Test a trading idea before you risk money on it.
Pick a stock, a simple rule and a stretch of years. Finertia shows what would have happened — and then checks whether that result was skill or luck.
How it works
Three steps. About a minute.
Pick
A stock, a simple rule, and a stretch of years.
AAPL · momentum · 2019–2024
Run
Finertia replays the rule on real prices, with real trading costs, and shows what it would have made.
+68% · just holding made +408%
Check
Then it re-tests the same rule on years it was never tuned on, and tells you if the result was skill or luck.
2 of 6 checks passed
A real run
This is what a result looks like.
A momentum rule on Apple, year by year, against simply holding the stock. It beat holding in one year out of five — which is exactly the kind of thing you want to know before trading it.
Beat buy & hold in 1 year of 5 — the one the market fell. Over the full period the strategy returned +68.4% against +408% for holding the stock. That is the engine's real output, not an illustration.
Lost to holdingOpen the full example, with every chart
Read the write-up: the best-looking strategy was the worst on new data
The strategies
Three ordinary rules, written out exactly.
Trend following and mean reversion, with every parameter you can change and how each did on data it was never tuned on, including when it failed.
Trend following
Momentum
Buys a stock that has been rising and sits above its moving average; sells it short on the mirror condition.
On AAPL: failed, then held up
The rule and its results: MomentumTrend following
MACD
Long while the MACD line is above its signal line, short while it is below: a smoothed trend signal.
On AAPL: failed, then failed
The rule and its results: MACDMean reversion
Bollinger
Buys when the price falls below its lower band and shorts when it rises above the upper one: it fades extremes.
On AAPL: held up, then failed
The rule and its results: BollingerWho it's for
Start where you are.
New to this?
Start from the example run. Every number has a ? beside it that says what it means and whether it matters.
Open the exampleAlready trade?
Stop trusting a curve you fitted yourself. See whether your rule holds on data it never saw before you put money on it.
Test a ruleLearning quant?
No black box. The engine is plain pandas and numpy, every check has its formula written out, and over 700 tests keep it honest.
Read how it worksWhy the checks matter
Any rule looks good on the years it was tuned on.
So Finertia tunes on the first 70% of the period and scores on the last 30%, untouched. Every figure from that stretch is in green, like this, so you always know which numbers were actually checked.
Real prices
Daily closes from Yahoo Finance, with a trading cost on every position change.
Open maths
No backtesting library. Plain pandas and numpy you can read, with over 700 tests on the engine.
Honest numbers
Every metric carries a confidence interval, and results that beat holding are the exception, not the sales pitch.
Try one idea today.
Free plan runs the full engine on any stock. No card, nothing to install.