Recommended Reading

Google Search

Google
 
Charts courtesy of stockcharts.com

Monday, February 18, 2013

Price Relative Scan Extended Post

I made a couple of changes in the way I scan for stocks in 200 day price relative up trends that may reduce the amount of turnover I get every week:

1. I eliminated stocks which do not have at  least 101 days f price data. Most of these are recent IPOs. Since they do not have price data for the time period of 101-200 days, the maximum value of price relative for that time period is zero, which was causing a lot of false positives on the scan. That eliminated about 80 stocks from this weekend's scan. That probably does not, however, address the issue of false negatives.

2. Price relative is simply stock price divided by S&P 500 price. For most stocks, that results in a small number, but for stocks under 1, it results in an extremely small number, and I was thinking there may have been rounding error in the spreadsheet. I multiplied the results by 100 to get a higher number and less likely hood of rounding error. So far I have noticed no difference in the scan. I will try different methods (such as multiplying stock price before dividing it by SPX), but so far this is starting to look like it might be a dead end.

Hopefully I have eliminated most of the false positives, but the issue of false negatives remains, and it is the one that concerns me most. One problem is splits: I have been doing a fairly good job of catching them, but some stocks split without making any announcement, however, that should at most affect one or two stocks a week. Another problem might be erroneous data, either from Finviz or Yahoo finance. If that is the case, I am pretty well screwed, because there is far too much data for me to check it daily. I will keep you posted on any progress made.

I am starting a new project of the price relative scan. Since I calculate the difference between this week's price relative scores and last week's, I can sort them by highest difference. This week I am going to look at the 30 stocks which have the highest gain in advancing score (here is a Finviz screen). The theory is that stock with a high gain in score are undergoing accumulation: for some of these, that might actually be true. I have historical data, so i will be going back to see how past stocks did, but that will be pretty time consuming, so I won't have results there for a while, but I can track these in real time, so we should see results quite soon.

No comments:

Google Analytics