 |
GS
2012-10-10 11:32:14
|
|
Hi,
I am using a screener to filter cross(ema5,ema14), I have observed that for many cases on the chart we can see that the crossover has come, however screener is unable to filter it at that time. In case of 15min chart it takes another two candles to actually filter it. If I change the time frame to one hour, for many cases it is unable to even filter the crossover, which otherwise we can "see" visually on the chart. Is it a bug? or some other issue? Looks like all crossover which are visible need not to be filtered, but all filtered crossover can be seen on the chart. Please correct me on this.
|
|
|
 |
QuantShare
2012-10-10 11:49:16
0
|
|
First of all, make sure that the EMAs in your chart are using the same scale ("Plot" without StyleOwnScale). Otherwise, the visual confirmation may be wrong.
Are you downloading data using a script or are you using a real time data feed?
|
|
|
|
|
GS
2012-10-10 12:05:47
0
|
|
I am using script to download the data, will it create problem? I use the following statement to plot:
ema5 = ema(close,5);
Plot(ema5, "EMA5", colorBlue, ChartLine, StyleWidth2);
Is this the correct way to plot?
|
|
|
|
|
QuantShare
2012-10-10 12:21:36
0
|
|
What about the other EMA?
Please note that the screener analyzes bars when they are formed. You can see a crossover in a chart during the formation of a bar and this crossover may disappear when the bar is formed.
|
|
|
|
|
GS
2012-10-10 12:32:03
0
|
|
For other EMA also it is the similar statement, like the following:
ema14 = ema(close,14);
Plot(ema14, "EMA14", colorRed, ChartLine, StyleWidth2);
Lets say I am running screener today at 10:15am (actual clock time) with the time on the screener as 9:45 (half an hour before). If there is a crossover at 9:45am, which I can see on the chart, what should I expect?
Similarly, if I am doing back testing over a 15 minute time frame, is there any possibility that some valid crossovers might be missed out?
|
|
|
|
|
QuantShare
2012-10-10 17:22:07
0
|
|
You should expect a signal on the screener.
I have test a simple EMA crossover and can see both signals on chart and screener at the same time.
Make sure that the screener date (not time) is set correctly.
is there any possibility that some valid crossovers might be missed out?
No.
|
|
|
|
|
GS
2012-10-10 18:56:18
0
|
|
OK, Looks like this problem occurs mainly just at the time of opening or closing the market, I have seen it at 9:30am (market opens at 9:15am) or at 15:30pm; in between it works well.
|
|
|
|
|
QuantShare
2012-10-10 19:33:23
0
|
|
It could be because you are displaying charts with regular trading hours only and you are using the screener with all the data.
In that case, the calculation of EMAs may become a little bit different.
You can have to screener perform calculation based on regular trading hours only by clicking on "Options" (next to "Select a time frame") then unchecking "Use data outside regular hours".
|
|
|
|
|
GS
2012-10-10 19:53:50
0
|
|
I think within "Options", "Use data outside regular hours" is available with more recent version of QS, I have 2.8.5, and I do not see any thing to uncheck within screener. Do let me know if this is the case. On the other hand I display charts by "hiding data out side the market hour".
|
|
|
|
|
QuantShare
2012-10-10 20:13:45
0
|
|
Yes, this option is not available in version 2.8.5.
Another important thing you must check is the number of days loaded in your chart. If you aren't loading all the data then the calculation of EMAs may differ a little bit.
|
|
|
|
|
GS
2012-10-11 02:47:40
0
|
|
OK, so now I am keeping 500 as "Number of days worth of intra-day data" and within "watchlist/screener" I have "0" to load all data. Hope this helps, but will it not slow down the speed?
|
|
|
|
|
QuantShare
2012-10-13 10:22:01
0
|
|
It will not slow down the app because the formula is very simple.
|
|
|
|
|
0
