I'm interested in finding out how indicators that utilize past data are affected by what past data is present or available, particularly the ATR indicator. For example, 'ATR(200)[0]' incorporates data going 200 bars into the past, so how is it affected by a chart that doesn't have that many bars showing ?
I've noticed on a Strategy Analyzer chart that an SMA indicator responds only to the data shown, which may be restricted by the session times used. In the case where non-market hours aren't shown and the price gaps across the day-to-day boundary, it takes a number of bars for the SMA to "catch up" with the data in the new period because the indicator inaccurately looks back across a gap in the data as if there aren't any missing bars.
Conversely, sometimes it would be desirable to isolate data in this way from the indicator's look-back. For the ATR indicator, looking back on small pre-market or weekend bars will result in an undesirably small value when the indicator is being used during market hours.
It seems that the session times defined in a Strategy Analyzer chart restrict what data an indicator uses -- can I assume that a regular chart contains data from around the clock, including weekends ?
My main concern is that the indicator values that arise in my backtesting do not match what they are when used in a live strategy assigned to a chart, since they are "looking back" at different sets of data at certain times of the day. Is this a legitimate consideration ?
I've noticed on a Strategy Analyzer chart that an SMA indicator responds only to the data shown, which may be restricted by the session times used. In the case where non-market hours aren't shown and the price gaps across the day-to-day boundary, it takes a number of bars for the SMA to "catch up" with the data in the new period because the indicator inaccurately looks back across a gap in the data as if there aren't any missing bars.
Conversely, sometimes it would be desirable to isolate data in this way from the indicator's look-back. For the ATR indicator, looking back on small pre-market or weekend bars will result in an undesirably small value when the indicator is being used during market hours.
It seems that the session times defined in a Strategy Analyzer chart restrict what data an indicator uses -- can I assume that a regular chart contains data from around the clock, including weekends ?
My main concern is that the indicator values that arise in my backtesting do not match what they are when used in a live strategy assigned to a chart, since they are "looking back" at different sets of data at certain times of the day. Is this a legitimate consideration ?
Comment