Ask any trader what the most important rule is and they’ll say “risk management.” Ask them what their per-trade position size is and you’ll get a hesitation that tells you everything.
The asymmetry that kills accounts
A 50% drawdown requires a 100% gain to recover. A 75% drawdown requires a 300% gain. There is no symmetric maths in compounding — drawdowns are punished disproportionately by the geometry of returns.
The trader who risks 1% per trade can survive 20 consecutive losses with their account barely scratched. The trader who risks 5% per trade is functionally dead after the same 20 losses. Both might have the same edge; only one is still trading.
The Kelly answer (and why most people don’t use it)
The mathematically optimal position size is given by the Kelly criterion. For most retail edges, this works out to ~1–3% of capital per trade. Half-Kelly (0.5–1.5%) is the more practical version, because Kelly is brutally sensitive to edge mis-estimation. If you think your edge is bigger than it is — which you do — Kelly over-sizes you.
The emotional dependency
Here’s the part nobody writes about: the size of your position determines how emotional the trade becomes. A 5% position cannot be held through normal volatility — your nervous system won’t allow it. You’ll exit at the first uncomfortable wiggle. A 1% position, you can sit through.
Position sizing controls drawdowns. Emotional regulation controls position sizing. The order matters: shrink the size, and the emotion shrinks with it. The traders who survive long enough to compound are not the ones with the strongest will. They’re the ones who sized small enough that will wasn’t required.
The narrative that “AI has eaten trading” is half right. Machine learning beats humans decisively at certain things and loses to them at others. Knowing which is which is the difference between giving up edge and competing intelligently.
Where AI wins
- Latency-sensitive arbitrage. If your edge is measured in microseconds, you cannot compete with co-located algorithms. Don’t try.
- Cross-asset statistical patterns. Multi-factor models can ingest hundreds of correlated signals simultaneously. No human can.
- Sentiment aggregation. Reading 10,000 news articles and tweets per hour for tone is trivial for AI, impossible for you.
- High-frequency order-flow microstructure. The order book moves faster than human cognition.
Where humans still win
- Regime change. Models trained on regime A break the day regime B arrives. Humans notice “this feels different” before the numbers confirm it.
- Rare events. Tail-risk events are by definition under-represented in training data. Models underweight them; humans with memory of 2008 / 2020 / 2022 weight them appropriately.
- Narrative formation. Knowing why a move is happening (and whether the why has legs) is still a human skill.
- Time-frame patience. Most AI models optimise for short horizons. Position trades held for months still favour patient humans.
The compounding architecture
The best 2026 setup isn’t human-or-AI. It’s AI for the things AI does well (signal detection, pattern recognition, latency execution) wrapped by human discretion at the regime/narrative layer. Our indicators sit deliberately in this layer: machine-grade pattern detection, human-grade decision authority.
Every trader has read about herd behaviour. Few apply it correctly — because the framing makes the crowd sound like someone else. “They” panic at the bottom. “They” chase tops. The whole edge of crowd psychology is realising that you are part of the crowd, and you’ll feel exactly what they feel at exactly the moment they feel it.
The setup that catches everyone
A market trends up for weeks. You watch it from the sidelines, feeling foolish for not being in. The pullback comes — small, then medium, then sharp. You wait for it to bottom. Just as the fear gets loud enough that you decide “this is the buying opportunity,” the bottom is in. Or — worse — just as the relief rally gets exciting enough that you finally commit, that’s the bull trap.
The emotion that drove your decision is the same emotion that drove the exit liquidity into your position. By the time the feeling is strong enough to act on, the move is over.
The mechanical workaround
The solution isn’t to feel less. It’s to act before the feeling peaks — using rules that fire based on objective conditions, not the intensity of your conviction. Pre-commitment beats willpower every time.
This is why our indicators emphasise setup detection over signal interpretation. By the time you’re “sure,” the move is already mostly done. The setup is what you can act on while still uncertain.
Every trader has read about herd behaviour. Few apply it correctly — because the framing makes the crowd sound like someone else. “They” panic at the bottom. “They” chase tops. The whole edge of crowd psychology is realising that you are part of the crowd, and you’ll feel exactly what they feel at exactly the moment they feel it.
The setup that catches everyone
A market trends up for weeks. You watch it from the sidelines, feeling foolish for not being in. The pullback comes — small, then medium, then sharp. You wait for it to bottom. Just as the fear gets loud enough that you decide “this is the buying opportunity,” the bottom is in. Or — worse — just as the relief rally gets exciting enough that you finally commit, that’s the bull trap.
The emotion that drove your decision is the same emotion that drove the exit liquidity into your position. By the time the feeling is strong enough to act on, the move is over.
The mechanical workaround
The solution isn’t to feel less. It’s to act before the feeling peaks — using rules that fire based on objective conditions, not the intensity of your conviction. Pre-commitment beats willpower every time.
This is why our indicators emphasise setup detection over signal interpretation. By the time you’re “sure,” the move is already mostly done. The setup is what you can act on while still uncertain.
Every trader eventually faces the same realisation: the harder they try to predict the next move, the more often they’re wrong about it. Markets are non-stationary, news is endogenous, and human pattern-matching evolved to spot tigers in tall grass — not regime shifts in price.
The compounding advantage of process
A trading system doesn’t need to be right more than 50% of the time. It needs to be consistently applied, with positive expectancy over a meaningful sample. The boring truth is that most retail traders abandon profitable systems after three losing trades in a row — not because the system stopped working, but because their discomfort with drawdown exceeded their conviction in process.
What a system actually is
- A pre-defined entry condition expressed in objective price/volume terms.
- A pre-defined exit logic (profit target, trailing stop, time-based exit, or condition-based reversal).
- A risk management rule that caps per-trade and per-day loss.
- A trigger that is not subject to interpretation in the moment.
If any of those are missing, you have a strategy idea, not a system. Strategy ideas don’t compound; systems do.
Why this matters for the indicators we publish
Every system in our catalogue is built around this discipline. We won’t ship something that requires real-time human judgement to execute — that’s not a system, that’s discretion with extra steps. The tools work because they remove the place where humans fail: between the signal and the trade.
