July 12, 2019
How to Optimise Your Trading Strategy
Imagine you entered an international soapbox race.
You have a series of heats before the grand final where you’ll be pitted against the top soapbox cars in the world.
At your disposal for the first heat, you have a car made from an old beer crate, some wheels from a wheelbarrow and some string pulleys to steer with.
Somehow in your first race, you scrape through in the top 3 and make it through to the next round, but you realise that you have some work to do if you want to avoid losing the next heats.
You find a steep hill near your house and start testing. You knew in your first heat that your steering pulleys were slightly primitive compared to the other guys with their steering wheels, so you’re ready to upgrade.
In your next race with your new steering wheel, you perform slightly better. This was a good decision. But you now realise that your wobbly wheels can’t match the solid tyres of your opponents. It’s time to test that and upgrade.
This goes on for a few months; every time you have a new race, you test and upgrade a new section of your car. The races begin getting easier and easier.
This happens up until the semi-final, when suddenly you’re far out-competing your opposition. Apart from a few niggles during the race, your car is now entirely dependable.
Time to iron out those niggles and optimise, ready for the final event – there’s a cash prize involved here, so you want to arrive at your very best.
On the big day, you arrive and realise that everyone is staring at you. You look to your left and see an opponent sitting in a wooden chassis with ‘fresh eggs’ written on the side. The opponent on your right is holding a steering wheel that could only be described as a picnic plate. Both of them are green with envy.
You take a good, long look at your ‘soapbox’ car and realise that after all your testing, after all your weeks of optimising, after all the time spent focusing on the individual areas, you had finally optimised your car as far as it could go.
You look down at the front of your mean-machine and admire the yellow badge you painted on with a black horse – nice touch.
As the flag is about to drop for the start of the race, it suddenly dawns on you…You’re sitting in a Ferrari!
Ok I admit that was a very over-elaborate and unrealistic analogy, but I wanted to get the message across. Trading is an international competition; do you want to be competing with the Ferrari of trading strategies, or the wooden egg crate?
Just like our soapbox car in the analogy, we need to single-out each weak area and optimise it. We don’t want to scrap an entire strategy if it isn’t performing; let’s look at which areas work and which ones don’t. After all, a badly performing strategy may be one small tweak away from being an excellent one.
Hermit Crab Traders
Some traders behave like hermit crabs. Rather than finding which aspects aren’t working, they move into a whole new strategy that looks better than their existing one.
However, you could have a fantastic strategy but some slight mistakes in interpretation (sometimes just a couple of pips either side) could lead you to taking a long trade, when it’s actually a short opportunity.
Therefore, we want to use our trade tracking to find trends in the performance of our strategy and discover which aspects work consistently or need upgrading.
Rather than finding a new strategy, we can develop one that perfectly suits ourselves and is entirely comfortable for us to trade.
Creating Robotic Perfection
When I was developing my automated trading system in 2013, my team and I would go through various stages of iterations in the overall concept before we built a working prototype.
Even once we had a live model, we would still back-test and live-test until we found the variable that needed tweaking.
This inevitably meant we went through versions 1, 2, 357, 1093 etc. of optimisation (it was a long process!) during our search for perfection.
Aiming for perfection is not necessarily the goal in this process. The real target is to tweak areas of your system to reach the desired performance levels that match your trading ambitions.
This means achieving the success rate, risk/reward ratio and trade volume that work together to reach your targets.In this next section, I’ll break down how we optimised our automated system, whilst explaining how you can do the same with your manual trading strategy.
The Game of Probabilities
With manual trading it’s easy for us to forget what trading is all about. Essentially, it’s a game of probabilities; we open trades because we perceive there to be a high-probability of success and avoid situations with low probabilities.
However, a lot of traders don’t even have an estimated probability for their system – in many cases traders are using a losing probability, without realising they’re fighting a losing battle.
By tracking your trades you will begin to establish the probability of your system; this is essential for the next points we will discuss, as we start to raise your probabilities and build your Ferrari.
Stage 1: The Logic Test
I’ve mentioned before that your trading strategy needs to make sense logically. This means every component must have a specific function; if not, you may want to consider cutting rather than adding to your strategy.
Are there any mutually exclusive or mutually necessary aspects to your system?
In other words, are there parts that will only work well when combined with something else?
You may also find that certain parts are under-performing as a result of being combined with other functions that don’t work well in synergy.
Sometimes you need to give your components enough space to ‘breathe’ and not be overshadowed or competing with other parts of the system.
An easy way to discover this is to ask yourself “what information about the market is this part meant to be showing me? Is this being duplicated unnecessarily?"
When building the automated system, our first version was looking for a very specific setup in the market, which looked like a big blue triangle on the screen. The setup was derived from the surrounding price areas and a calculation was made based on these to find an entry point.
However, we soon found that the surrounding price areas weren’t entirely necessary, which meant the blue triangle was just restricting the system.
Instead, we were able to build the system based on a stand-alone calculation that provided a model for the markets. This led to an increase in opportunities, more versatility and a much higher success rate.
Stage 2: Performance Issues vs. Strategy Issues
We often distort our own system results by executing the plan incorrectly.
Without tracking our trades, it would be impossible to know whether we’re doing something consistently. Remember: consistency is the key to good trading.
We would check this with the automated system by printing screenshots of entire days of ‘paper trading’ with arrows representing the entries and exits.
These print-outs would then be compared to the same trades taken using manual trading. This allowed us to check for consistency – was the automated system following the correct plan?
My team ran these manual checks 24 hours a day, 5 days a week to ensure we had a good sample pool to work with. (This isn’t essential for your own trading, but I wanted to point out how the more you track trades, the more dependable and consistent your data will be).
It doesn’t matter if you are performing well or not. All that matters is that you are following a consistent plan because this is the only way to find the areas that need strengthening.
Even if you make money by deviating from your plan, this isn’t a good thing. In a similar way to patting your dog and playing with him, after he’s just left a huge load on your new carpet, the positive reinforcement will create mixed signals and lead you to sin again!Iron out your performance inconsistencies, even if it means you make less money. You can always optimise towards better performance, but not if you have inconclusive statistics that contain a heavy dose of “+/- luck”.
Stage 3: Chunking the System
Think of your overall strategy as a human body; each organ works individually but the overall effectiveness is generated by how they operate as one entire organism.
If your system passed the logic test, you should be operating in a similar structure.
Each section has a unique purpose and role within the system, but the parts create synergy when working together. Therefore, any useless parts can be removed if they cause issues – like the human appendix.
As an example, your system may be defined in the following ‘chunks’:
- Section A determines if the market is overbought or oversold.
- Section B determines the direction the market will move.
- Section C determines the expected size of the move.
- Section D determines the ‘entry area’.
- Section E is the micro-strategy determining the precise entry within the ‘entry area’.
Once you have ‘chunked’ your system, you can uncover the variables from each section that you’ll be testing in the next stage.
With the automated system, we had an underlying code which generated potential reversal levels, based on a complex calculation. This was useless on its own, but combined with the rest of the system, it was very powerful.
The rest of the system was made up of a strategy built on top of the code, which determined whether a trade could take place and in which direction.
The final part mapped-out the ideal stop loss and take profit levels, as well as more complex calculations about the size of the trade and risk level.
Stage 4: Testing the Variables
The fourth stage is the fun part – this is where you make minor (or occasionally major) adjustments to find the optimal value.
You should have a plan for working your way through your entire system, one ‘chunk’ at a time. Only changing one variable within that chunk as you go along, until you find the sweet spot for each one, bearing in mind any dependencies you noted in Stage 1.
At this point, tracking is essential – don’t be tempted to work on gut feel or your interpretation of what works best. Let’s stick to raw statistics.
For example, if you are basing entries on a trend line and have always gone for a Type 1 close on the first test of the level. What happens if:
- You try Type 2 entries?
- You trade the second test of the level?
- You define the required candle and wick sizes on these tests?
I know it seems like a lot of work – but we are talking about building a Ferrari here!
The more effort you put in, the better results you will get out of this. Is this a hobby for you or a chance to live the life of your dreams?
With our automated system, one interesting discovery was made when testing the reversal levels which were calculated as the base of the entire system. We wanted to prove that they were as strong and precise as we thought.
Based on extensive testing, we found that the price movements around the levels were similar to air density when getting closer to the ground.
Our levels were producing a gravitational pull which meant more activity happened the closer the price got to them – this allowed us to rule out the need for any leniency around the line, since we understood the power of them. We didn’t require a buffer for any margin of error. Either the price reaches our level or there’s no trade.
Stage 5: Live Trading and Continuous Improvements
Once you have run through Stages 1 to 4, you should be in a situation where your system is ready to be set free, live on the markets.
It’s your big race day and you’re sitting in the driver’s seat of your brand new Ferrari.
However, the process shouldn’t end there.Markets change their dynamics frequently, so we have to stay ahead of the curve and make sure we’re always developing to remain profitable.
You don’t want to be sitting at your desk one day and suddenly realise your system isn’t working anymore.
My business partner Dr. Alexander Cassar used a technique to encourage continuous development during our automated system development.
During the day he would take screenshots of the opportunities he had traded, whether they were in-line with the system or not. He would also take screenshots of the opportunities he should have taken but missed.
This would soon become a stack of papers and he would check these at night before going to sleep.
But what was he checking for?Firstly, he was double-checking that he had been meticulously following the strategy to a tee, in case he started to deviate from it and form bad habits.
Secondly, he was seeing where there were trends of good or bad performance that could lead to a further system development.
I really admired this smart, forward-thinking and dedicated approach.
In any successful company, regular performance reviews feature as part of the process for achieving continuous improvement. You should be taking this attitude with your exposure to the markets.
If we rest on our laurels, we’re always one small step away from losing money. In the financial markets, money will always move towards the pockets of the smart and best prepared traders.