Dogs of the Dow Rotation malfunction?
Author: HYPr2
Creation Date: 11/25/2008 4:38 PM
profile picture

HYPr2

#1
I am still fairly new to WLP but love the potential. I have been learning the best that I can but am not a full time programmer. There is something about the function of the program, buy methods, or Dogs of the Dow Strategy (WLP5.1) that I just don't understand. Perhaps someone can shed some light on it for me.

I run the preloaded Dogs of the Dow strategy as stated on the summary tab. To review: Scale = daily, data range = 6 years, position size = 10% equity with just over 1.1 margin $100k starting capital, number of dogs parameter = 10.

The backtest runs ok but I get several inconsistencies.
1) There are 1000's of trades reported missing from the simulation (there are only supposed to be 60 trades TOTAL).
2) Examining the trades there are more than 10 positions taken each year (remember the parameter was set to 10). There are always more trades than the parameter. Seems to be due to the margin.
3) Trades are taken multiple times. For example the same stock is purchased 3 or 4 times.
4) The dividend rates for several years are 0% and seem unrealistic. More recent years are reasonable.

Now I went through the code the best I could (no expert for sure) but it seems the script works as intended. It calculates the number of dogs properly and does the loops properly. It appears to be something behind the "script" that is causing the difference from expectation.

Why should these inconsistencies happen? Am I missing something? Could someone point me in the right direction?

Thanks
profile picture

Cone

#2
You didn't follow the last usage note, i.e., select only 1 symbol in the DataSet, not "Dow 30". Otherwise you'll be running the script 30 times, once for each symbol, and it's not designed for that.
profile picture

HYPr2

#3
Thanks Soooooooooo much! It did work selecting just one symbol. Previously, I was fooled when only the single chart was provided. I skipped over looking at the trades tab to confirm. Thanks for helping me catch this oversight!