OK, thanks for the links to the slides. Will watch them this weekend ... didn't know they were available.
Hi guys, based on what GAT wrote: Let's take a simple example. Suppose I had Gold prices from 1st January 1980. To ensure I can measure rolldown I want to be in the second contract, but no further. My roll cycle is GJMQVZ which are the liquid months, all of which are around 2 months, 60 days, apart. To make things simple let's ignore the possibility of having a front month not in the roll cycle to measure rolldown off, and switch 70 days before expiry. In January I'd want to be in April Gold, so I can measure rolldown off February. 70 days before April Gold expires I would then switch to June Gold. 70 days before June Gold expires I would move to August Gold, and so on. However, if I were to take quotes on Jan 08 2018, Feb is the month that seems to be the best for trading. Can someone enlighten me as to why we should be in APR gold?
Hi Rob, I noticed in the presentation (~17:25) you mention that the framework you discuss in the presentation doesn't particularly 'play well' for mean-reversion strategies, or if limited leverage is available. Specifically you suggest making some tweaks to make it work for these type of strategies / conditions. I would be very interested if you could provide some further information / suggestions with regards to these 'tweaks'
'Tweaks' is perhaps an underestimate; certainly the adaptations required would require a blog post to explain properly (I'll add it to my list...) GAT
Thank you for this. Different, probably dumb, question. Again a yes or no would be fine as no doubt your academic duties are keeping you busy. I am currently filtering instruments with the help of the various tiebreakers in your book plus some shortcuts. For example, even before optimising forecast weights, I notice that the returns from some instruments are massively negative in a particular rule class for TF strategies (ranging from -15->-30% over all history). Could I be criticised for backfitting by excluding these at the outset and not letting the optimiser downweight (presumably to zero)?
Yes, it's wrong to delete those 'in advance' and it will inflate your likely realised SR. The optimiser will only downweight to zero if there is significant evidence that they're negative - there may not be. GAT
In that case I am a bit stumped. The problem I am facing is that two instruments in particular (BRL, sugar) have such large -ve annualised forecast returns (anyone else see this?) that they overwhelm the forecast returns of every other instrument when pooled together. So if I pool, I end up with -ve returns in virtually every speed variation except one. This presumably will have the undesirable consequence of concentrating the weight in a single speed (with the +ve return) if I am not mistaken?