Backtesting
What do historical tests actually show?
Examine how defined rules behaved in selected past data without treating the output as a forecast.
A test built around stated assumptions
Biotrons configures a historical test for your stated instruments, timeframe and trading rules. We record the data period, assumed costs, order handling and any gaps that could change the interpretation. Initial setup commonly takes one to three weeks once suitable data and a complete rule set are available. You receive a reproducible configuration, a trade-level report and notes on the assumptions. We can explain the mechanics of a result, but a past pattern does not establish what future trading will produce.
Read a backtest in this order
Check the data period
Ask which market conditions appear in the sample and whether the instruments were actually available throughout it.
Inspect transaction costs
Commissions, spreads and slippage can change a simulation substantially, particularly when a rule trades frequently.
Look for missing observations
A clean chart can hide gaps in price data. Review how the test handles those gaps before interpreting its output.
Review individual trades
Trade-level records reveal implementation problems that a single summary number cannot show.
Where a backtest stops being useful
A backtest is a controlled examination of a rule set, not a rehearsal of every live market condition. Historical prices may not capture order-book depth, outages or the precise timing of a broker response. Repeatedly adjusting parameters until one sample looks attractive can also make the result less informative. We flag these limitations in the report and encourage a separate observation period before anyone considers live execution. The decision to trade remains yours.
A useful test makes its assumptions visible. If you cannot tell which data, costs and order rules produced a chart, you cannot tell what the chart is measuring.
Biotrons team
Start with the rule set
Tell us which instruments, timeframe and historical data you want to examine.
Ask about backtesting