February saw the market continue its slow trickle up, “climbing-the-wall-of-worry” behavior. Movement in the gaps and anemic intraday ranges were the order of the day, and unfortunately such an environment doesn’t offer intraday systems many opportunities. The system’s equity curve saw most days coming in under +/- 0.50% and many offered no trades at all, leaving the system flat (-0.40%) for the month. Yawn.
It seems this system is not alone, as a fellow algo trader friend of mine joked “Welcome to the $300 club!” after listening to me complain about February’s returns. His system has been seeing anemic returns as well, yielding him pocket change of plus/minus a few hundred dollars a day (vs the thousands per day he was seeing last Fall). I guess misery loves company, lol.
Here’s the system’s equity curve from inception (click to enlarge):
Here’s a histogram of daily returns:
Here’s the matrix of monthly returns to date:
And here’s the systems performance against its benchmarks:
The system underperformed all of its benchmarks during the month of Feb (even Berkshire!), which isn’t all the surprising given that much of the movement was in the gaps. It managed to squeeze out a couple of >1% positive days, but these were offset almost exactly by an equal number of >1% negative days near the end of the month. This is in stark contrast to last Fall, when things were moving +/- 2% per day on average, with a few days coming in at >5%.
As I’ve mentioned in another post, this system tends to make the majority of its money in just a handful of months per year with the others being relatively flat or canceling each other out. That said, two flat months in a row are enough to put one to sleep – hopefully we’ll get the usual Spring/Summer sell-off that will widen intraday ranges again and give this system something to sink its teeth into.
I’ve been running System B using IB’s simulator and it actually outperformed the live system by quite a bit during February (+5.00% vs -0.40%). This could have just been luck, but I’m going to take it live shortly and if it continues to outperform as I scale it in, I’ll most likely replace the current live system with System B.
I’ve continued to experiment with dynamic hedging techniques and have been able to see Calmars above 10:1, but at the expense of much lower absolute returns (as expected). This concerns me a bit; I like to have a bit of a buffer in my returns since live slippage is always worse than simulated slippage. I have a few other ideas want to play with, but the results thus far make me wonder if at the end of the day the extra hassle, cost, and complexity of hedging is really worth the bother. We’ll see…