Back to Community
The Simplest Algorithm

This "algo" will never win a contest or be chosen for allocation, but it is likely used in some form by a lot of hedge funds and most individual investors, knowingly or not. The key points here are that by being always 90% invested in stocks and bonds equally, keeping 10% cash for emergencies, an investor is getting almost the same returns at half the volatility, and less than .3 beta. There are no "indicators" used, just the most basic axiom in capital allocation theory that an investor should own some % of stocks, bonds and cash.

I am very much interested in what you guys think about this "algo".

Clone Algorithm
67
Loading...
Backtest from to with initial capital
Total Returns
--
Alpha
--
Beta
--
Sharpe
--
Sortino
--
Max Drawdown
--
Benchmark Returns
--
Volatility
--
Returns 1 Month 3 Month 6 Month 12 Month
Alpha 1 Month 3 Month 6 Month 12 Month
Beta 1 Month 3 Month 6 Month 12 Month
Sharpe 1 Month 3 Month 6 Month 12 Month
Sortino 1 Month 3 Month 6 Month 12 Month
Volatility 1 Month 3 Month 6 Month 12 Month
Max Drawdown 1 Month 3 Month 6 Month 12 Month
# Backtest ID: 576154be1c4f7711cdef74d2
There was a runtime error.
17 responses

I can tell you that what I learned through long sample testing of various asset allocation methods during my MSF thesis I was a bit surprised how well a 1/N portfolio performs. There are definitely times when making less assumptions, particularly with correlations, will pay off.

Maybe try adding more asset classes.

Thanks Georges, and that was my main point - complexity is often costly, and as beginner, I am questioning how far I should take my "indicator" development, and how large a "price" I should pay for it, in terms of time spent.

By the way, reducing volatility and drawdown by half is no small feat - a lot of people are paying 2/20 to some hedgie for it.

Anyone else working on a VERY simple strategy which outperforms in terms of vol and drawdown but does not leave an investor behind without participating in equity markets?

Unlike most people, I am not interested in designing a faux high frequency algo (as I suspect Quantopian wants to extract from users coding on the site, and most coders on Q are working on - sadly I think this is a fool's errand), but I am searching for a simple heuristic driven "algo" which can beat the S&P500 (and thereby beat 95% of all hedge funds).

May be Plain vanilla fixed ratio stock-bond portfolio is what you are looking for.
Clone it, backtest to the latest date and let me know the out of sample results.

Thanks Vladimir, the linked thread has a VERY simple example of exactly the same code as the one I used...what I am saying is after you beat the market with something simple, which is what most people use in their retirement funds, and what most hedge funds charge 2/20 for, what is there to gain by adding complexity?

I know Quantopian hopes that someone will sumble upon an HFT algo which will beat the market in any time frame, on any tradeable security, but that is not going to happen, as I said that is a fool's errand.

So the question for discussion remains - how much time/effort/money would and should you invest in coding an "algo" when the simplest algo can beat the market, and it will take you 10 minutes each month to implement, i.e. rebalance?

@ Behavioral Trader -

Minutely trading is not HFT. Also, the "bet" by Q, as I understand it, is that they'll be able to compile a portfolio of many algos that, as a whole, will be more attractive than what other hedge funds can offer. The idea, as I've heard it articulated, is that there is a market for "pure alpha" actively managed funds, and Q is looking to play there. They won't be building a business on competing with Vanguard, etc. with the kind of asset allocation approach you posted above. For an individual investor, it makes sense, but my understanding is that it won't be attractive to institutional investors, who think in terms of buying beta and alpha separately, hence Q's preference for beta near zero.

"compile a portfolio of many algos that, as a whole, will be more attractive than what other hedge funds can offer"

I need to disagree on a couple of points.

  1. They are not in the business of reselling intel (licensing crowdsourced algos)
  2. They are in the asset gathering business
  3. They must compete with Vanguard and everyone else who gathers assets
  4. They must beat the market and not "hedge funds", as most hedge funds do not beat the market to begin with
  5. If using the "Simplest Algo" can beat the market in terms of vol and drawdown, then why spend time and money in pursiuit of other complicated (costly) algos?
  6. Seems to me that anything other than naiive simple diversification is a fool's errand Quantopian is pursuiing...
  7. Sorry to be negative on the entire premise...

Grant makes a great point about differentiating between returns from beta and returns from alpha.

Ray Dalio on the "basic building blocks" of portfolio construction. (return = cash + beta + alpha)

http://orcamgroup.com/wp-content/uploads/2012/10/pmpt-engineering-targeted-returns-and-risks.pdf

I'm not in the industry, but the idea of beating the market (S&P 500?) would seem to be irrelevant here (and one has to define, as a goal, what "beating" means). In my mind, the idea is to find inefficiencies (money left on the table) on whatever time scale, that can be exploited to generate a return on capital. It is an active trading approach, for which investors will pay a premium (versus beta, which is a commodity), if justified by a track record. This could be dope smoking, for all I know, but I doubt it. There is such a thing as a hedge fund industry, and the numbers are big (see http://www.barclayhedge.com/research/indices/ghs/mum/HF_Money_Under_Management.html). In fact, if Q achieves their goal of $10B AUM, they will only be 0.37% of the current overall hedge fund AUM. It is a pretty big pie. I don't have the patience to count to 1,000, let alone 2.7 trillion.

Grant, Georges

Why this Simple Zero Beta Plain Vanila may not be considered as "Pure Alpha"?

Clone Algorithm
35
Loading...
Backtest from to with initial capital
Total Returns
--
Alpha
--
Beta
--
Sharpe
--
Sortino
--
Max Drawdown
--
Benchmark Returns
--
Volatility
--
Returns 1 Month 3 Month 6 Month 12 Month
Alpha 1 Month 3 Month 6 Month 12 Month
Beta 1 Month 3 Month 6 Month 12 Month
Sharpe 1 Month 3 Month 6 Month 12 Month
Sortino 1 Month 3 Month 6 Month 12 Month
Volatility 1 Month 3 Month 6 Month 12 Month
Max Drawdown 1 Month 3 Month 6 Month 12 Month
# Backtest ID: 57670029cf000c0f890b7605
There was a runtime error.

It's all such a stupid bullshit industry. Even the stupid phrases alpha and beta are thoroughly irritating.

And what is the market anyway? If it is the S&P 500 then is is a market cap weighted S &P 500? If so why? Smaller more nimble institutions can do better with equal weighting. Is equal weighting also beta? Is smart beta smart? Or is it really alpha? Or does it matter?

Look at hedge funds: do you really think they have anything SO special? I don't think so.

There is no such thing as beta because there is no such thing as THE market. There are many different markets. There are many different sectors. There are many different national markets - don't be narrow minded and assume the US is THE market. It is not.

And what about commodities? And currencies? And real property ?

Returns are about asset allocation, largely.

And market timing.

Beats me. I guess it would.

Vladimir,

I would consider it beta because of its long only / buy and hold / passive management nature. You are simply taking market risk in both consumer staples and US government bonds.

I believe most of the alpha of these returns are coming from the low volatility anomaly but I could be wrong.

Honestly, if you can stand the volatility and have the capital just look at a simple trend following approach on a balanced portfolio of around 25 futures markets.....stock indices, currencies, energy, grains, software. Use modest leverage.

Probably my favourite is the simple Bollinger band breakout with a holding period of a few months and a profit taking mechanism. I achieved 16% cagr between 2006 and 2012 when I was blown out of the water by MFG. Max DD around 40%.

Nothing clever about it and used a very standard volatility based money management routine .

What is it? Well, probably simply asset allocation with market timing.

Plenty of track records out there you can look at.

Too simple for most of you guys perhaps. And too volatility.

Gosh....sorry about the typos.....it's autocomplete at work.

Thanks guys....just to clarify my note...

  1. Market - since the default benchmark in all backtesting is S&P500, then S&P500 it is
  2. To beat the market - this is up to debate, since I would argue that achieving close to the market return while reducing vol and drawdowns can be defined as beating the market (see the stats for my original "algo" above).
  3. Active versus "naiive" investing - this is just it - if by utilizing naiive 1/n diversification models one can achieve the same or better returns than the market, and cut vol and drawdown by half, then the ANY step in the "active" direction must be justifiable in terms of effort and risk.
  4. Horse beat to death.