Algo trading's new Holy Grail?
Hmm, I wonder if this is what James Simons does?
Trading has a ton of indicators, Bollinger bands, Support, Resistance, MACD, and Fibonacci ratios just to name a few.
But two Professors from Indiana University, Johan Bollen and Huina Mao, along with the University of Manchester's Xiao-Jun Zeng, may have just found the ultimate indicator.
Twitter.
The Twitter calmness index to be exact, and according to this report it is accurate up to 87.6% in predicting the daily ups and downs of the Dow.
Using mood tracking tools such as the Google-Profile of Mood States (GPOMS) and OpinionFinder, and then taking 9.7 million tweets posted by 2.7 million tweeters globally between March and December 2008, then comparing their calmness levels to the DJIA of the same period, the trio has found surprising correlations.
Lagging by just 2 to 6 days, their time series graphs show that the two frequently overlap or at the very least, point in the same direction. "We find an accuracy of 87.6% in predicting the daily up and down changes in the closing values of the Dow Jones Industrial Average" Bollen and company said, totally repeating what I wrote earlier.
They are unsure however how this could have happened. Their findings have offered no explanation for their correlations and well, calmness levels from worldwide tweets predicting the Dow is about as random as anything I've ever heard.
Nevertheless, there’s already some speculation that this would attract a lot of attention from the financial community, and may even become quite influential.
So what's your take on this? Will this be the new Holy Grail of indicators? Will funds be throwing millions at this?
Me? I'm not so sure.
Got me a little curious on behavioral finance though.
of course not.........................
Classic case of correlation vs. causation...
I"ll second DontMakeMeShortYou, though my first impression of reading this post was, "Can they do the same thing with fb, perhaps with more accuracy, than with twitter?"
What does James Simons have to do with this? He is the father of Algo trading.
I'm sorry...does this have some sort of predictive power of which I am not aware? It seems it retroactively looks at Twitter and then compares the number of tweets to the moves in the markets on that date. This isn't ground-breaking work. It is an effort to seek a cause for an already observed effect. Or am I reading this incorrectly? Are they saying that an increase in tweets indicates a move in the markets, and that the markets lag by a few days? Or are they saying that due to large market moves, there are many more tweets than normal?
If it's the latter, who gives a fuck?
Read the paper. It compares the mood of the general public expressed through tweets with the DJIA. It says nothing about "the number of tweets". The general public is much larger than the investing public. In fact, most of the people tweeting used in this study said things like "I feel" or "I am" and then some sort of emotion/mood related statement. The paper compares the collective mood by aggregating the data and comparing it with the DJIA. Your analysis that it is an effor to seek a cause for an already observed effect is off base: the two events are seemingly unrelated. Do you really think that me tweeting something like "I'm upset because I didn't get a hot dog at the baseball game" actually has an effect on the DJIA? It does not.
All it reveals is a strong correlation between the mood of the general public (based on the content of tweets, not the number; also, the only mood that showed any significant relationship was calmness) and the DJIA.
Next time read the paper before you make a comment on it.
Shouldn't professors at Indiana University know that correlation isn't causal? This seems like a completely useless study unless they can figure out some way to look at it other than retroactively. Definitely interesting though.
data mining
Agree that this is certainly correlation and not causation. However, I don't think the idea is totally without merit. There is some very valuable information hiding in the Twitter real-time stream. I personally follow 5-10 excellent traders (both independent and professional) and equity analysts that tweet often about stocks they're following. I've even traded on their recommendations several times and made $1000s. There's gold in them tweets.
Checkout stocktwits.com for an aggregator of Twitter sentiment around individual tickers. Twitter is a data miners wet dream.
@CaptK, which ones do you follow? Always on the lookout for good trade ideas.
Est voluptatibus quo quam voluptatem voluptatem et dignissimos nobis. Corporis est doloremque amet quod eveniet et suscipit. Nam voluptatem accusamus veritatis id eos modi. Sit aut repellat omnis aspernatur magni.
Aliquam ea voluptatem sunt aliquid ut harum totam. Quia laudantium repellendus quos.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...