TVMcp
TradingView Backtest vs Server-Side Backtest: What Actually Changes
Where results diverge, why they diverge, and how to compare both environments without fooling yourself.
Same script, different outcomes? Welcome to reality
You run one strategy in a chart and then on a server. Profit factor changes. Win rate changes. Your confidence changes. Usually downward.
This is normal. Two environments can interpret execution details differently unless you force strict parity conditions.
Most common divergence points
Different candle windows, symbol mapping, timezone assumptions, and order-fill logic are the usual suspects.
Even small differences compound over hundreds of trades. Tiny drift early often becomes huge drift later.
How to run a fair comparison
Use one fixed test case and compare trade-by-trade before comparing high-level metrics. Metrics can match by accident while execution logic still differs.
A clean parity check is boring, methodical, and absolutely worth it.
Operational rule for teams
Treat parity drift as a failing test. Do not proceed to optimization while parity is unresolved.
Scale comes after correctness. Always.