Nominal Trading Account: Php 1,114,630
No Signals – ALI, AEV, AP, BDO, BEL, CEB, EDC, FGEN, GLO, ICT, JGS, JFC, MBT, MER, MWC, PX, RLC, SM, SMDC, SMC, SCC, SMPH, TEL
Buy Signals – AC, AGI, BPI, DMC, MEG, MPIC, URC
Sell Signals – None
March 1, signals triggered were AGI, BPI, DMC, MEG. What should be bought were 22,000 AGI @ 11.45 (Php 251,900), 4,000 BPI @ 71 (Php 284,000), 6,900 DMC @ 48.6 (Php 335,340) and 133,000 MEG @ 1.83 (Php 243,390). Of course, if you add up all 4 amounts, you get Php 1,114,630.
What on earth is going on? Isn’t that the same as the nominal trading amount? Yes. That was intentionally modified post-hoc. Originally, the nominal trading amount was set at Php 1,000,000, which was 114,630 Pesos less than what was required for the March 1 string of buy signals.
In real trading, you take the signals as they come. If you’re running out of trading ammo, the last position you take before you are out of cash-on-hand may be only partially filled. Here, since my goal is to show you around the workings of a primitive trend-following system, I’m just fixing the problem post-hoc to make the computations easier. As they say in professional wrestling, “don’t try this at home.”
Now, in random order, let me show you some of what just happened. 4 signals triggered on March 1, and now the system is all-in the market. AC, MPIC, and URC triggered on March 2. With no cash-on-hand, none of these trades can be taken. In real trading, if I’d run into the Php 1,000,000 cash vs. Php 1,114,630 worth of signals problem, I’d have used tie-breakers.
For example, the the 20-day highs are the following % higher than the 10-day lows at the time of the March 1 trading signals: DMC +6.34%, BPI +7.57%, AGI +8.63%, and MEG +8.92%. In order of priority, I would take MEG, AGI, BPI and only then try to fill in the DMC trade with what was left of my trading cash.
I can hear some astute readers ask, “Wait! Isn’t this all post-hoc again?” Well, yes and no. If you’re trading a very short-term system (like what we’re demonstrating now), yes, that would be post-hoc. It wouldn’t matter which stocks were going faster since you would simply take the trade signals chronologically. However, I use a slower system for my personal trading. It’s easy to infer why I can prioritize the taking of trading signals this way without resulting in a post-hoc fallacy. Although I could, I would rather not spoon-feed you the solution to that small mental puzzler. You can do it. It’s easy.
Back to the system. How did we determine these position sizes? Simple. Let’s take AGI:
March 1, it set a new 20-day high at 11.45 (20 days refers to 20 trading days, highs count intraday highs of that period). The 10-day low was 10.54 (10 trading days, intraday lows).
2% of Php 1,000,000 (the original nominal trading account) is Php 20,000. To risk 2% of portfolio on each trade, 2% should be the maximum lost if the sell signal triggers. 11.45 – 10.54 = 0.91. 20,000 / 0.91 = System recommends buying 21,978 shares of AGI. Rounded off, it’s 22,000.
Why 22,000 shares? If, within the day, the worst case scenario (cut loss stop triggered) occurs, 10.54 is hit, remember that the system risks 2% for a failed trade. 10.54 would represent a loss of 0.91 Pesos per share. 0.91 x 22,000 shares = Php 20,020 or about 2% of Php 1,000,000.
Yes, there was no accounting for slippage at all. You can do that in your free time on your Bloomberg terminal. What I wanted to do was take you through an example of volatility-based money management. The part of the system that answers, “How many shares do I purchase?”. It’s different from parts of one’s system that tell one “When do I buy/sell?” and “What do I buy/sell?”. Money management is often misunderstood and overlooked by many traders.
As of today then, March 4, 2012, the trend-following demonstration system has the following trading positions:
22,000 AGI; 4,000 BPI; 6,900 DMC; 133,000 MEG; No cash.
I hope you enjoyed this initial entry. I actually don’t recommend anyone trade these exact parameters. Personally, I find a 20-10 system too noisy for my liking. It also takes too much time and generates too much stress. One should also look into using better trading signals than these channel breakouts. Volatility breakouts can often outperform these old-school tools. The standard deviation concept, I think, is also a better alternative than the % tie-breaks I like to use. Take the ideas you like, optimize them to your preferred trading horizon, your personality, time available for trading, markets traded, etc.
There’s a lot to be learned from simple trend-following systems. They don’t generate the best returns. But anyone can learn them and make decent returns.
Questions? Comments? 🙂