I remember cassette tapes. The clunking sound when you closed your Walkman, the art of fast forwarding just enough to catch the next track, but not too far to miss half of it. I remember being in awe when we got our first CD player at home, and hearing it for the first time. It was a Police record (excuse the pun), Zenyatta Mondatta. Until today, I didn’t know precisely why it was called Zenyatta Mondatta (an explanation is here). The first track of the album is Don’t Stand So Close To Me. The track starts quietly and slowly. Then the sound builds up. It was like a ritual loading up the CD player, looking at the album sleeve and pressing play.
I’m currently listening to the same track, whilst writing this. Although this time, I’m listening to it on iTunes. I suspect in the age of streaming such a track would never have been written. Today, the emphasis is on getting to the chorus quickly, before the listener impatiently skips to the next song on iTunes, rather than having a slow build up. I can’t really remember the last time I used a CD player. My stacks and stacks of CDs sit somewhat idle filled with dust, a memory of the past. Whilst streaming music is super convenient, and I listen to far more music as result, I do sometimes bemoan the demise of CDs and records. There was a certain wonder about going to a record shop, and browsing through the music. Vinyl has had something of comeback in recent years, but I doubt we’ll see the same for CDs.
The funny thing is that the notion of using physical media to hear music is a relatively new idea from the twentieth century. Obviously, up until that point, all music was live. The closest thing we had to “physical media” was sheet music. Today, the biggest musicians make the most money from what: live music. Despite all the technology and ease of streaming, we still want to see rock stars live (myself included). We haven’t forgotten the past.
In finance, there’s all this talk of artificial intelligence and machine learning and its impact. It’s undeniable that it is having an impact and will certainly do so. However, I do think it’s mistaken to think that we suddenly need to “forget” everything about what we know about finance. Instead, tools like machine learning should be seen as complimentary to domain knowledge. Domain knowledge allows us to leverage quantitative tools more effectively.
Indeed, I’d argue that domain knowledge about the markets and understanding is crucial to developing successful trading strategies, and helps us to leverage the most from concepts like machine learning and alternative data in finance. Admittedly, many techniques from machine learning aren’t really “new” and the same might be said for the search for unusual datasets. Markets still need traders, its just that those traders are increasingly coding and using computers to automate their trading. Banks are still hiring folks to trade the markets but they are looking for different skills, more skewed towards tech.
I’ll still be listening to iTunes, but it doesn’t mean I don’t want to see U2 playing live (or indeed the Police, who I missed on their last tour… maybe next time I suppose). The same goes for trading, just because I code in Python and use quant analysis to develop trading strategies and analytics, it doesn’t mean I want to stop reading Bloomberg News, scouring my Twitter feed and talking to my friends in the market to try to understand and learn more about what’s going on. Each part complements the other.