Search
× Search
Menu
  1. Home
  2. Members
  3. archive
Saturday, December 21, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

What can I do with Social Media Sentiment on Securities??

photo

 Kevin Close, Managing Director of Business Development and Research

 Friday, January 16, 2015

Often one of the largest hurdles with new content in the marketplace is understanding how to derive value. I have seen this issue occur across many types of financial firms. We can show plainly there is alpha resident in our content, but that is not enough. The process of mining value from new content is typically more complex and the answer changes depending on how the data is utilized. At SMA we see sentiment in our S-Factors as an additive item within a decision model that can enhance the performance or risk profile of a system. The hurdle for providers of this new content is to clearly illustrate multiple areas of value and applications for sentiment. This will reduce the inherent risk of devoting resources to incorporating a new data set in algorithmic decisions and spur on wider adoption. Recently, we have seen promise in utilizing our S-factors to predict the likelihood of volume increases in a specific security. This makes sense since an increase in positive or negative sentiment for a specific security will likely result in focus and activity in that specific name. Changes of note in these metrics/factors should be a precursor to increased activity in that security. So our initial findings proved positive. When we look at our history of S-Factors and identify conditions when our SV-Score is greater than 2 or S-Delta is above 2 or lower than -2, we can identify significant changes in the underlying security volume. With these S-factor conditions we found increased volume occurred 78% of the time during the event conditions and for two time periods after the event. The volume increase averaged 2.5 times more than the average of that same time period for that security. This illustrates my point that sentiment can provide value in multiple areas. In this case, the fund manager or buy side trader will find this a simple and effective tool to increase the capacity of a trading system or reduce transaction costs through managing their execution algorithms. A sell side trader can set alerts to indicate securities active in social media or to advise clients on opportunities to move inventory. Alpha generation in short term momentum models, option volatility prediction, risk management alerts, short term portfolio selection are some of the initial uses of sentiment. It may be that one use case is ample reason to invest the time to integrate sentiment into your decision model, but this would be short sighted. There is more to be mined and new data is typically where the gold is found. If interested in more detail on our volume study please download our research note found at http://bit.ly/SMA-research .


More links

Print

4 comments on article "What can I do with Social Media Sentiment on Securities??"

photo

 Larry Kase, Financial Analyst and Hedge Fund Principal

 Sunday, January 18, 2015



Responding to the question posed; avoid so called social media sentiment like the black plague unless inclined toward trading the rumor or opinion of the moment from the seat of one's pants.


photo

 Kevin Close, Managing Director of Business Development and Research

 Monday, January 19, 2015



Larry I think that Social media sentiment providers have evolved and improved their process of eliinating spam or bots deriving tweets. We watch the tweeting behavior for an account a few months and if the behavior meets our criteria we then use their comments in our sentiment evaluation. Additionally firms are getting better at determining intention not just sentiment. It is the intention that has greater value in predictive metrics versus reactive sentiment on events. There is allot of focus in this area and we should expect the value of this data to increase.


photo

 Larry Kase, Financial Analyst and Hedge Fund Principal

 Monday, January 19, 2015



Evolved into what is the question. Reading sentiment is highly subjective. Reading intent is an exercise in futility. Who decides what is important or relevant and what is not. Measuring volume of rumors may produce a trade motive. However, measuring tweets, tumbles and other such social media buzz is not different and no more valuable as empirical evidence than rumor mills past. Without doubt the buzz tracking is fun and entertaining but developing predictive metrics is highly improbable when considered in the most favorable light. Respectfully, there is no basis for placing much value upon the aggregate commentary. Rumor and innuendo hardly qualifies as data. I have no problem accepting the stream for what it is. Anything that moves price may present a trade opportunity for those inclined toward rumor or story based speculation. Nothing wrong with doing it but must be recognized and accepted for exactly what it is.

A note regarding sentiment seems appropriate. Generally accepted market sentiment measures are the product of actual surveys and polling processes. There are reports based upon unscientific anecdotal scans such as the nonsense appearing regularly on the CNBC instant opinion tallies. Then there is the matter of using the sentiment reads. Sentiment is often confirms the contemporary experience on a relatively short run basis but tends to be a contrary indicator on a secular basis. Then there is the challenge drawing distinctions between retail level sentiment surveys and those polling asset managers, advisers and publishers. Measuring sentiment is hardly new and remains a highly subjective art. There is no new content. The delivery forms change but there is nothing new under the sun. Best of luck with the pursuit. I admire the adventurous spirit.

One suggestion; read The Crowd by Gustav LeBon while pondering the possibilities for reading intent. It is one of the two most important books ever written for anyone researching sentiment.


photo

 Jim Damschroder, Chief Investment Officer at Gravity Capital Partners

 Saturday, January 24, 2015



there are a couple of firms doing this as a data vendor. i think one is called ravenspack

Please login or register to post comments.

Newsletters

Get up-to-date industry news and the latest algorithmic trading insights when you subscribe to our free newsletter.

Subscribe me

Already a Subscriber?
To manage your current subscription settings, enter your email address.

Go

We value your privacy. We will never rent, sell, or otherwise use your email address for anything other than the services you specifically request. Period.

About our Association

The Algorithmic Trader’s Association, established in 2009, is the world's leading professional organization and resource center for the discussion of algorithmic trading strategy, methods, software, and technical analysis. 

Get in touch

  • Address: Delaware, United States

  • Phone: (888) 668-0571

  • Email: info@atassn.com

Media

Follow Us

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association