The Not-So-Hidden Secret to Building Robust Trading Strategies
Systematic strategy creation requires a systematic development path!
A wonderful chat with Dr Tom Starke who is on Substack here! We’ll be doing workshops with Tom in the Algo Collective membership, check it out! Stay subscribed as these workshops are dropping soon and are extreme value for money in the Collective for the first 50 members!
1. Why Most Traders Fail: The Missing Third Dimension
Too often, traders start by looking for signals or strategies that seem promising. However, without a clear objective or understanding of the underlying risk, success remains elusive. This is a trap many retail traders fall into, starting with high hopes but lacking a plan. A critical step in building a robust trading strategy is to define your objectives upfront.
For instance, are you trading your life savings, or is this speculative capital? If it’s your retirement fund, your strategies need to be far more risk-averse, with a long-term horizon. On the other hand, speculative trading may allow for more risk but still requires a disciplined approach.
1.2 Trade Like a Business, Don’t Gamble
Think of your trading portfolio as a business. It’s not just about finding a shiny new signal. You need to account for everything that goes into running a successful operation: data acquisition, costs, market knowledge, tools, counterparties, and more. Similarly, institutions like hedge funds or pension funds don’t just focus on one strategy. They focus on the entire business—how they allocate capital, manage risk, and sustain their edges over time.
1.3 Move Beyond Narrative: Embrace Scientific Thinking
The retail trader is often drawn to narratives—the stories behind the trades or the next big indicator. But Tom Starke emphasizes that institutional success comes from evidence-based research and rigorous hypothesis testing. The best quants don’t chase stories; they rely on data to uncover the truth. This requires learning how to interpret data scientifically and using statistical techniques to validate signals.
A common mistake made by retail traders is the pursuit of a “magic signal,”. But these signals, or a single metric (like ‘win rate’) are often devoid of context and fail to account for the complexities of the market. As Starke notes, scientific thinking leads to an understanding of how signals function within broader systems—what risk they carry and whether they remain robust across different market conditions.
- Start with the end in mind. Things will take time, don’t expect perfection at the outset.
- Stay simple & robust at the outset. Keep expectations low. Buy yourself time to learn the ropes and succeed. Don’t burn out.
- Trade real money as soon as possible.
- Keep your day job, it keeps the pressure off.
- Run it like a business. Have a business plan.
2. Define the Game Before You Play It: Objectives & Constraints
Your trading capital significantly influences your strategy’s viability. For instance, small retail accounts often face issues like slippage and higher execution costs compared to large institutions, which have access to better liquidity, lower spreads & advanced order execution algos. Tom Starke stresses the importance of understanding the limitations of your capital. A strategy that works with $100 million may not work for a $10,000 account. Furthermore, cost drag and turnover are hidden killers that retail traders frequently overlook. Conversely, many strategies that trade fine on a few hundred thousand dollars, are hard to scale even to $5 million. An important consideration if your objective is to trade for a fund.
2.2 Clarify Your Risk Profile
The next question you must answer: How much risk are you willing to take? Starke advocates for clear risk profiling, not just focusing on potential returns. Institutions prioritize stability and lower risk over time, aiming for consistency, not “home runs.” Retail traders, however, often chase high-risk, high-reward strategies that lead to catastrophic losses when things go wrong. By defining your risk tolerance early, you can ensure that your strategies align with your long-term goals.
- Capital is not homogenous: high risk returns <> low risk returns
- In institutions, mandates (objectives) anchor everything
- Account size determines viable strategies
3. The Institutional Trifecta: Risk, Robustness & Execution
Risk management is the bedrock of institutional trading. The strategy signals themselves are less important than how those signals are weighted in a portfolio. A profitable strategy can still lose money if it is sized incorrectly. This is why Starke mentions the Kelly criterion—a formula that helps determine optimal position sizing based on your edge (there are other options and additional considerations, but it’s a start). Without the right sizing, even the best strategy could fail due to improper risk allocation.
3.2 Robustness Over Optimization
Robustness means that a strategy will continue to work in different market environments, not just under the conditions it was tested. Too often, retail traders optimize their strategies to fit historical data, forgetting that past performance is not always indicative of out of sample results. Starke argues that (most) institutions prefer lower-risk, stable returns over “home runs.” Robustness testing—using walk-forward analysis, Monte Carlo simulations, and out-of-sample testing—helps ensure that your strategy won’t just perform well on paper, but the edge will continue to play out in live markets.
For an institution, ‘edge’ isn’t just ‘expectancy’.
“Expectancy”=Probability of (”win” )×”Avg Win” - Probability of (”loss” )×”Avg Loss”.
Expectancy is a vital metric and a good theoretical proxy for edge – it captures the idea of having odds in your favour. But a truly valid edge in trading is more than just a number; it encompasses risk-adjusted performance, consistency, and the ability to actually realize that expectancy in live markets. A comprehensive definition of edge might be: a repeatable strategy or advantage that produces positive expected returns after accounting for risk and costs, and that can be executed with discipline to yield persistent outperformance. Achieving that is the holy grail that every trader, from the ivory-tower quant to the instinctive macro guru, is ultimately seeking.
3.3 Execution: A Hidden Pillar
Many retail traders overlook the importance of execution. Slippage, spread, and market impact can all eat into profits. Starke highlights that even small improvements in execution can make a significant difference over time. Institutions spend significant time refining their execution processes, reducing transaction costs, and minimizing slippage. Retail traders can level up by improving their execution methods, ensuring that they pay attention to every aspect of their trading costs. A good order management system or a basic knowledge of IB order execution algos might help enormously.
- Strategy signals matter far less than portfolio construction
- Tiny errors in position sizing can destroy edges
- Retail pay higher costs & must work with less capital. They must compensate through smarter strategy design and efficient use of capital
- Retail traders have access to strategies that institutions are too large to bother with
4. The Real Work: Building a Process for R&D
Starke stresses that back-testing should come at the end of the research process—not the beginning. Far too many retail traders fall into the trap of optimizing strategies for past data before they even know what they’re trying to accomplish. Institutional traders, however, build hypotheses first. They define their objectives and constraints, test hypotheses, and then validate them using back-testing last. Back-testing is only useful when you’ve already answered key questions about risk, strategy design, and market conditions. Why? Because it’s just not good research. It’s prone to overfitting; ignoring blind spots; missing context. Research will give you an intimate understanding of the edge you are trying to capture, and that will drive your strategy design.
4.2 Evaluate Signal Significance
Once you have a hypothesis, you need to determine whether your signals are statistically significant. The goal is to determine whether a signal is predictive, not just a product of market noise or the select group of filters you attached to it. Retail traders often base their strategies on signals that appear strong in nicely ‘filtered’ back-tests but fail in real-world conditions. Tom recommends using statistical tests to assess the strength of signals before committing to them. Are your signals actually predictive, or are you just lucky with the sample that was in your back-test?
4.3 Build Robustness into Your Process
The heart of an institutional research process is its focus on robustness. Institutional quants build multiple layers of testing into their process, such as walk-forward analysis, stress testing, multi-market testing, Monte Carlo simulations and more. These are baked right into the research process, not applied after the fact.
These methods simulate various market conditions to ensure that the strategy performs well when it is supposed to. Retail traders can do the same, running their strategies through different market regimes to see where performance is generated, and when risks emerge out of the dark.
4.4 Path Dependency: An Example of a Back-Testing Flaw
One of the most common mistakes retail traders make is seeing the single path of a back-test as a definitive guide to the future. But back-tests can often be “path-dependent”—a strategy may look profitable because of the specific time frame or sequence of events chosen. A single back-test is just a single path among many, many possible paths. Instead, we must look for strategies that are not dependent on specific starting points, or a highly specific sub-set of the total universe of possible trades. Don’t be afraid to look at worst case scenarios, worst possible trades, etc. Try to invalidate your hypothesis. Focusing on the weak spots is the only unbiassed way to know whether it can be reinforced to handle the likely strain or not.
- Develop a systematic pipeline for the development of systematic strategies
- Automate it where you can to compound your search area
- Spend more time researching a concept, prior to pinning exact filters, entries and exits to it and calling it a strategy
- It’s trading, not true science, so stay creative
- Reduce degrees of freedom early, invalidate quickly, move on to greener pastures
- Remember that statistics are averages, usually ignorant of sequence risk or events we haven’t seen before
- Build around the concept of avoiding catastrophic failures
- Assess all possible trades, not just the filtered ones
- Consider sequence of trades, and correlations
- Map to regimes, and expectations
- Use as many robustness testing methods as possible, but not the ones irrelevant to your particular model
- Step away from the screen and sit under an apple tree for a while meditating on what you’re trying to achieve
- Know your metrics, and which are relevant to the task at hand
- You will think more in terms of ‘risk drivers and return profile’ than in ‘mean reversion’ or ‘trend following’
5. Combining Strategies into a Unified Portfolio
A robust trading system is more than just a collection of individual strategies. It’s about portfolio construction—how different strategies work together to reduce risk and increase consistency. Think of each strategy as a “product” with its own risk profile and return characteristics. When combining strategies, you should consider their correlations, how they behave in different market environments, and how they complement each other.
5.2 Diversification in Action
Starke stresses that diversification is not about adding more assets to your portfolio but about finding uncorrelated drivers of risk and return. For instance, you might combine mean-reversion strategies with trend-following strategies, as they tend to perform well under different market conditions. A sufficient number of discrete strategies aggregated can behave like one continuous, dynamic portfolio. Positions can be netted, multiple variants can throttle exposures, bet sizes shrink (as can commissions) and risk can be more easily managed.
Additionally, your ‘average returns’ are more likely to play out in the future since you aren’t trying to pick the ‘one thing’ that is going to succeed next year. Hint, it’s never what succeeded last year!
5.3 Sizing and Turnover
Position sizing is crucial to portfolio construction. Even a good strategy can fail if the position size is too large or too small. Stop and think about what it means to use capital efficiently. This is especially important for retail traders who often have less capital and higher transaction costs. Efficient portfolio construction ensures that each strategy works harmoniously within the larger framework. For examples, check out the podcast with Corey Hoffstein (Ep 27) where he talks about ‘return stacking’ – a stock portfolio can often add futures (or other leveraged products) without requiring additional capital.
- Diversification doesn’t mean more tickers, it means uncorrelated drivers
- Seek to quantify diversification
- At every possible opportunity, reduce your reliance on any single edge
- Design strategies with the portfolio in mind
6. Mindset Shift: From Retail to Institutional Thinking
The first mindset shift is moving away from focusing solely on trading signals and indicators. Institutions think about the process—how strategies are built, how they fit together in a robust portfolio, how to size positions, and how to manage risk. Starke emphasizes that without a defined research process, you risk making decisions based on biases, narratives or overfitting.
6.2 Get Comfortable with Iteration
In institutional settings, research is an ongoing process. Institutions don’t expect strategies to work perfectly on the first try. Starke’s experience in academia and industry has taught him to be comfortable with iteration—testing hypotheses, refining models, and validating results. As a retail trader, adopting this iterative mindset (and automating where possible) allows you to build strategies with more rigor and avoid emotional decision-making – dropping bad ideas quickly.
6.3 Don’t Fall for Quick Fixes
Another critical mindset shift is to avoid falling for quick-fix solutions. Retail traders are often sold “simple” strategies or tools that promise instant success, but these rarely deliver in real-world markets. Institutional traders take a long-term view and focus on building sustainable, evidence-based strategies. They work in teams, they spend time on automation, and they stay accountable – excessive risk taking isn’t an option.
- It takes time to know what you don’t know
- Build your network: cross-pollinate ideas; stay accountable
- Take the mindset of a professional sceptic
7. Conclusion: Key Takeaways for Traders Ready to Level Up
- Define your objectives clearly: Understand your risk tolerance and the capital you’re willing to invest. This will shape your strategy and trading approach.
- Focus on risk management: A good strategy is only as good as the way you manage risk. Proper position sizing and portfolio construction are essential.
- Build a process: Follow a structured, evidence-based approach to research and strategy development. Don’t jump into back-testing without a clear hypothesis and an understanding of your risk.
- Diversify across strategies: Combine different strategies that complement each other to reduce risk and smooth returns.
- Shift your mindset: Focus on building a process, not chasing a signal. Be prepared to iterate and refine your strategies based on real-world data.
If you adopt these principles, you’ll be well on your way to building institutional-grade strategies that can stand the test of time.
Never give up!
Get in touch with Tom