mechanical systems


Screen Shot 2015-11-18 at 4.55.47 PM

Just putting on trial a new stock + options strategy. Have been looking for the missing piece on this basic strategy for a while now. By chance, just heard something on the internet that clicked, and with a couple of simple modifications, I had the missing piece to make my strategy complete.

Anyway, going to run it for a couple of months and test whether the reality matches the theory. The market has a horrible way of trashing theoretically beautiful strategies.

nfl_cheerleaders_02

Snoopy-Typing-Away-1-CVV14J0D95-1024x768

IN THE SEARCH FOR THE NEW and different indexes that will power a new and different ETF, back-testing plays a critical role. Index providers, including S&P, Dow Jones, MSCI, Russell, Zacks and others can index just about anything. You want to rank the companies in the S&P 500 by earnings growth, then take the 50 top firms and weight them equally? Weight them by market capitalization? Go long the top 50 and short the bottom 50? They can build it. And then they back-test it. Indexes that look good in hindsight have a shot to become ETFs. Those that don’t, don’t.

Given our quantitative roots, we are sympathetic to the fact that backtests are often used as an input into making investment decisions. But past returns, as we all know, do not predict the future. And we think backtested results may be particularly problematic today. Very little fundamental data for US equities extends back more than 30 years, but the last 30 years were a period generally accompanied by two related phenomena: increasingly easy monetary policy and falling interest rates. In particular, the wave of liquidity and stimulus provided in the wake of the Tech Bubble coincided with unprecedented levels of credit expansion, rising asset correlations and record earnings volatility.

Snoopy-Typing-Away-1-CVV14J0D95-1024x768

Some of the world’s biggest quant hedge funds have suffered steep losses in the past two weeks following the sell-off in global bond markets.

So-called “CTAs”, which use computer models to automatically spot and ride market trends, were caught out as investors anticipated an end to the Federal Reserve’s measures to stimulate the US economy, triggering a global rout in fixed income investments.

Which is the nature of a mechanical system. It is both a strength and a weakness. A strength in that it takes trades to stated criteria without emotion, which, if the code is good, is profitable on aggregate. A weakness if the code is bad, or, the market changes significantly, thereby evolving past to tested for parameters.

Snoopy-Typing-Away-1-CVV14J0D95-1024x768

In the longer term portfolios that I run I make use exclusively of ETF’s. With an ETF you can gain a diversified exposure to a market segment. The diversification is important as following individual companies is fraught with risk as their financial statements are often hideously distorted and even then the ‘market’ sometimes runs with the ones that you would discard.

In the larger ETF’s, should a company [stock] go bust, sure the ETF takes the hit, but that stock is replaced and life goes on, the dip becomes a buying opportunity. The risk of selection is removed and leaves you only with market timing risk.

Market timing risk is a lot easier to manage than market timing risk and stock quality risk. It lets you focus on the big trends that play out over years rather than the short term wiggles that incorporate earnings risk, legal risk and a myriad of other risks.

Of course the market ETF, say SPY, can be broken down into sector ETF’s which will allow the same market, viz, SPY to be outperformed by the market composed of sectors. I haven’t actually implemented this idea yet, I’ve just been thinking about it after watching a bit of a stoush on CNBC the other day with regard to just this question.

In the discussion healthcare and utilities were identified as outperforming, while materials was identified as one of the under-performers.

All that is required is a methodology for balancing allocations, which is another way of saying ‘market timing’. If I’m correct, then a market composed of individual ETF sectors should over time outperform exactly the same market composed of all the same sectors.

I’ll set up my little experiment and trade it over the next four years and see how the theory stacks up in reality.

The inspiration behind the ppt?

From this post.

My trading partner and I have been working on a pattern recognition algorithm as a side project for over a year now. We’ve been chipping away at the various problems and bugs associated with new software, but we’re definitely getting closer to completion. For those who haven’t read about it here’s an over simplified abstraction of what the software does from a previous post:

To visualize how the software works, imagine drawing a line graph of every single trading day (intraday) of the Dow on a separate sheet of paper. Now take the humongous stack of paper and press the it up against a sunny window like a kid with tracing paper. Even if the sun could shine through the stack of paper, you’d still have an unrecognizable blob, right? Wrong. Our software can intelligently processes HUGE amounts of data and reveal the true nature behind the noise — even if the nature of the data is in fact actual random noise (in which case it will generate no patterns, smart huh?).

Except for the fact that truly random noise, as generated by a Monte Carlo engine, will generate ‘patterns’ which rather makes a nonsense, and a potentially dangerous error of the entire concept. The two key concepts here are [i] independence and [ii] path dependency. True ‘independence’ or actual random noise, will generate ‘patterns’ but possess no ‘prediction’ qualities other than the odd random one generated via ‘luck’. ‘Path dependence on the other hand can account for the sometimes bizarre ‘trends’ that develop in the markets against all rational analysis.

I’ve quietly been working on my adaptive system for a while now, the results are just starting to show through in the last month or so. Previous incarnations have had periods of outperformance, then problems. Currently, those problems have been ironed out.

What is the difference between a mechanical backtested system and a forward tested adaptive system?

Essntially the amount of data that one can access immediately. A mechanical backtested system can backtest as much data as one can get hold of. Woodshedder has this historical period:

The data runs from 1990-Dec 2004. This is a very discrete period of market history. Does it adequately represent the 1969-1975 period? How about 1929-1937? You take the point.

Adaptive, or future testing, can only test 1 day at a time, in real time, thus the data is unoptimised, it is, exactly what it is. You succeed or fail, there is no inbetween. Of course due to this variable, there can be no lovely data generated predicting expected returns based on historical data.

From marketsci

I’ll take a month of real-time trading over a 10-year backtest any day of the week.

That may be a surprise coming from someone whose blog almost wholly consists of backtests, but at the end of the day I know (and so should you) that NONE of this really matters. The only thing that matters is what you do in real-time, audited.

That’s why I don’t release backtests for our own proprietary strategies and I raise a wary eyebrow anytime I see a sexy backtest bandied about.

It’s not that I don’t trust the backtester. There’s a whole universe of good folks whose workmanship I respect very much, but at the end of the day ‘the best laid schemes of mice and men often go awry’.

The table above lists the evil little demons that lead our backtests astray. None of these are new…just a reminder of what we already know (but sometimes forget).

There are things we should be able to control…the hard-and-fast, black-and-white “math” of backtesting.

Have we accurately modeled the trading environment including transaction costs, slippage, realistic quotes, and survivorship bias? Small mistakes here compounded = hugely inaccurate results.

Have we built a mathematically-sound model? You would be shocked to know how often I test published strategies only to find that the results rubbish (reader beware).

And there are things we can try to control but never totally will…the far more fuzzy “art” of backtesting.

We have to cope with curve-fitting and other biases (read more from CXO), markets that are constantly evolving at the most fundamental level (an example) and markets that inevitably go through abrupt “abnormal” periods where nothing works the way it’s supposed to (another example).

The only way – the ONLY WAY – to judge how well a trader has responded to this myriad of challenges is real-time audited results.

It’s easy to churn out fancy charts showing what has worked in the past. It’s a very different thing to put yourself on the line each and every day.

Done right (independently-audited without cherry-picking) there are no mulligans. Your moments of glory and defeat, of brilliance and stupidity, are laid bare.

We the investing community get way too excited about sexy backtests and make way too half-hearted a demand for the real-time.

In my mind, in your mind, in all of our minds, 1-month of real-time audited trading should mean more than 10-years of backtesting any day of the week.

My month of forward testing:

# of trades…………………4
# of winners……………….3
# of losers…………………1
Total trades……………….4

Largest % winner………..+50%
Largest loser……………..[-10%]
Net profit [loss]………….+18.75%

So that’s an 18.75% return since 14 July or three weeks. If I annualise this then we have a 325% annual return [always looks impressive] I shall contine to trade this live certainly over the next month or so. The % are due to the leverage of using Options rather than common stock.

Of course this return includes the internet blow-up, which would have eliminated the 10% loss, into a profit, which would have taken me north of 20% for the period.

Woodshedder has a graph depicting what a 48% compounded annual return looks like:

Imagine if you will 325% compounded annually.

Enjoy the free updates [testing in real time] while they continue, as, should the results continue as they are at the moment I’ll be selling the signals direct to some smaller Hedge Funds.

« Previous PageNext Page »