Maximum Adverse Excursion

I’ve reached the point with the SPY VIX trading system where I need to analyze maximum adverse excursion (MAE).

MAE is a technique published by John Sweeney in 1997 (John Wiley & Sons, Inc.) that measures the farthest a trade ever goes against you. For long (short) trades, this is calculated with the difference between entry price and the lowest low (highest high) price ever seen during the life of the trade.

MAE is very important from a risk management perspective to help determine potential benefit of a protective stop.  A scatter plot of maximum drawdown (MDD) for each backtested trade by final trade profit/loss should be studied.  If there is an MDD level beyond which few trades ever end up profitable then it may be useful to keep a protective stop at that level.  This may lessen some extreme losers while preserving most of the winners.

The difference between MAE and MDD is that MAE pertains to each individual trade whereas MDD may pertain to an individual trade or a series of trades made by a system.

From a risk management perspective, MAE and MDD both address trader tolerance.  Both measure “the farthest X ever goes against you.”  Consider the broad stock market as an example.  Over the last several decades, the broad market has returned roughly 6% per year.  Every now and then, the market crashes and loses 40-60% (or more!) of its value. The MDD is essential to understand because that moment of worst loss is when I will be a nervous wreck, unable to sleep at night, or contemplating a jump out of the nearest 10-story window.  If I can’t tolerate the MDD (or MAE) then at best, I won’t be able to reap the benefits of a profitable trade[ing system].  At worst, I may end up suffering catastrophic loss.

In my next post, I will begin to analyze MAE for the SPY VIX trading system.

Highlights of System Development

Today I want to review a couple major points that have been covered in my last two blog posts.

Trading represents a dream because I am able to work for myself from the comfort of my own home on my own schedule.  As hard as it is in this economy to get any job, to get a gig like this will demand out of me something the vast majority of others are not putting forth.  This just makes sense because otherwise, everyone would be doing it!  For starters, I must be willing to put in extensive legwork at every turn whether or not it ends up being required.

Trading as a business can never coexist with laziness.

The trading landscape for independents is like a hot desert littered with mirage.  I am constantly bombarded with alluring returns and trading strategies that sound good.  Greed pushes me to gamble with speculative trades on a whim.  Wishful thinking accelerates the process.  Other cognitive heuristics also combine to stack the deck against me.

System development must be extremely thorough to flush out any oversights that may lead to false profit expectations.  One could say the whole point of system development is to highlight the arbitrary and determine it to be lucky or good:  discard the former and implement the latter.

This was the idea behind realizing the need to backtest 3-, 4-, 6-, and 7-day trades before moving forward with a 5-day SPY VIX trading system.  Failure to perform these additional steps could leave me unknowingly chasing an illusion.  When lucky, traders will eventually find out by losing precious capital and often going bust.  When good, traders are able to stay in the game.

All trading recommendations, whether technical strategies or stock screens, are due only one response:  backtesting within the confines of a complete system development process to assess how likely they are to perform in the future.  This is a labor-intensive process where many false leads and dead ends potentially await.

Laziness Dissected

Back in Part 1, I described the SPY VIX trading system as having three variables.  The system trades when VIX is y% extended from its z-SMA with an x-bar stop.

Full disclosure:  I don’t want to consider the x-bar stop as a system variable.  So the trade lasts five days.  That seems routine and unlike a technical indicator I employ where I set the critical value to be tested.  Why bother checking x for validity? Besides, I have already entertained y-values from 5 to 25 (five total) and z-values from 6 to 15 (10 total).  Allowing x to vary from three to seven, for example, would increase the number of potential systems from 50 to 300.  Furthermore, keeping one variable constant would now leave two to vary.  This suggests employment of college Calculus rather than high school Algebra in order to study performance.

It doesn’t take much introspection to discover that this bellyaching is simply a matter of laziness.  At some level, most people tend toward laziness.  This may be the primary reason why most new traders fail in the first 1-5 years.  They think this is an undiscovered country where hard work doesn’t exist.  They are eager and willing to pay hundreds to thousands of dollars on trading books, investment webinars, or educational programs that promise consistent profits in 10 minutes per night or a couple hours per month.  They are fertile ground for scammers and con artists, alike.

I have written this before and I’ll write it again:  no free lunch is available in the world of trading and investment.  I’m up against institutions with billions of dollars at their disposal and highly-educated quant teams with thousands of advanced degrees between them all focused exclusively on gaining the slightest bit of Edge.  The easy way out is no longer a viable option.  Screw the ego-laden label of “busywork…” when it comes to labor-intensive, painstaking tasks, I need to shut my eyes, puff out my chest, and rush headlong into the pile so I can get the job done.

The x-bar stop:  in focus with my next post.

Motion to Dismiss System Development?

I now have the subjective function from http://www.optionfanatic.com/2012/10/12/the-subjective-function-part-6/.  Before I move on and apply this to the SPY VIX trading system, I want to briefly discuss one other point that threatens to derail system development.

In my last post (see above), I mentioned the following with regard to high-profit, infrequently-trading (i.e. “surgical”) systems:

> Ideally, I would like to take advantage of the time one such system is not in the market
> by trading with another surgical system that is in the market. I am therefore in the
> market more frequently with systems that deliver concentrated profits.

When trading with surgical systems, the only way to achieve a reasonable percent return on the entire portfolio in a risk-controlled manner is to combine them.

Statistically speaking, this is the employment of non-correlated systems (i.e. low R-squared values).  With two non-correlated systems, one system may be making money while the other is flat (i.e. out of the market) or one system may be losing money while the other is making a bit more.  The catch is that R-squared is calculated based on historical data.  During explosive market conditions, one system may make money when the other system suddenly loses much more.  Even worse, both systems may become correlated and lose money together.

While these are threatening possibilities, without more details I can’t say it’s reason to dismiss system development altogether.  You could look at a long historical backtesting period to determine how often these conditions occur.  When they occur may be generalizable to a particular type of market condition.  You could explore the possibility of including a failsafe filter–for example, suspending trade of a mean-reverting system in case of powerful breakout or breakdown.  Whether the markets traded by these systems are the same or different could also affect the severity of this threat.

I will continue analysis of the SPY VIX trading system in my next post.

The Subjective Function (Part 6)

I unmasked the systems in yesterday’s post (http://www.optionfanatic.com/2012/10/11/the-subjective-function-part-5).  Prior to that I made all my observations about the graphs shown in http://www.optionfanatic.com/2012/10/09/the-subjective-function-part-3.  Once and for all, the time has arrived to select the subjective function.

On one hand, I like a system that trades more and presents more opportunity for profit.  This sort of system has a profit factor (PF) just over 1.00 and grinds out a small profit in each of many trades.  This would be insert #3.

On the other hand, I prefer systems that are surgical in their efficacy. These generate infrequent trades and have larger PFs (e.g. over 1.50 or 2.00).  Ideally, I would like to take advantage of the time one such system is not in the market by trading with another surgical system that is in the market. I am therefore in the market more frequently with multiple systems that each deliver concentrated profits.  I prefer this model.

The risk of using surgical systems to make the equivalent dollar profit as a frequent trading system is that larger position sizing must be employed.  Although the likelihood is that these trades will end up profitable, in case the next trade results in MDD the larger position size could result in catastrophic loss.

I want a subjective function that takes into account both profit and DD.  A system may have had minimal DDs in backtesting but as one author on system development described, your worst DD is always ahead of you.  If the subjective function does not factor in DD then I fear the Risk of Ruin.

Therefore, I will choose RAR/MDD (insert #1) as the subjective function.  By using RAR instead of compound annualized return, I also like the fact that the more a system is out of the market, the more it is rewarded.