:33 Flag: Howard, what happens when Stoc based on Close crosses 50 and then retracts its signal? Is it recorded in profit or erased, for example -- I stumble with this unless I use Open based indicators. Obviously H/L price based exits are fine.
18:34 Ensign Howard: All of my DYOs and Study Alerts have the CLOSE ONLY option checked so that the signals are evaluated ONCE per bar at bar close. Trade uses that signal bar's close as the execution price. That way in real-time you do not have a signal give and then taken away. With Close Only checked, then Signals given will be permanent
18:35 Kongfuzi: Wonder how changing it to open would change the results.
18:35 Ensign Howard: DO NOT CHANGE the execution price to Open because that can badly inflate results
18:36 Ensign Howard: If you are Selling, chances are it is a down bar because signal is a cross below 50 and thus the OPEN for this bar 90% of the time will be higher than the close and you cannot get filled at such a favorable higher price.
18:37 Ensign Howard: So when you see the signal, place the trade. You can probably get a fill like unto the bar's close. Sometimes your fill will be at a better price and sometimes at a worse price. I think it averages out.
18:38 Ensign Howard: What you are trying to do is backtest whether one idea is better or worse than another and results based on Close price do that.
18:39 Ensign Howard: No warranty is made that this system will continue to make money like the results published in the article. However, in 3 days of using it on the EUR/USD its results for 3 days are:
18:42 Ensign Howard: A move from $6375 to $7962 with NO commissions or slippage deducted. Looks like it did 16 trades in 3 days. Some of the ledge entries are for the partial lifting of half of the position on a profit exit.
18:43 @Buffy2: Why did you choose the EUR/USD?
18:43 Ensign Howard: I chose EUR/USD for the illustration because I have received requests for more forex examples. EUR is an active contract with good vertical movement each day, yet does not get abusive in wild swings.
18:45 mikey: How did the system perform in non-trending conditions? Did you find determination of trending vs. nontrending significant in backtesting results?
18:45 Ensign Howard: Position was always initialized with 2, and then possibly lift 1 at profit exit, which then splits the ledger log into 2 trades recorded.
18:45 Ensign Howard: Line LIne 241 and 242 -- you see QTY of 1 for both.
18:46 Ensign Howard: See the Notes on those lines show trade time for both began at 8:35. One was lifted at 9:25 and other at 10:00
18:46 scooter5: Just curious , was this watched tick for tick real time to see where one would actually get filled or is it all up to the computer where the system fills you?
18:46 Ensign Howard: yes I watched them real-time scooter
18:46 scooter5: Okay thanks
18:47 Ensign Howard: Go look at the examples and see the next bar open or how it covered the signal bar's close with its range to get a feel. I think it is very doable.
18:47 Ensign Howard: Site has 2 templates: the one for the EUR/USD as shown in the article is EnsignStochastic
18:48 Ensign Howard: 2nd template modified for hours and profit objective for ES is named EnsignStoES
18:48 Ensign Howard: Now you know as much as I do and can continue your research or adaptation of the concepts taught to work with other symbols, time frames, profit objects, study parameters, etc. I am trying first most to be a teacher with this example. And after its release, at least I am not embarrassed......;-)
18:49 @Buffy2: this is excellent for learning how to use dyos and alerts for systems/backtesting.
Adjustment to Template to enable use on All Symbols
18:50 croc: Hi Howard, it's Chris, I wrote the second article and I played a bit with your template today. It's actually very simple to expand the EUR/USD template to a universal template applicable to all kinds of contracts. Just use no fixed parameter for the profit objective but calculate it as (profit in ticks)*[$Q]
18:51 bg: Fantastic work there Howard
18:51 croc: Here's a copy of the DYO modification -- just 2 lines added. [Q] = The minimum tick size as set in the Symbols settings. Makes it extremely easy to have the templates based on ticks.
18:51 Ensign Howard: Thanks croc, you are exactly right about that issue. Oh, thanks again, you even show the example and personally I would probably move the Number count from Line D to the top of the DYO to find it easier or if room in 1st DYO of the system to put it there so I have just one place to edit user parameters like that.
18:53 Ensign Howard: Let me show a similar little parameter trick involving the Top and Bottom counter being 2 or 3
18:54 Ensign Howard: On the DYO line with the Number = 3 you can change that to read a GV content like [9] and then in the 1st DYO add a line like croc's Line D where Number category puts a value in GV[9] such as 3 or 2. Then you edit the Number field and it is used by 2 places in later DYOs.
18:56 Ensign Howard: There will be no end to the custom tweaking that could be done for various reasons. But for the educational purposes, the example is sufficient and succinct.
18:56 Flag: It's a great contribution for us -- we amateur systemizers. Thank you, Howard.
18:56 Ensign Howard: you are welcome, and there are lots of concepts illustrated such as one study condition preceding another and counts occurrences to wait for a subsequent signal instead of taking every one
18:58 Ensign Howard: croc, thanks for the article you wrote for the newsletter. That was really appreciated and got me going to add a companion article using just DYOs and Alerts
18:59 croc: your welcome
18:57 Ensign Howard: ok, lets call it a day.
18:57 bg: Howard this is great thanks -- fantastic work
18:58 mikey: excellent. thanks.
18:58 Ensign Howard: so you have a great week,and we can have another exciting class next week.
18:58 bg: take care Howard
18:58 @Buffy2: Thank you so much for really walking us through it -- have a good evening
18:58 DYoung: Thanks Howard
18:58 croc: a really great template Howard...with a lot to learn from
18:58 cyndc: tks -- Shar: Thanks so much Howard. -- aww: thanks Howard
18:59 scooter5: thanks Howard -- Kongfuzi: thx HA, great stuff