OmniVest Forum OmniVest Forum
forums calendars search
today this week
 
register logon control panel Forum Rules
You are currently browsing as a guest.
You should logon to access more features
A Self-Moderated Community - ALL MEMBERS, PLEASE READ!
Vote for Members who contribute the most to your trading, and help us moderate content within the Forums.


Only Forum Moderators, Administrators, and the owner of this message may delete it.

 
Jim Dean

Elite
10002525
Posts: 1059

Joined: 10/11/2012
Location: L'ville, GA

User Profile
 
Subject : RE: Portfolios based on FUNDAMENTAL analysis
Posted : 3/26/2014 7:06 AM
Post #29591 - In reply to #29590

I have been a proponent of this for years, with the caveat that we need back testable fundamental data to work with, or it's all potentially smoke and mirrors. There are two issues:

1. N does not want to have to purchase a huge truckfull of historic fund data. However, as I understand it, they have daily data going back to around 2005-6 or so, whenever they first started doing OData/OScan. In my mind that's a decent basket to work with - about nine years. There are places they could "scrape" historical fundamentals back to 1990 or so if they chose - tiny cost but some considerable hassle.

2. The second issue has been that of bringing OScan and backtesting up to speed to dynamically handle the historic data. My guess is that the CPU overhead as well as the data storage involved (maybe 50 data items per symbol per day versus just five OHLCV) has been the sticking point.

Advantages and solutions:

A. Amelioration to massive data issue is to store the data once a week vs once a day.

B. Solution to massive CPU hit is to process the Fund Scan rules on the cloud OVest servers.

C. To make it work on local machines (for ET testing), upload the completed daily symbol lists created on the OV server to the local PC for use by OT. This method works equally as well for fund scans as it does for tech scans. Essentially eliminates both the local CPU speed and storage problems in one fell swoop.

D. Bluesky suggestion: Benefits are enormous if a hybrid techno-fundamental engine is created - ie being able to apply formulae (like MA crossovers, std deviation, etc) to the raw fund dataseries, rather than just working with discrete data items. There are a couple of other vendors that do this - it's cool! It would however call for the ability to "plot" the hybrid fund "indicator" so the person can visualize their logic. If all the discrete data stays on the server, that means OV would need a "Filter Lab" with data and charts, to "paint" the OScript formulae results on those charts. That probably would be too much to add to the pot near term, but it's worth keeping in mind for when OV is a moneymaking engine for Nirvana.

Summary - I think implementation of ABC above would be sensible, helpful, and might attract some "motif" investors. I suppose it could be done fairly near term, especially in light if the current work being done on dynamic lists - this would be a natural adjunct. However, compared to many other things currently on the table, I'd rate the relative value of this to be "moderate".
Deleting message 29591 : RE: Portfolios based on FUNDAMENTAL analysis


Nirvana Systems
For any problems or issues please contact our Webmaster at webmaster@nirvsys.com.