Methodology & data sources

Price data

OANDA (via TradingView): H1 bars for gold and the yen pairs, 2005–2026, used in both the TradingView Strategy Tester and the Python engine
Dukascopy: real 1-minute bid/ask data for Quant Prototype (2019–2026) and News Radar (2013–2026)

Two ways of testing

TradingView Strategy Tester: the real script runs on TradingView with Bar Replay window by window, and the trade lists are joined. Used for Gold Swing, Gold H1, Gold Breakout and Trend Rider.
Python engine: the same rules rewritten in Python and checked against TradingView until entries match 99.7%. Used for Ichimoku MTF because it can run 20 years continuously like the real bot (TradingView has to restart every window).

The two methods give slightly different numbers, so we only ever compare within the same method.

How the rating works

The rating (0–5 stars) is built to be fair across systems that use different risk per trade, so it ignores raw return and uses ratios instead:

30% Profit factor — PF 1.0 scores zero, PF 2.0 or more scores full
30% Return ÷ max drawdown — 4 or more scores full
20% Consistency — no losing year scores full; each losing year costs points relative to the test length
20% Sample size — 1,000 trades or more scores full (fewer trades = less reliable statistics)

Grades: 4.5+ Excellent · 3.5+ Good · 2.5+ Fair · below that Experimental. “#1” marks the top-rated system in each group, “best” on a version marks the highest-rated version of that system, and “Recommended” is our own pick with the reason shown on the card.

What each number means

Avg. return per year (compounded): position size follows the latest balance and each system's own risk per trade. Risk differs between systems, so returns cannot be compared directly across systems.
Max drawdown: the largest fall from a previous high, measured on closed-trade balance
Profit factor (PF): gross profit of winners ÷ gross loss of losers. Above 1 means profitable.
Longest losing streak: the most consecutive losing trades, ordered by exit time
R: a trade's result relative to the risk taken, e.g. +2R means twice the amount risked

Entry/exit charts on the trades pages

Charts are drawn from the real prices of the same source used for testing. Arrows mark entries, circles mark exits, dashed lines are SL/TP when available. For systems tested on TradingView, the entry price is the open of the entry bar.

Limits of backtests

Backtests do not include every real cost, such as wide news spreads, slippage and swap
Systems were developed on past data. Even after many checks, future results may be lower.
Statistics verified by a live run are on the Live page, which is still short