IMPORTANT NOTE: PLEASE READ THIS FIRST!
I was alerted (by alert reader Satish) that the data that I have for ticker ULPIX is incorrect!
There is Bad News and Good News
The Bad News is that – and yes, this is embarrassing but – an alert reader noticed that the data that I have for ticker ULPIX is incorrect on several dates and as a result the performance numbers in the article below are incorrect.
The Good News is that there is still “a Happy Ending” in Part 3 which appears here. So while you can read the article below for the calculations that will ultimately be used in Part 3, remember that the performance numbers are incorrect.
For the record, here is what happened: Under the category of “from sources believed to be reliable” I have not reviewed the data that I have for ticker ULPIX in some time. Which is a shame because as it turns out the data that I have for ticker ULPIX and appears to contain random dates with incorrect prices. This ended up having the effect of mathematically improving performance, but did not reflect reality. Instead of using SPY data in my initial test, I went straight to the ULPIX data. Had I tested the results initially using SPY data instead of ULPIX data I would have noticed the discrepancy when I then went to use ULPIX data.
So two takeaways:
1) Yes, this is embarrassing and I should have verified that the data I was using was accurate before “going to press”. So I apologize for that.
2) Don’t give up on this whole “Trading Stocks using Bonds” thing until you read Part 3.
Jay Kaeppel
Jay: I cannot seem to replicate your results at all. I am using yahoo data. I have tried both adjusted close and non adjusted close…..44 day and 252 day results are very different from yours for adjusted close prices. My results are quite a bit lower. 252 day results are close for non adjusted prices but 44 day results are still quite different. Could you please check your results, I will be glad to send you my excel file.
Satish, thanks for the comment. One possibility – and perhaps I did not spell this out clearly enough – is that with mutual funds there is a one day lag. Allow me to explain. If as of Wednesday’s close, a model is “bullish”, then ULPIX will be bought at the close on the next trading day (presumably Thursday) and the “long trade” starts fluctuating on Friday. In this meantime I am reviewing my data and calculations just to make sure. Jay