The Motley Fool Discussion Boards

Previous Page

Investing/Strategies / Bonds & Fixed Income Investments


Subject:  A Simple Trading System (for ETFs or Whatever) Date:  1/21/2013  4:55 PM
Author:  globalist2013 Number:  34715 of 36327

The term ‘trading system” is just another way to describe your ‘investing plan’. It’s the set of rules you’ve committed yourself to following as you put capital at risk in the rational expectation of gaining profits. Neither ‘Hope’ nor ‘Good Luck’ play a role in generating profits for the account (though ‘Bad Luck’ does have to be defended against). This isn’t to say that your plan will necessarily succeed in obtaining those expected profits, nor that the profits might not well exceed your expectations. You just can’t know ahead of time, for markets being the complex, chaotic, ever-changing things they are. But there are three constants you can count on. One, human nature (hence, the role of Fear and Greed in buying/selling decisions, both yours and your counter-party’s) never changes. Two, however much over-bought/over-sold securities (stocks, bonds, commodities, currencies, real estate, etc.) become, supply-and-demand dynamics will eventually prevail. Three, “Markets can stay irrational longer than you can stay solvent”, as Keynes famously quipped.

What those three market constants mean in practice is this. Even if you’d like to echo the cutesy pretension that “your favorite holding-period is forever”, you’re just a gambler making bets about the level/direction of prices, and the sooner you figure out for yourself a disciplined way to be making those bets, the better are going to be your chances of survival. In practice, that means two things. One, you’ve gotta find an investing/trading game with a positive-expectancy that is simple enough for you to understand and execute on. Two, you can’t over-bet your hand. The customary --though unprovable-- assumption is that stocks (as an asset-class) offer a positive-expectancy over the long haul. So, theoretically, if you did nothing more than buy a representative basket and hold that basket a sufficient time, you’d capture the gains observed in the past. That’s the bet stock indexes are making, that past performance predicts future performance when the holding-period is long enough to replicate the characteristics of the sample. But as Keynes also quipped, “In the long run, we all are dead.” So, ‘holding forever’ in hopes that losses can be made up is a crazy way to be making bets, as is ‘doubling down’ (and other Martingale strategies).

Instead, the better strategy and tactic is to bet on what’s doing well right now, or what’s going to be doing well soon, and let them that have the money to lose bet on the rest. That means, in practice, two things. (1) “Bet with the crowd when the crowd is right.” (2) “Bet against the crowd when the crowd is wrong.” And it’s actually not that hard to tell ‘when’ is ‘which’, as the following charts suggest. [to be filled in later if I get around to it] But here’s the guts of a family of trading systems. The simplest version uses ‘brute-force’ trend-following, going long only. A slightly more sophisticated version trades both from both sides of the market. The next level up adds counter-trend trading.

Simple version.
The truly easiest way to keep yourself out of trouble is to accept the fact that when prices move against your position, you’re losing money. You can argue with prices if you’ve got the time and money to do so, or you can accept the fact that your bet isn’t working out as you expected and cut your losses. For sure, a lot of adverse price movements really are no more than news-driven noise and need to be ignored and suffered through. But the situation you don’t want to find yourself in is discovering --after the fact-- that the signal should have been acted upon because the worries expressing by sellers selling had a sound basis. What a good exit point for each stock and/or each account holder might be is for each investor/trader/better to decide for her or himself. When a position doesn’t work out for Buffet, he liquidates, so should you. Selling losers is simply good investing/trading/betting, especially if it can be done while the loss is still small. There’s at least two ways to set hard stops. One is to use a fixed percentage, as does William O’Neil. Another is to exit on the reverse of the conditions that got you into the trade, and those conditions could be ‘fundamental’, or ‘technical’, or a combo. It just doesn’t matter. Flip a coin, and pick one or the other. Your results are going to be equivalent. What you think will work for you is what has the best chances of working for you. It just doesn’t matter.

But basing investing decisions on technical considerations is an easy way to go. So let’s work that angle, and let’s do our work at Yahoo Finance this time, instead of or Pick a security to chart, any security. It doesn’t matter for developing the rules for a trading system. But since this is a forum for ‘fixed-income’ investing, let’s throw TIP up on the screen.;range=2010...
If you’re not used to charting, the lines will be confusing. So do two things. One, turn off the Bollinger Bands (by selecting that indicator and deleting it). Two, grab the slider bar at the bottom of the chart and move it back and forth in time. Hugely better, right? And hugely obvious, too. When prices are in a clearly defined trend (up or down), the three moving averages ‘stack’ and display themselves like ribbons. OTOH, when market action gets ‘choppy’, the moving averages become intertwined with each other.;range=2009...

Now, here is where user-input becomes necessary. The parameters for those three moving averages (MAVs) are 16-24-32. Each successor number has a geometric ratio of 1.5x to each predecessor number. And if that terrorizes you math phobes, think about the well-known Fibonacci series, 1-2-3-5-8-13-21-34-55 etc., in which each successor is roughly 1.6x its predecessor (and also, the sum of its two predecessors). The intervals could have been 1.4x, or 1.7x, or they could have been an arithmetic series like 20-30-40. It just doesn’t matter. For sure, through back-testing and data-mining, one could find the “best” values to use. But if those values are used out-of-sample, you can bet they will fail to perform as well, because they were ‘curve-fitted’ to a data sample that won’t likely reoccur in the future. That doesn’t mean the parameters chosen are totally arbitrary, nor are the types of averages in which they are used. ‘Faster’ averages produce more signals, but they are much more vulnerable to ‘noise’. Slower averages tend to be more ‘robust’, but they also give back profits toward the end of a trend. Also, ‘simple’ MAVs behave differently than ‘exponential’ ones, than ‘weighted’ ones (which both Scottrade and E*Trade offer their clients), than ‘triangular’ ones (which are just a doubly-smoothed simple MAV), than ‘volatility-based’ ones, etc. There are probably a dozen moving average varieties, and all of them can be made to perform pretty much like each other simply by varying the look-back period. But I’d say this. Stick with simple moving averages. They are simple [pun intended], classic, easy to understand, easy to compute, and they are robust. What some thought should be given to, however, is how many MAVs you want to see on the screen and how fast you want your signals to occur. Reportedly, institutions (like the big pension funds) depend on a pair of MAVs, the 50-day and the 200-day, and they make their moves in and out of their positions accordingly (typically, also using ‘fundamentals’ to also justify the move). If that sort of glacial pace appeals to you, go for it. If you want something more responsive to present market conditions, use faster parameters. Chief’s choice.

Now, here’s the chief problem with trying to use moving-average crossovers as your signals for when to buy and when to sell. You’re going to get ‘whip-sawed’, which is a logging metaphor that describes the back-and-forth, in-and-out that can occur when markets are ‘choppy’ and lack sustained trends. However, there’s also an answer to that problem. “Live with it.” Seriously, any successful trend-follower will tell you that trying to tweak your rule set to avoid whipsaws is just going to get you into worse trouble than enduring them. Here’s the link to an hour-long talk that makes that point. Or here’s a quick, six-minute version, Ed Seykota’s “Whipsaw Song”,

OK, this post has gotten to be longer than I intended, and I’ve still got a list of points an arm long I need to make. But another day. Now, it’s time to get out on a bike before the afternoon’s shadows lengthen and temps drop back down below freezing. [to be con't]

Copyright 1996-2018 trademark and the "Fool" logo is a trademark of The Motley Fool, Inc. Contact Us