Do Good Strategies Really Work in all Timeframes?

May 23rd, 2018

It’s the sentence that comforts and fills our hearts with confidence:

“This system works perfectly on all timeframes!”

Whew. That’s good news. If it works on all timeframes, it must be awesome.

I’m definitely buying that!

But what does the phrase “it works on all timeframes” really mean?

And is it true?

First, if a system has a numerical profit target and/or a numerical stoploss, then that system definitely will change in different timeframes.

An entry signal on a Daily chart is vastly different than an entry on a 15M chart. The 15M chart has a lot more noise and is based on short-term market fluctuations. We can’t expect a signal on a 15M chart to travel as far as a signal on a Daily chart.

Plus we have the problem of the stoploss. If our stop is to be placed under a previous low, for example, then the low on a 15M chart is probably only a few points away.

A stop under a previous low on a Daily chart is probably dozens of points away. That, of course, changes everything from risk to reward to trade sizing to target distance.

So, if the system has a hard, numerical stop or target, it’s pointless to talk about that system working in multiple time frames. That type of system changes into an entirely different system with each timeframe change.

If a system doesn’t use numerical stops and targets, though, then we can reasonably, potentially, possibly make a claim that the system works in all timeframes.

Let’s go back to our example last week on the Golden Cross system (last week’s post is here).

The Golden Cross is a basic trend following system and it definitely produces profit (at least it did on Dow Jones stocks). You go long when the 50 simple moving average (SMA) closes above the 200 SMA and exit the position when the 50 SMA closes back down below the 200 SMA.

Those rules make philosophical sense and can easily be applied in the exact same form on any chart we choose.

Thus, it would be easy to claim that this system works in all timeframes.

But does it?

If we put the Golden Cross on a Visa (V) Daily chart, we get a profit of $20,800 from 2009-May, 2018 (trading $10k worth of stock each time) and we’d be currently in a massively profitable open trade. The max drawdown would have been $5,300.

Now let’s transfer this system, with no changes whatsoever, to other timeframes of Visa. Here are the results:

• Daily chart: $20,800 profit, -$5,300 max drawdown
• 240 minute chart: $15,500 profit, -$5,700 max drawdown
• 60 minute chart: $2,800 profit, -$8500 max drawdown
• 15 minute chart: $2,690 profit, -$7,300 max drawdown
• 5 minute chart: -$9,800 profit, -$13,400 max drawdown 

As you can see, when you move to smaller and smaller timeframes, the results get worse and worse, culminating with the 5-minute chart being completely unprofitable.

This system uses a timeless, proven trend following philosophy yet it clearly doesn’t work on all timeframes. Anyone saying so is trying to sell you something.

It begs the question: if a simple trend following system doesn’t work on all timeframes, what does?

Anything?

Here’s the real question, though: does it even matter?

We’ll discuss that on Thursday’s YouTube video.