“We need our trading strategies to adapt to the market to maintain their effectiveness!”
What does that statement mean and how do we achieve it? For our trading strategies to adapt to the markets, they need to have adaptive components.
Defining Adaptive Components
Adaptive Component: A component in the trading strategy that adapts to market conditions.
Adaptive components could be in any part of the trading strategy – entries, exits and position sizing e.g. signal generating rules, indicators, take profit and stop loss etc). Let’s look at some examples to get a clearer picture.
Example of Non-Adaptive Components (Not Good!)
- Hard stop of 100 pips
- Enter a trade when market is 50 pips above yesterday’s high
- Enter 1.5 standard lots per trade
Example of Adaptive Components (Good!)
- Hard stop of 2 ATR (adapting to volatility)
- Enter a trade when market breaks 20-day high (adapting to market range)
- Risk 1% per trade (adapting to account balance)
Importance of Adaptive components
We want our trading strategies to adapt to market conditions to maintain their efficiency. Efficiency is defined as the ability to capture market inefficiencies.
Let’s look at a case study of a component adapting to market range:
Fig 1: Non-adaptive mean-reversion strategy.
In the scenario above, we have a non-adaptive entry rule. We short at the top red line and long at that bottom red line. This rule works well between time = 0 and time = 1. As the market behaviour changes, the strategy is unable to adapt.
The aim of this strategy is to capture mean-reversion tendencies, but it fails to do that between time = 1 to time = 3. Between time = 1 and time = 2 it does not capture any trades. Between time = 2 and time = 3 it has a tendency to enter some of the trades too early.
Fig 2: Adaptive mean-reversion strategy.
In this new (ideal) scenario, the entry rules are changed such that they adapt to the market range. In every time period it changes itself to adapt to the new market range.
What Should We Adapt To?
Our main priority is to adapt to the market inefficiency we are capturing.
- On a mean-reverting trade, our entry/exit rules must adapt to market range.
- If your strategy depends on tweets, your rules may need to adapt to the rate of tweets or overall tweet quantity/ratio.
- If you trade co-integration between stocks, you may want to find out how the extent of price discrepancies affects your strategy’s performance (linearly or exponentially etc) and design your rules accordingly.
Besides the market inefficiency, there are other elements to adapt to. What these elements are depends on our trading strategies' characteristics. There isn't a one-size-fit-all list for this but here are some popular ones.
- Volatility - Few strategies are volatility independent. A common component that should adapt to volatility is our position sizing.
- Account Balance - Trade less when your account is smaller and vice versa! (Unless it’s a martingale betting strategy, which is generally a bad idea)
- Current Risk Exposure – Total amount you can lose if all your positions get stopped out. We need to find the sweet spot.
Adapting Our Adaptive Component
We need to go one step deeper in order to adapt. Adaptive components may have fixed components within them, so we need to adapt our adaptive components. To illustrate this, let’s look at some strategy rules.
- Enter on 20-day high
- Stop Loss of 2*ATR(15)
What determines these fixed parameter values: 20, 2 and 15?
In order to adapt our parameter values, we need to conduct a special type of optimisation called the Walk-Forward Optimisation (WFO). However, we won’t be talking about it in detail here.
For now, just know that the WFO enables us to forecast optimal future parameter values using past data, with minimal curve fitting if done correctly.
 Price levels that the market is trading within.
AlgoTrading101 is the first comprehensive online course on algorithmic trading. Learn the fundamentals of algo trading and put your ideas into action along with more than 25,000 students. Learn more about us at AlgoTrading101.
Lucas has been designing and building algorithmic trading robots since 2010. He worked at a proprietary trading firm and taught programming for financial applications to Government of Singapore Investment Corporation (GIC), one of the largest sovereign wealth funds in the world. He runs this blog and the AlgoTrading101 course.