Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : TA-Quotes Plus -- Ignore unavailable to you. Want to Upgrade?

To: Bob Jagow who wrote (9147)	3/29/1999 1:32:00 PM
From: Nine_USA	Read Replies (1) \| Respond to of 11149

Bob, <<What requires you to "retain data on a daily basis">> In order to expose the variables and the weightings I assign to them in order to backtest on the basis of data which was available to the QP2 user on any given 'evaluation' date, I need to avoid letting QP2 updates from 'biasing' the backtest I might do after the 'evaluation' date. For example: Yesterday, (March 28, 1999), I wanted to test my variables on the QP2 data which was available to me Sept 27, 1998. I can't use the current QP2 database and just index references to things like functions of prices and volume and yield and periodic revenue or eps. To the extent that price LEVEL is evaluated, then splits may prevent an accurate calculation. Splits also will make problems relative to volume OR dollars traded. There are variables like book value which are not retained as vectors but are modified with the former value being discarded. Also, how can I evaluate earnings or sales growth or level 6 months after the fact? We all know the that the quarterly and annual stuff is released and updated to QP2 in a very irregular fashion. I can't just reference the quarterly data 2 quarters back and be correct for all the stocks. The only practical way I can backtest what I'm doing is to export a 'mini database' of all the variables I use for each stock for each day I want to use as a possible 'evualation' date (in order to see what the performance was N months hence). Analysis done on that mini database with a custom program is working with data that was available 'that' day, since it was created from the QP2 database ON that day. To find the investment return for 6 months forward, I have to refer to the mini database I created for the 6 months forward date, find the stock in question and capture the return percents for all the months (or weeks or days) I want to explore. Sounds like a large effort, but the data manipulation part is done. I still play with the variables and the weights but that is not a large programming rask.