Learn

Short articles about how automated trading actually works — written for someone who is good at their job, curious, and has never placed a trade. No jargon without an explanation, and a worked example in every one.

Concepts

• Getting started — From your first sign-in to your first approved paper trade, in four steps and about fifteen minutes — none of which can touch real money.
• Holding a stock through its earnings report — Why a scheduled report is a different kind of risk from an ordinary bad day, why a stop doesn't cover it, and the three honest ways to handle one.
• How much can I lose? — The arithmetic of losing, why recovering is harder than losing, and how to work out the worst case before you're living in it.
• Letting the engine trade for you — How much freedom to give the software, why approvals expire, and the difference between halting and stopping.
• Look-ahead bias — When a backtest accidentally uses information that wasn't available yet — the most common reason a great result turns into a losing strategy.
• Market orders and limit orders — The two ways to ask for a trade — one guarantees it happens, the other guarantees the price, and you can't have both.
• Overfitting, and how to test around it — Tuning a rule until it fits the past so well that it has learned the noise instead of the pattern — and what to do instead.
• Paper trading, and how long to do it — Running a strategy against real prices with imaginary money — the only test of a trading idea that can't be fooled.
• Position sizing — Deciding how much to put into each trade — which affects your results far more than deciding what to buy.
• Reading a decision record — Every time the engine looks at a symbol it writes down what it saw and what it concluded — including the overwhelming majority of times it decided to do nothing.
• The spread and slippage — The small, invisible costs on every trade that no backtest charges you for — and the reason frequent trading is harder than it looks.
• What a backtest is, and why yours looks better than it is — Running a strategy over history to see what it would have done — a useful sanity check, and the single easiest thing to fool yourself with.
• What a strategy actually is — A written-down rule for when to buy and when to sell, applied the same way every time — including on the days you'd rather not.
• What an edge is — The small, repeatable advantage that makes a strategy worth running — and an honest look at why most ideas don't have one.
• What the optimizer actually does — An AI rewrites your strategy's rule, tests the new version on data it wasn't tuned against, and keeps it only if it survives.
• When a strategy stops working — Telling the difference between a normal bad patch and a strategy that has genuinely stopped working — and what to do about each.
• Why win rate misleads — How often you're right matters much less than how much you make when you're right — and the two often pull in opposite directions.

Glossary

• Benchmark — The ordinary thing you measure a result against — usually a fund that tracks the whole market — so you can tell skill from a day when everything went up.
• Blackout — A period when you deliberately don't open new positions in a stock — most often the few days around its earnings report.
• Drawdown — How far an account fell from its best point before recovering — the dip you had to live through, rather than where you ended up.
• Earnings date — The day a company publishes its quarterly results — scheduled weeks ahead, usually released while the market is closed, and the most common reason a stock opens far from where it closed.
• Excess return — A stock's return minus what the market did over the same days, so a rising tide doesn't get counted as skill.
• FOMC — The Federal Reserve committee that sets US interest rates. Its decisions come eight times a year at 2 pm New York time, and they move nearly every stock at once.
• Gap — The jump between one day's closing price and the next day's opening price, which happens while the market is shut and nobody can trade.
• Paper account — BotStock's own simulated brokerage account, where every new account starts — real prices, practice money, no brokerage needed.
• Paper trial — A new version of a strategy trading with simulated money beside the version that's running, so the two can be compared on the same days.
• Protective stop — A standing instruction to sell if a position falls to a price you chose, so that a disappointing position can't quietly become a disastrous one.
• Resting order — An order that's been accepted but hasn't filled yet, because the price it needs hasn't arrived — a limit waiting for its price, or a stop waiting to trigger.
• Sample size — How many observations a figure was calculated from — the number that decides how much the figure is worth.
• Setup — The conditions a rule needs lining up before it will trade — a setup is forming when they're most of the way there and formed when the rule fires.
• Tape — A running record of what happened, in the order it happened — originally the paper strip a ticker machine printed prices on, now any live feed of trades and decisions.
• Watchlist — A named list of stocks you've decided are worth paying attention to — the pool a trading rule is allowed to pick from, not a list of things you own.