Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Contest 19: No one in the top 10 has higher returns

My contest submission for #19 is beating all the top 10 leading funds in annual % returns, but my fund is ranked #306. LOL.

I understand completely that Quantopian doesn't care as much about returns as it does all the risk metrics etc, and adjusts scoring in such a way to reflect that, but if you knew what my fund holding it would be rather pathetic that no one is beating it. In fact, I'd never invest in any algorithm that couldn't beat that particular fund.

My trading strategies, over time, will generally show that there is no academic basis for what Quantopian is trying to accomplish. That is, a fund that can outperform the S&P 500 with lower risk. It is not possible and it can never happen (mathematically, philosophically, economically, etc).

Volatility and draw down are the price you pay for higher returns. Alternatively, you can leverage to the max with low, stable returns (i.e., borrowing millions of dollars to buy CDs that pay 1%...lol) but the leverage you are using carries the risk component, so the returns come at a high risk. A one day unanticipated bust up could send you to the cleaners and have you OWING more than you are worth. Further, you could even trade on inside information and "never be wrong" and make stable, high returns, but the day the FBI comes and puts you in cuffs, you realize that in fact those "risk free" returns did in fact carry a high degree of LEGAL risk.

I choose volatility. I choose draw downs. I choose RISK to get HIGHER RETURNS. That commitment to higher risk is the only valid way to achieve higher returns. Period. End of story. The difference between me and other wall street hedge fund charlatans is that I'm honest about the risk I'm taking and don't try to create cute metrics that don't accurately show risk, then sell them as being lower risk, higher return funds.

6 responses

Jeffrey, I am in the same boat...here is a comparison of my "algo" to your algo:

user_name
Jeffrey Robinson
Behavioral Trader

score_contest_rank
306
287

score_contest
72.26987793
74.24942263

rankAnnRet_pt
36
39

annRet_pt
1.838184322
1.797461863

I echo all of your points exactly. I hold a not-so-secret combination of ETFs. Makes you wonder why so much money is being thrown at the hedgies...it must be that they are better at sales than trading/investing.

The annual return for Contest 19 reflects 3 days of out of sample performance, but some wannabe walls street charlatans would have you believe that's being honest about the risk in the portfolio. LOL.

Timothy, this is a discussion about principles, i.e. Quantopian allocation criteria, and it is not a premature yardstick measuring contest, as some charlatans would like to spin it into.

Jeffrey your right. No algo like this exists and never will.

I am really trying to understand what your point is Jeffrey?
Quantopian has a business model, lets say it is awful and flat out wrong, but that is their prerogative. In pursing whatever it is they are, right or wrong, they have provided tools and data streams that many others would not have easily access to for free. Your logic of high risk high reward is pretty sound, I am in the same boat myself, but what is the point of constantly criticizing their judging criteria?

"It is not possible and it can never happen (mathematically, philosophically, economically, etc). "

I'm sorry but this is just laughable. To say that an algorithm cannot possibly beat the S&P without having more risk or leverage is borderline autistic level thinking. The S&P500 by it's nature is a completely "dumb" average, meaning it does not have any logic driving it's value other than the 500 components that it is made of, it can't buy and sell like an algorithm, it's running all the time. By simply adding -any- level of intelligence to buying and selling or risk management techniques you automatically beat the S&P. For instance: during a market crash, the S&P crashes. Any algorithm worth it's salt would be able to mitigate this risk, at least partially, by hedging with cash or even better, a more stable asset like bonds. Holding cash does not increase your leverage or volatility... which according to you shouldn't be possible.

Also, there is nothing illegal about making amazing, mind-reader level returns. Insider trading only applies to people who are insiders: "In the United States, Canada, Australia and Germany, for mandatory reporting purposes, corporate insiders are defined as a company's officers, directors and any beneficial owners of more than 10% of a class of the company's equity securities. Trades made by these types of insiders in the company's own stock, based on material non-public information, are considered fraudulent since the insiders are violating the fiduciary duty that they owe to the shareholders." All regular people, like you and me are not insiders and can act on whatever information we want, no FBI at our doors.