The Twitter Mood Stock Market Study and Hedge Funds

At this moment, hedge funds are pouring over the Twitter (and StockTwits) API and they are testing their models against the market directly. Distinct from their distant cousins at the universities, they are not sharing their results (you will not hear much about this) and the veracity of their methodologies are not subject to the same human shortcomings.

Since writing about the Twitter mood study by Bollen and colleagues from IU a week ago Friday, I have seen a bunch of articles on this topic with many words and many of the same details recounted.

Writers just love to repeat each other and they do it to sell product made by Procter and Gamble ($PG) and Ford ($F) while instead they could simply be linking. 🙂

Meanwhile, no one has made the money point and even though I did not intend to write twice on this, I am left with little choice as I only indirectly alluded to it in my first missive.

Academics, and especially social scientists, participate in a process of discovery which occurs publically only slowly over time. This process involves multiple researchers from multiple universities focusing on similar sets of hypotheses, formulating studies, critiqueing themselves and these peers, refining methodology, rerunning the studies with new methodologies, republishing, tweaking still further until finally reaching some semblance of a conclusion.

The kicker here, of course, is that even while this supposedly objective process occurs in the open, still confirmation and political biases abound.

Meanwhile, in this particular instance, there is another cohort working on the same problem.  They are not publishing and their measure of success is much less inefficient and not at all dependent upon the political whims of academic institutions.

At this moment, hedge funds are pouring over the Twitter (and StockTwits) API and they are testing their models against the market directly.  Distinct from their distant cousins at the universities, they are not sharing their results (you will not hear much about this) and the veracity of their methodologies are not subject to the same human shortcomings.

They ask: Does the model distill alpha?  This is a simple question with an outcome set that is inarguable.

Tonight's Market Shrinkology: Cutting Losses

On Tuesday evenings, I do a program on StockTwits.TV called Market Shrinkology which focuses on the psychological aspects of trading and investing.

Tonight, I will flesh out a recent post on cutting losses quickly and answer questions about this topic.

If you have questions, you can ask in the comments or feel free to ask me directly on the StockTwits stream.

Below is a paper by Shefrin and Statman titled The Disposition to Sell Winners too Early and Ride Losers too Long. Its a critical piece of  behavioral finance research, germane to tonight’s discussion and essential for those who are interested in market participant experience.

Disposition Effect

Tonight’s Market Shrinkology: Cutting Losses

On Tuesday evenings, I do a program on StockTwits.TV called Market Shrinkology which focuses on the psychological aspects of trading and investing.

Tonight, I will flesh out a recent post on cutting losses quickly and answer questions about this topic.

If you have questions, you can ask in the comments or feel free to ask me directly on the StockTwits stream.

Below is a paper by Shefrin and Statman titled The Disposition to Sell Winners too Early and Ride Losers too Long. Its a critical piece of  behavioral finance research, germane to tonight’s discussion and essential for those who are interested in market participant experience.

Disposition Effect

Choice Getaway for October 17: Warwick Valley Winery

If you live in or around New York City and would like to escape for the day then I’ve got your plan.

The Warwick Valley Winery and Distillery is a choice spot that’s not too far from the concrete and this is the perfect day to do it.  The weather forecast predicts sunny and 65 and Autumn is going bizonkers out here.

Theyve a huge lawn with live music from 2 to 5, vendors selling ridiculous lunch fare, an assortment of wines and the best apple and pear ciders ever.

I will be there so ping me if you make it.

Learning To Take Small Losses Quickly (Revisited)

While taking small losses may seem like a simple matter of discipline, there are a number of reasons why this is not the case. To summarize briefly, holding a losing position tends to spur a set of psychological processes which interrupt the rational choice of realizing the loss.

There are a few posts from my old tumblog that continue to get a bunch of hits and, god bless the $GOOG, they look to me like some of the best ones i did. So I figured I would edit them a bit and repost them here on the new blog.  This first one fits beautifully into the StockTwits University Project and might be useful to some readers.

I have had the opportunity to talk with a number of very successful traders over the years and have found that even while their trading styles and asset classes of choice differ, they all tend to share a few common attributes which separate them from less successful traders and which I believe contribute to a significant portion of their profitability.

I have traded equities profitably for a number of years myself and have applied similar rational behaviors to my own routine that have served me well.

One in particular stands out as crucial and so I will begin there.

Every winning trader that I know habitually takes losses quickly before they become outsized.

Continue reading “Learning To Take Small Losses Quickly (Revisited)”

Twitter Mood Predicts the Stock Market (But Don't Bet On It Yet)

since feeling is first who pays any attention to the syntax of things

– ee cummings

By now, there is enough evidence suggesting mood materially affects individual decision making that we can assume this to be the case.  With regard to markets, so much depends upon one’s break even point as prospect theory has shown us and as losses loom larger than gains. Indeed, one can make the case that this is why support and resistance levels seem to work as these tend to be areas where a large number of positions break even.

The leap though from empirical validations of the individual’s behavioral-affective dynamic to making assumptions with regard to the collective market psychology is not so easy.  For one thing, it is much easier to construct a sound scientific method for measuring complex relationships within individuals and for another, measuring the market mood alone is fraught with construct validity problems.

This brings us to a wonderful study published by Bollen, Mao and Zeng in which they measure the correlation between mood states derived from large-scale Twitter feeds and the $DJIA over time.

Instinctively, I love this study because I have observed qualitatively a possible relationship but also because it provides a new potential framework for measuring near real time collective market sentiment that might be less encumbered by indecipherable motivations than current popular measures.  Here is a copy of the study with some concluding comments below:

Twitter Mood Predicts the Stock Market

There is a big difference between discovering a method to measure close to real time collective sentiment and its relationship to market behavior and exploiting it for market gains.  For one thing, the methodology described in the study above, while provocative, will require scrutiny from the scientific community, reproduction and refinement.

But, and perhaps more importantly, the individuals who will attempt to trade off of such data will be subject to not only market structural limitations but their own emotions as well.

Twitter Mood Predicts the Stock Market (But Don’t Bet On It Yet)

since feeling is first who pays any attention to the syntax of things

– ee cummings

By now, there is enough evidence suggesting mood materially affects individual decision making that we can assume this to be the case.  With regard to markets, so much depends upon one’s break even point as prospect theory has shown us and as losses loom larger than gains. Indeed, one can make the case that this is why support and resistance levels seem to work as these tend to be areas where a large number of positions break even.

The leap though from empirical validations of the individual’s behavioral-affective dynamic to making assumptions with regard to the collective market psychology is not so easy.  For one thing, it is much easier to construct a sound scientific method for measuring complex relationships within individuals and for another, measuring the market mood alone is fraught with construct validity problems.

This brings us to a wonderful study published by Bollen, Mao and Zeng in which they measure the correlation between mood states derived from large-scale Twitter feeds and the $DJIA over time.

Instinctively, I love this study because I have observed qualitatively a possible relationship but also because it provides a new potential framework for measuring near real time collective market sentiment that might be less encumbered by indecipherable motivations than current popular measures.  Here is a copy of the study with some concluding comments below:

Twitter Mood Predicts the Stock Market

There is a big difference between discovering a method to measure close to real time collective sentiment and its relationship to market behavior and exploiting it for market gains.  For one thing, the methodology described in the study above, while provocative, will require scrutiny from the scientific community, reproduction and refinement.

But, and perhaps more importantly, the individuals who will attempt to trade off of such data will be subject to not only market structural limitations but their own emotions as well.

Load More Posts