New Year 2020 quant resolutions

January. Cold. Rain. There’s little to be joyful about this month. However, amidst the gloom, there is that shard of light which crosses over into the future. In particular, the New Year gives you an opportunity to start again, that imaginary line of sand, separating you from the past 12 months. It’s that time when the gym suddenly swarms, diets become en vogue and so on, all fruits of New Year’s resolutions. However, resolutions can also encompass your work, as opposed personal ones. I recently tweeted and posted a short LinkedIn comment on my New Year’s quant resolutions, which got a lot of interest. Hence, I thought I’d elaborate on that (I won’t bore you with my gym based resolutions, which are probably more and exercise in futility!)


Resolution 1: Teach more Python (and learn more Python libraries)

I’ve presented my research work many times over the years, as well teaching at workshops. Over the past year, I also started at Queen Mary University of London teaching a Python for finance course. I also taught the same course to practitioners in house at financial firms. I very much enjoyed the experience, and hope to do more lecturing in 2020. On that note (with a subtle plug!) if you’d like me to teach my Python for finance workshop at your firm let me know (maybe it’s your New Years resolution to learn Python or to improve your Python coding)! Over the past year, I’ve learnt many new Python libraries, ranging from Dask for computation to many NLP libraries. I hope to continue that process this year, where I’m planning to get more stuck into libraries like TensorFlow.


Resolution 2: Learn Julia and q

Julia is a newer language which promises the ease of use of Python with more speed. It’s still early days and it doesn’t yet have all the supporting libraries which Python has. However, the syntax looks quite cool, and not too dissimilar to Python, so I’m keen to learn it (even if I don’t think I’ll necessarily move my stack to it). Another language I’d like to learn is q which is use in the kdb, a database which is ideal to work with high frequency time series. Nearly all my work involves time series. It is a very different language to something like Python, so it will be tougher to learn, but I think it’s worth it. It’s also not open source, which is worth bearing in mind. I’ve already done basic q code, but keen to learn much more. Plus you can call it from Python using qPython, so I should be able to integrate it into my workflow. I’ll be reading the new book on kdb+/q by Jan Novonty, Paul Bilokon, Aris Galiotos and Frederic Deleze to learn it!


Resolution 3: Create more macro alternative datasets

I’ve been immersed in alternative data for many years doing projects for firms including Bloomberg and RavenPack. Alexander Denev and I are also coauthoring The Book of Alternative Data which will be published on Wiley in a few months. One point that is quite apparent, is there’s much more alt data for equities than macro assets. Cuemacro already distributes a dataset for Fed communications which does NLP on what the Fed says. I’m keen to create more alt datasets for macro this year to distribute to clients.


Resolution 4: Create more intraday FX trading strategies

I’ve developed many strategies over the years for FX, including many daily strategies. I’ve also made many intraday FX strategies. The capacity of intraday strategies tend to be lower, but the Sharpe of them tend do better. Trying to see if I can find some inspiration to develop some newer ones for intraday FX. I already have a few prototypes and ideas for more. Fingers crossed, I can put the final touches on some of these strategies in 2020!


Also, as ever, one of my resolutions, is to find the best burger this year. I’ll keep you posted on that! I wish you all the success on your New Years resolutions.