Here's a sentence that would have been science fiction fifteen years ago: You can now describe a trading strategy in plain English and have a working, backtested system running on your computer within an hour. No coding degree. No quant background. No expensive data subscriptions.
That sentence changes everything about this chapter. Every trading book written before 2023 treated algorithmic trading as an advanced topic — something for maths PhDs, hedge fund engineers, and people who dream in Python. The barrier to entry was enormous.
That barrier has collapsed. It didn't lower gradually. It collapsed, almost overnight, when large language models became capable enough to write, debug, and explain code at a level that previously required a professional developer. This is the single biggest democratisation event in the history of retail trading.
But before we get to the revolution, you need to understand what algorithmic trading actually is — because the term is surrounded by more mythology than almost anything else in finance.
Demystifying “Algos”
An algorithm is just a set of rules. That's it. If X happens, do Y.
If the 20-day moving average crosses above the 50-day, buy. If the price drops below yesterday's low on above-average volume, sell. If the macro regime shifts from Regime 1 to Regime 3, reduce equity exposure.
You're already thinking algorithmically if you've been reading this book. The Grade A system? That's an algorithm. The opening range breakout? Algorithm. The swing trading entry rules? Algorithm.
The difference between what you've been doing and what quant funds do is simply this: they write their rules in code, so a computer can execute them automatically, across thousands of assets, millions of times per day. You write your rules in your head and execute them manually across a handful of positions.
Neither approach is inherently better. What matters is whether the rules are good.
How Institutional Algorithms Work
Understanding how the big players use algorithms isn't just academic — it directly affects your trades every single day.
Every time an ETF rebalances, algorithms execute the trades. If a stock gets added to a major index, the ETFs must buy millions of shares. These aren't discretionary decisions. They're mechanical, forced, and predictable.
This creates predictable patterns. Stocks being added to indices tend to rise before the inclusion date. Stocks being removed tend to fall. The moves have nothing to do with the underlying business.
Quantitative funds run factor models — algorithms that rank stocks by momentum, value, quality, size, or volatility. When a stock's momentum score crosses a threshold, the algorithm buys. No human reviews the trade. No analyst reads the earnings report. The machine acts on the numbers alone.
This is why momentum works so reliably as a factor. Billions of dollars of algorithmic capital is programmed to buy things with strong momentum. The factor creates a self-reinforcing loop.
Modern market making is entirely automated. Algorithms post bids and asks, manage inventory, adjust spreads, and hedge positions thousands of times per second. A human couldn't do this job even if they wanted to.
These algorithms are the ocean you're swimming in. Every time you buy a stock, a market-making algorithm is probably on the other side. It's providing liquidity but extracting a tiny fee (the spread) on every transaction.
When a pension fund wants to buy 2 million shares, it feeds the order to an execution algorithm that slices it into hundreds of small pieces dripped into the market over hours or days. This is why you see the iceberg orders and steady institutional accumulation patterns from Chapter 14 — those are the footprint of execution algorithms.
When you learn to read those patterns, you're essentially reverse-engineering what the institutional algorithms are doing.
The 80% Reality
This has profound implications. Price moves in the short term are dominated by algorithmic flows, not human sentiment. Chart patterns that worked thirty years ago — when markets were dominated by human traders reading the same textbooks — have become less reliable. The algorithms don't read candlestick patterns. They read order flow, statistical relationships, and factor signals.
But here's the opportunity: algorithmic flows are predictable in ways that human behaviour isn't. ETF rebalancing happens on a schedule. Factor rotations follow measurable signals. Execution algorithms leave readable footprints. If you understand what the machines are doing, you can position yourself to benefit from their actions.
The Revolution That Changed Everything
Until recently, the gap between understanding algorithmic trading and doing algorithmic trading was enormous. You could read about factor models all day, but unless you could write Python, build data pipelines, and debug code at 2am, you couldn't build anything.
Large language models demolished that barrier.
Here's what you can now do with zero coding experience, using nothing more than a conversation with an AI assistant:
Scan all S&P 500 stocks for specific moving average crossovers, RSI levels, and momentum thresholds. Working code, not pseudo-code.
Test your entry and exit rules against 10 years of data. Get total return, max drawdown, win rate, and average gain per trade — visualised in charts.
Portfolio positions, sector exposure, drawdown tracking, and alerts if any position loses more than 3%. A tool that used to cost five figures.
Download latest GDP, CPI, and unemployment data. Automatically categorise the current environment into one of four macro regimes.
Each of these tasks used to require hiring someone or spending months learning to code. Now they require a clearly worded prompt and a few minutes of iteration. The cost went from thousands of pounds and hundreds of hours to essentially zero. And if you'd rather use a ready-made system while you build your own, you can try our signals free for 14 days.
What This Means For You
Had been swing trading for three years using manual charts and screening. Profitable at ~15% per year, but the process was tedious — 45 minutes every morning scanning charts, updating entries, checking positions.
In early 2024, he started using an AI assistant to automate his workflow. Over a series of conversations, he built a system that scanned 3,000 stocks nightly, ranked them by his grade system, generated entry/exit levels, and emailed him the top 5 setups each morning.
He couldn't write a single line of code. Every script was generated through conversation with the AI. Iterative. Conversational. Like working with a developer who never gets tired.
Result: morning routine dropped from 45 to 5 minutes. Annual return improved to 23% because he was seeing more opportunities and reacting faster to grade changes.
Not coding. Communicating with AI clearly enough that it can code for you.
The ability to describe your trading idea in precise, unambiguous language is now more valuable than the ability to write Python.
What AI Can and Cannot Do
Taking your clearly defined rules and turning them into working systems. Speed, code, tireless execution.
Deciding which rules to use, when to override them, interpreting ambiguous market conditions.
This is why the combination of human and AI is so powerful and why pure AI trading systems consistently disappoint. The human provides the macro awareness, the market intuition, and the judgment calls. The AI provides the speed, the code, and the tireless execution of defined rules. Together, you get something neither can achieve alone. Our free tools and calculators are a good starting point for building your own systematic edge.
I've seen traders ask an AI to "tell me the best strategy" or "predict where the market is going tomorrow." The AI will give you an answer — but it will be generic, backward-looking, and potentially dangerous.
AI doesn't have market intuition. It has pattern matching on historical data. The smartest traders treat AI the way a surgeon treats a robot-assisted scalpel — the machine has superior precision, but the human directs every cut.
The next chapter shows you exactly how to build your first automated system — step by step, using AI as your coding partner.
- 1.An algorithm is just a set of rules — if you've been following this book's system, you're already thinking algorithmically.
- 2.AI has demolished the barrier to building trading tools — you can now describe a strategy in plain English and get working code.
- 3.Use AI for implementation (screening, backtesting, monitoring) but keep judgment (macro, context, final execution) with the human.
This content is for educational purposes only and does not constitute investment advice. Trading and investing involve substantial risk of loss. Past performance is not indicative of future results. Always do your own research and consider seeking professional guidance before making financial decisions.