Quantopian – Learning the Lessons

I hope Quantopian rises again in its new guise of fund manager – which looks to be one of the ways they are repositioning themselves.

Amateur punters using a free back testing engine can not have been the ideal client base.

I have been back testing quantitative trading systems since 2003 and so I probably know a little about the topic.

I traded a wide range of world futures for some years, but the glory days of inflation driven commodity trends were long gone by the time I entered the game.

I wrote a book on investment systems for Exchange Traded Funds. Which, much to my surprise, still sells a few copies 12 years after its publication.

I have lost track of the meandering history of Quantopian and its ever changing course. But it can not be an ideal business plan to provide a free platform to amateur quants, in the hope that they will produce quantitative systems robust enough for use by hedge funds of the caliber of Steve Cohen’s.

Data science is all very well. But unfortunately it does not allow us to predict the future of financial markets. You may believe that day will come and that we live in a deterministic universe. Or not. You may believe that if we can only understand and track all the variables which go towards determining price, we will unlock the doors of Midas’ treasure house. Good luck with that.

Certainly, many of Quantopian’s amateur punters (who seemed to be new to financial markets) believed that Eldorado was in sight. They flocked from other fields where some of them had presumably thrived and they believed that they could “solve” the markets. So apparently did the management at Quantopian. So perhaps did their one time backer, Mr Cohen and Point 72.

Clearly none of them had studied Voltaire. The cautionary tale of Candide might have saved them much effort and quite a lot of money.

My 17 years of coding quantitative trading systems should have taught me that it is largely a waste of time. Or rather, being clever is largely a waste of time. And yet I continue to fool around as a silly intellectual exercise.

Professor Brainstawm and his many doppelgangers in and around the Quantopian ecosphere perhaps felt differently. Perhaps still do.

Let me clarify my position.

I have every belief in quantitative investment. One of the first stock market indices was the Dow Jones Industrial Index, first published in the late 19th Century in the US.

A broad based index captures the value which the investing public attaches to major businesses, and by definition therefore it also reflects a view of the prospects of an economy as a whole.

Stock markets are still driven by fear and greed, a reflection of their all too human participants. But by and large they have provided an almost perfect quantitative approach to investing.

These days, you don’t even have to track an index for yourself – for some 50 years, pioneers in the field have provided index tracking funds which, by and large, is all most of us should use for our investing.

The best stock indices simply invest in the top X companies in a given economy or economy. Weighted by market capitalization.

The vast majority of stocks ever listed have crashed and burnt over the years. Or been absorbed by competitors in the ferocious and cannibalistic world of mergers and and takeovers.

A stock index follows this trend – as new companies rise and prosper, they become included in an index once they become big enough. As they inevitably decline they are removed from their seat of glory. An index fund is thus a trend follower – it remains invested in strong growing companies and exits the weaklings.

If the world economy continues to thrive as it has done for hundreds of years, the stock index will continue to rise and investors will find it a good place to invest money.

All an investor has to do is to go with the flow. How very Zen; what a perfect reflection of the Tao.

The problem arises because as smart apes we always think we can do better. History has failed to teach us that we can not.

The correct use of quantitative investing techniques is to provide a simple asset allocation method and there can be a certain flexibility here. But the overriding rule, the key objective, should be to follow the world economy and not to try to isolate winners or losers. Let alone stock pick – god forbid.

Wide diversification over the entire global economy should be the object for an investment manager. He should not be seeking to shoot the lights out. He should have learnt his lesson long ago: the fate of Imperial Russia, Weimar Germany and tin pot banana republics in South America should have taught him that concentrated bets would eventually lead to disaster.

Leverage is lunacy, but many failed to heed the warning provided by LTCM. And market neutral? I’ll do the jokes. There is no such thing.

There is a road to riches provided by quantitative trading but few of us will be able to tread those paths for one reason or another.

Such routes usually entail inside information or some other structural advantage unavailable to the mere mortal.

High Frequency Trading appears to be a prime example – but there are those far better to tell this story than I. It appears such traders have information about order flow and can thus deal ahead of clients to their own advantage. Doubtless they have many other tricks up their sleeve but they do not appear to be playing on a field which could be described as level.

Quantitatively speaking, Initial Public Offerings lead by the big US houses have been underpriced historically and those able to get stock at the issue price have reaped handsome gains. Buying in the aftermarket provides a very different and less favorable story.

Buying secondary offerings at a large discount, or junk bonds with the benefit of the syndicate commission are other routes reportedly leading to profit. And no doubt occasional large losses.

With relevant data and a good back testing engine you may be able to spot some such structural advantage for yourself. You may even be able to profit from it.

My friend Alan Coppola at Hot Chili Analytics provides a worthy successor to Quantopian, using basically the same software. But you will need to provide your own data and pay a modest monthly fee. No free lunch as they say. And you can use the software to place trades with Interactive Brokers.

But if you do decide to continue your programming journey, find yourself a structural market fault line of some sort. Don’t expect to find Eldorado using the dull, boring compass used by so many for this journey.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s