Read time: 3 min

And Another Thing…

I flew from Tokyo to London earlier this week, it reminded me that I actually flew on the first ever non-stop flight between the two cities back in the late 1980s. I was, for some obscure reason that I cannot recall, travelling in first class, and I was the only passenger in that cabin who wasn’t; a/ on the board of directors of British Caledonian (as it was then); or b/ a “Sir” (a Knight of the Realm to non-Brits). I only mention this because I am about to venture into territory unknown and, as I did back then on the plane, I feel a little out of my depth!

I sometimes think we can get a little too analytical in this business, although I suppose the desire to enumerate everything – to give it a value – is becoming a fact of life in all areas of the modern world. I am interested specifically in the release, reported on in this issue of Squawkbox of a human emotion analysis tool. Now I am sure it is an excellent product, but I really struggle to understand how you model human emotion. Actually, I think I understand why you want to do that, but I would have thought that was already handled by the relative technical analysis tools out there.

What I am arguing is that human traders reflect their emotions, their opinions is probably a better word, in their positions, therefore is this new tool merely telling people whether the market is long or short? I understand the interrogative aspects of the service, by asking questions and getting answers from the Twitter-sphere and other social media platforms, but I am not convinced by the direct value. I mean, has anyone read some of the stuff on Facebook, Twitter or other social media outlets? There are some proper, bona-fide lunatics out there with the most bizarre opinions (most probably still trading spot!)

It strikes me that any analysis of this type of data will have to have a very careful sample set to make sure that the “bonkers of Bradford” (with apologies to that fine city, this is just a random example) are not included. As I noted earlier, this is more a (tongue in cheek) cry for help in understanding how this is going to work than actual analysis of the service, but I wonder if we aren’t making things too complicated.

I have written previously of the Twitter-following hedge fund and how if I had the chance I would bombard the Twitter-sphere with bizarre opinions to confuse it. I still see this as something of a gimmick, whereas the new product seems to be an attempt to analyse market sentiment. That is to be welcomed I suppose, but it does strike me there are easier ways than listening to some of the lunatics that populate social media sites.

In FX terms, I have always thought the sentiment and human emotion about the markets and news events is reflected in market positioning and the order book. One told you that people had sold, and the other told you where people were going to sell (or buy of course). The new trade repositories will also tell somebody how the market is positioned – surely that is the best analysis of market sentiment?

Probably my biggest concern is that the analysis can be skewed by outsiders. As an example I would point to certain high profile hedge fund managers of the 1980s and 1990s. It was amazing how often a certain manager was quoted in the Financial Times or other major media outlets expounded the benefits of a certain position. Were we really to believe that the manager in question didn’t have the position on board and wasn’t trying to squeeze an extra percentage point out of it? More often than not, after going public, there was a move in the direction called by the manager, and then the market reversed. Now far be it for me to question anyone’s tactics, but I always had a shrewd idea who it was behind the reversal.

So by all means try these tools, and I am sure they can add value in some ways. But I would also argue the case for treading carefully when dealing with public opinions – they are public for a reason.

Profit & Loss

Share This

Share on facebook
Share on google
Share on twitter
Share on linkedin
Share on reddit

Related Posts in