Ensign Mirror

Q: The "Net Change" is between this bar's close and the previous bar's close? I have seen implementations where the difference is between open and close, but it seems more logical to use close vs. previous close.

A: Our Ergodic property form allows you to select a data point, and what ever is selected, such as Close, uses the same data point in the prior bar to make a Net between the two prices. I do not have a selection that would use Open and Close of the same bar.

Q: As I understand it, the ERG is usually specified as three parameters, X, Y, Z where X is the length of the first average applies and Y the second. Z is the length of the moving average that is finally applied to the ratio you mention. So, if I specify Weighted in the ERG property form, does it mean that the averages for X and Y will be weighted? What about Z? Which type of averaging is used for Z?

Yes, I realize that the indicator will always behave slightly different depending on where exactly the first bar is constructed on your chart. That will most likely vary between different charting application for sure.

A: Ave 1 and Ave 2 are the parameters for the 2 averages in the numerator and the 2 averages in the denominator. This makes the ERG study line using the average formula selected.

Now a 2nd study line which is an average of the ERG line can be plotted, and it will use the Average parameter on our property form. This average can use a Simple formula if the Simple option box is check. If Simple box is unchecked then it will default to being an Exponential formula.

Q: I have much enjoyed my experience with Ensign even though my usage is at a relatively simple level. One technique that I've found quite helpful is using Stochastic to screen out false candlestick signals. Of course Stochastic has real limitations in the overbought/oversold arena. One potential solution to this problem is to also use the indicator known as 'true strength index.' I was reminded of this via a recent discussion by Mark Phillips beginning on page 25 of the June 2005 issue of Technical Analysis of STOCKS AND COMMODITIES. This indicator appears to have a niche role as a window on the strength of price action at times when stochastic is frozen in the overbought or oversold zone--especially when it's flatlining there. So......can you give us this indicator?

A: The formula you show is known in the industry as the Ergodic study, and the Ergodic study is already in Ensign Windows.

Q: I must have the Ergodic indicator as consistent as possible. I followed to the detail the discussion you had with Bill in the Market Logic room about the discrepancies in the calculation of the ERG generated by the chart refresh and the data source. I asked questions about the best approach to minimize the differences. The conclusion was to refresh the minimum and only when needed.

Knowing that IB turns their servers off everyday for 1 hour around midnight ET, I decided to turn my computer off at 4:15pm ET every day, turn it back on in the morning around 8am ET, run a chart refresh and leave it so for the rest of the day. With that in mind, I noticed marked changes in the ERG calculations today after trimming down the charts.

Please explain how is the ERG calculation affected by the chart max bar amount. I trimmed down the charts today and my ERG indicator differs more than it used to to other people in the room. It seems more sensible and it's missing quite a few oscillations now.

A: The problem is ERG is using exponential averages which means the current calculation includes some contribution from every bar in the chart, though admittedly the further away a bar is from the current bar, the smaller its contribution. Thus to have the same calculation as anyone else you both would need the same data set. The 2nd problem with your ERG is you all are using HUGE average parameters like 147. Huge parameters means that older bars contribute more to the current calculation and a small parameter would. So dropping old bars changes the ERG both by changing the initial starting point and changing the calculation because each bar is a contributor to the answer.

Q: What other variables in Ensign set up, chart set up, data feed set up, etc are taken into consideration or affect the calculation of the ERG indicator? I appreciate your insights, I'm a little confused, and I can't afford not to have the ERG indicator as consistent as possible. I want to find the best balance between not using unnecessary bandwidth for data I already have and having consistent charts/ERG indicators which are my main trading tool. Hence the reason of all my questions, it used to work well, however I wasn't considering the max bars in charts and the refresh function.

A: ERG is a ratio calculation of Net changes, so it is sensitive to the closing price. If you use constant tick bars, then any addition of ticks via a refresh is going to cause all bars thereafter to build differently, and cause the Net from one bar to the next to probably be different. This causes ERG to change. Usually the value of the study is similar to what it was before but it is not the value you are watching. You are using the slope of the line to be rising or falling to color your bars, and the slight changes can cause the values near zero to be rising or falling slightly differently, and to you the observation is a big change because bars color differently. The actual change can be out in the 6th decimal place, and in my opinion be trivial. The bar coloring makes you think the change was bigger than it really is. So it is all a perception thing. My suggestion is don't worship the ticks, and your ERG is just as valid as anyone else's though they seem to differ. It is like telling time. My watch says 10:55:23 and your watch says the time is 10:54:57. Who is right? We are both right because our watches perform the function of telling time and we both manage our affairs accordingly. The time on our watches is relative to what we need them for. Technically neither is the pure accurate time that matches the government's official time.