Compound Interest for Your Brain

Or, why trading gets easier when you stop taking shortcuts

In a past life, I was a backtest wizard. This was a complete waste of time.

I conjured flattering tear sheets from the ether of historical data. I tested thousands of combinations of moving averages and entry/exit rules, goal-seeking that perfect up-and-to-the-right equity curve in a game of imaginary drawdown whack-a-mole.

More than once, drunk on hopium and convinced that my latest opus magnum held the key to the gates of wealth, I took a strategy live... only to see it go absolutely nowhere.

I have no idea how many times this backtest cycle of doom repeated. I’m sad to say that it was a lot.

This is not about edge

You probably think this is going to be about trading with an edge. That’s a fair assumption; it’s a topic close to my heart.

But it’s not about that. At least, not directly.

It’s about something less obvious. It’s about missed opportunities. And their compounding cost.

The right kind of hard work

Here’s what I eventually realised:

I was putting in plenty of hours. I was grinding. Telling myself I was doing the hard yards.

But I wasn’t doing the right kind of work. I was taking shortcuts that cost me time, money and sanity.

Running endless backtests requires time and coding skills, but it demands almost no real mental effort. I was essentially outsourcing my thinking to my laptop. And that should have been a massive red flag.

Real progress in trading comes from a different approach entirely - one that feels harder at first but creates compound benefits that make everything easier over time.

How to make money trading

There are two ways to make money trading:

1. Taking advantage of risk premia (risky things trade at a discount to their expected value because they’re risky)

2. Taking advantage of mispricings (buying things for less than they’re worth and selling them for more than they’re worth)

I know this sounds obvious. But people do genuinely believe they can make money by predicting the future with lines on charts. Hell, I convinced myself that I could make money by backtesting all the trading rules.

Those approaches have something in common - they’re almost certainly not going to work.

If you accept my assertion, and I hope you do, then the obvious implication is that it’s helpful to know some things about markets, who participates in them, and how and why they trade. Without that knowledge, it’s difficult to find mispricings, and by extension, to make money (beyond what you can do with simple risk premia harvesting).

Compounding knowledge

When you first start thinking about how to find real edges, it seems impossibly complex. Market behaviour looks random - because it mostly is. You have no idea where to begin.

That’s normal. We all start there. Everything is hard, until it isn’t. Why would trading be any different?

The solution lies in systematic thinking - observe, hypothesise, test, learn, repeat.

That sounds more grand than it really is. What we do certainly isn’t quantum physics. Market effects are blunt, noisy, and imprecise. Hardly the domain of high science.

But adopting the mindset of someone who wants to understand what the hell is going on, rather than just hoping their laptop will figure it out for them, will get you on the right path.

It looks like this:

1. Notice something interesting in the market (or hear it from someone else)

2. Think about why it happens, what might cause it, and why

3. Test it with simple, focused analysis and good questions (what would I expect to see if this were true?)

4. Look at the evidence and see how your idea stacks up

5. Implement what works and learn from what doesn’t

There’s a subtle magic in this approach.

Every time you go through this cycle, you gain more than just a potential trading strategy. You develop pattern recognition, know-how, and intuition that makes the next cycle faster and more effective.

It’s like a snowball rolling downhill, getting bigger with each rotation.

Engage thoughtfully

The “curious scientist” approach takes a measure of creative thinking and mental effort. Distilling it into a mechanical process to be applied in every situation won’t work. You must take each one on its merits.

Thoughtful engagement with the market requires effort, but it’s also way more exciting than a formulaic process (which is unlikely to work out anyway).

The foundation of being a curious market scientist is thinking about why the edge exists. Once you get good at that, everything else falls naturally into place. The right approach reveals itself as a function of the quality of that first bit of thinking.

How it compounds

When you first get serious about trading, it will take weeks to research a simple hypothesis. You’ll spend days asking the wrong questions and trying to answer them with the wrong tools, and even more time interpreting what you’re seeing.

A year or two in, you’ll do this work in days instead of weeks. Fast forward further, and it becomes hours instead of days (especially if you know how to prove yourself wrong quickly).

Eventually, you’ll be able to decide if something has potential in minutes, not hours. You’ll have this ability because:

1. You have intuition about the sorts of things that are likely to work and why - aka, trader smarts

2. You know how to chunk a problem down into manageable components

3. You know which questions to ask first and how to use your data to answer them

4. You know how to disprove things quickly and move on without being tied to an idea

5. You instinctively understand which analysis patterns will reveal what’s hidden fastest

6. You can quickly connect new ideas to concepts you already understand

The knowledge compounds via network effects - each piece connects to everything else, making all the other pieces more valuable.

Data mining purgatory

Compare this to the data mining approach, where you scan for patterns without understanding what drives them.

Someone recently asked me if scanning commodity price histories for statistically significant seasonality effects would be a good idea.

The answer is no.

I guarantee that you’ll find something that appears to work. Scan enough price histories and you’ll even find something that blows statistical significance out of the water, p-hacking your way to backtest glory.

But almost certainly it won’t work where it matters - in the future.

Worse, if you trade that data-mined effect, you have no feedback about whether it’s a good idea other than your trading P&L. And trading P&L is noisy. Good edges go months or more without making money.

Worse yet, and this is the most insidious and costly of all the drawbacks, the data miner gains no transferable knowledge. Each new idea requires starting from scratch. He never develops that pattern recognition system. He never acquires his trader smarts.

It’s like having to learn to ride a bike from scratch every time you get on one. Exhausting, inefficient, and demoralising.

It’s messier than it sounds - just be a grown-up about it

This approach is science-inspired, but don’t get hung up on the rigours of the scientific method. It’s an ideal you won’t be able to live up to. But don’t stress about it - this is people buying and selling stuff, not cancer research.

Your hypotheses aren’t formed in a vacuum. They’re polluted by what you saw in the data last time you looked. Market data is finite, and you know what worked and what didn’t in the past. This will mess with your head in ways you won’t even notice. Sometimes your analysis will reveal answers for questions you hadn’t even asked, and you’ll backwards rationalise them into a trade.

You’ll do all sorts of things that a real scientist would cringe at.

But that’s OK. The market doesn’t give extra credit for being the best scientist. But it will certainly take it away if you bullshit yourself.

So the important thing is that you approach your work in the right spirit - paranoid that you’re wrong, rather than hopeful that you’re right. This will mostly save you from yourself, and you’ll be fine.

I’ve also made this sound way simpler, smoother, and more linear than it really is. It will very much be a case of three steps forward, two steps back. Sometimes you’ll make mistakes and be forced to unwind what you thought you knew to the point in history when you made the mistake. This is like ten steps back, and it really sucks, but it’s a great teacher.

Evidence will rarely be clear-cut. You’ll have to weigh uncertainty, decide which bets are worth taking, and live with the consequences.

Accept that it takes time. And enjoy the ride - because it’s the greatest ride in the amusement park.

Opportunity cost, compounded

In some ways, data mining has an opportunity cost curve similar to the 40-year-old who wishes he’d started saving at 20.

At 20, he knew intellectually that compound interest was a thing. But he was more interested in spending what little money he had at the pub or backpacking or otherwise living the life of a 20-year-old. So he missed out on years of compound interest.

But at this point, the comparison ends. Data mining actually costs way more.

At least the 40-year-old with savings regret has some lived experience to show for his decisions. I can mount a strong argument that those lived experiences are worth multiples on whatever dollars he might have compounded.

It’s folklore that Buffett would forgo a can of Coke because he felt that the dollar it cost, when compounded over many years, was ultimately worth way more than that hit of sugar and caffeine. There’s something unsettlingly monastic about that. And I wouldn’t want that for my own life. But there are lessons lurking in his insight.

I can’t say I have much to show for my “backtesting is research” phase. Maybe some coding skills that are slightly sharper than they’d otherwise be. It’s hard to come up with much more.

I feel it keenly as a wasted opportunity.

How to ride the compound learning train

Acquiring trader smarts takes some work. But so does anything worth having. Here are some tips:

1. Focus on understanding why before testing what
   When you spot a potential edge, spend time thinking about the market mechanics behind it before rushing into data analysis. Definitely don’t backtest it - that’s the final step, not the first.

2. Ask, “What would I see in the data if my idea has merit, and what would I see if it didn’t?”
   These two questions are highly underrated. They force you to think before you rationalise what you see in your data. They’ll act as reference points after the fact and highlight the exact gaps in your knowledge.

3. Deliberately connect new concepts to existing knowledge
   Ask yourself: “How is this similar to something I already understand? How is it different?” This is an effective way to gradually fill in the trading jigsaw puzzle one adjacent piece at a time.

4. Challenge your assumptions
   For each hypothesis, ask: “What assumptions am I making, and what would prove them wrong?” Another highly underrated question. Many assumptions are subtle - a common one is assuming persistence without gathering evidence for it.

5. Share ideas with other traders
   Explaining your thinking to others forces you to clarify your reasoning and exposes gaps in your logic.

I don’t consider my trader smarts to be exceptionally adept. I’ve worked with people who don’t seem even consciously to think through a problem - their highly developed trader brains do it automatically, just like you don’t consciously think about how to ride a bike.

That’s why point 5 above - sharing ideas with others, explaining your thinking, and clarifying your reasoning - is so underrated and important. It’s what allows average traders, like me, to stay on the narrow path, focus on what’s likely to work, and not bullshit ourselves with untested assumptions. This has been an unexpected but insanely valuable by-product of working with the RW Pro community.

The payoff is more than returns

The approach that compounds your learning doesn’t just make you a better trader. It makes trading more fulfilling.

You develop a deeper understanding of how markets (and humans) work. You start to see patterns across different markets, because you understand that it’s all just people buying and selling stuff. You build a powerful intuition. Your research starts to feel like muscle memory, even though you’re doing deep thinking and asking hard questions.

And perhaps most importantly, you gain confidence.

You swap the anxiety of someone clinging to a backtest that deep down they know is based on hope, for the real confidence that comes from understanding why things happen in the market and how to take advantage of them.