Notes on
disciplined trading.
How the systems work — the mechanics, the failure modes, the maths. Never what to buy. Nothing here is investment advice.
Fees, funding and slippage: the three costs a backtest forgets
A rule that looks fine on a chart can be quietly eaten by costs the chart never shows. A plain-language look at trading fees, funding and slippage, with the arithmetic to check your own numbers. Educational only.
Read itLeverage is a position-size decision wearing a costume
10x leverage does not make a trade ten times better. It makes the position ten times larger relative to your margin, and moves the point of forced closure much closer. A plain-language explainer. Educational only.
Drawdown: the number that decides whether you are still here next year
A 50% loss needs a 100% gain to recover. A plain-language look at drawdown, why the recovery math is lopsided, and how to decide your limit before the market decides it for you. Educational only.
A backtest without a control is a story, not evidence
In a rising market almost any rule looks clever. A plain-language look at why every backtest needs a boring baseline beside it, and how to read the two together. Educational only.
Practice profits are not real profits: why we keep testnet and live separate
Testnet and paper trading let you learn a strategy without risking capital. Folding those results into your live record is how traders quietly mislead themselves. A look at why practice fills flatter, and why Stralines keeps the two ledgers structurally apart. Educational only.
Is it safe to give a trading bot your API key?
The honest answer depends entirely on what the key is allowed to do. A plain-language guide to exchange API scopes, the four questions worth asking any platform, what a trade-scoped key can still cost you, and how to revoke access in under a minute. Educational only.
Why your stop-loss disappears — and what a system does about it
Exchanges drop, reject and auto-cancel stop-loss orders more often than most traders realise. A plain-language look at the failure modes, why placing a stop is not the same as having one, and what a three-layer recovery system can and cannot fix. Educational only.
How automated trading bots work — and what they can't do
A software-engineering view of automated trading bots: the signal-to-execution pipeline, where things fail, and the limits no bot can engineer away. Educational only.
What is Smart Money Concepts (SMC)?
A plain-language explainer of Smart Money Concepts — the vocabulary, the chart structures, and what the framework does and does not claim. Educational only.
AI-agent trading vs bot trading: what each gives you, what each costs you
A balanced look at two trading architectures — autonomous AI agents and deterministic bots — with the real strengths and trade-offs of each, and a way to decide which one fits how you actually trade. Educational only.
Why a 35% win rate can still be net-positive
A plain-language walkthrough of trade expectancy — how a strategy that loses 65% of the time can still finish ahead, and why "win rate" alone is the wrong number to fixate on. Educational only.
Position sizing and risk caps: the boring math that keeps you in the game
Survival in trading is decided less by your entries than by how much you risk per trade. A plain-English look at position sizing, drawdown math, and the risk caps that enforce it. Educational only.
The 3 mistakes that blow up manual crypto accounts (and how rules fix them)
Most blown crypto accounts don't die from a bad call — they die from three operational mistakes that repeat under pressure. Here's what they are and how rule-based automation removes them. Educational only.
Why the shift from manual to algo trading is essential in 2026
Markets in 2026 move faster, run around the clock, and punish hesitation. This is a software-engineering case for moving from manual execution to rule-based automation — for discipline and consistency, not predictions. Educational only.
Reading is one thing.
Watching is another.
Replay a DCA plan over real prices — no signup.