Sometime this year CIMB Securities started sending me a Daily Trading Ideas report every day by email. They’ve been touting it as investment advice that is immediately actionable. Now, CIMB’s analysts’ reports have a rather poor reputation and are often the subject of mockery on the LYN investment forum, mainly because of their extremely optimistic price targets assigned to stocks. But it’s hard to objectively assess how wrong they are since in theory it is perfectly possible for stock prices to deviate from what they ought to be worth even across extended time frames.

Technical analysis is a different matter. I may not know much about the subject and I certainly don’t know how to do it, but I do know that its supposed to predict market movements in the very short term based completely on momentum factors that are not directly connected to stock fundamentals. This means that it should be possible to assess the accuracy and usefulness of their stock tips. To be safe, I collected price information for their stock picks for the Malaysian market only for 21 days following the publication of each report beginning in around July 2015.

My presentation is organized to reflect how an individual investor would reasonably make use of the report. Ideally, we want to buy a stock at the opening price stated in the report and sell it shortly afterwards at its peak. But in practice, peaks can only be identified retrospectively and no one knows how high the price can go in advance. So instead I’m testing a mechanical strategy: buy a stock, set a target gain threshold and sell automatically when the target is reached. At the end of the 21 days period, if the stock never reaches your target, sell it at the current price. For such short periods, I think targets from around 5% to 10% are pretty reasonable.

My chart therefore shows the opening price and then for each 7-day period afterwards how high it manages to get plus the gain, if any, expressed as a percentage of the opening price. However since my reasoning is that you want to sell at the peak price, I bring the best price forward, that is if the peak price is achieved in the first 7-day period and it drops subsequently, I just show that peak price. If the opening price is in fact the peak price, I also just bring that forward. Finally, I show the closing price of the stock at the end of the 21-day period to reflect your selling price would be if the stock never managed to reach your chosen target. Note that the data are presented in the order sent by CIMB. In many case, the same stock appears more than once because they were picked on multiple days. I simply treat each one as a distinct buying event.

I realized pretty quickly that my methodology is flawed. In the cases where there are gains, much or even all of the gains occur during the first day following the release of the report. When this takes place in the form of a gap-up, no investor can buy at the stated price and take advantage of the information. Of course, one must also keep in mind that just because a stock hits a certain peak, it doesn’t necessarily mean that you can sell at that peak even if you have a sell order at that price and at that time. Conversely, it would be easy to set a cut-loss target for oneself so that the largest losses after the 21-day period could be minimized. Given that this is momentum based technical analysis, it would make sense to cut loss very quickly once you see that the momentum is going the wrong way. Due to these problems, I’m unwilling to add everything together and form any kind of overall conclusion.

Even so, it should be clear from this data that this strategy is far from fool-proof. It looks to me that it actually works decently when times are good but crashes hard when the overall market sentiment is bad. But since all stocks tend to do well when the market as a whole is buoyant, it isn’t obvious if this is of much use. Still, I have to admit that looking at the data, this kind of strategy isn’t completely hogwash and it may be interesting to spend some money that I’m willing to lose to experiment with it.