I hesitate to put this up, but I found a couple files to quickly try. I decided to post it as I think it will help stimulate further discussion.
ThinkorSwim has a quasi-backtester. You can write script for strategies and it will play it out. But there are a lot of caveats, too many to list here, so DO NOT take it as definitive or anything absolute. However, looking at price action (closing price) interacting (crossing above versus below) with an average of the Heikin-Ashi candle body averages (meaning the average of the OHLC of an individual HA candlebody plotted as a line) does show some promise.
This is QQQ weekly graph for 3 yrs. The bottom insert is TOS’s “floating P/L.” Using their measurements, obviously it’s positive. I wrote this script and included the capacity to very the interval and the type of moving average…except triangular EMA (Sorry Arindam, just wasn’t on my radar at the time). If you run the various moving averages you get their estimates of P/L investing $100 per “Buy” and “Selling all” for a sell signal. Below is the period-MA-Profit/Loss:
7-SMA 3304
5-EMA 9752
7-EMA 8898
7-Hull 7104
7-Weighted 6006
7-TEMA 12200
5-TEMA 15250
Interesting. DO NOT TRADE BASED ON THESE COMMENTS. I need to recheck the script and make sure there is not a simple error misleading everything. I have been here before, only to find the bug. I also prefer to confirm with Excel hard data, although, I haven’t written such for weekly intervals…will need to work on that.
The goal is to say that it is interesting and, more importantly, that any perspective and input is wanted, needed and appreciated.
Lakedog