Read the evidence
Look beyond the equity curve. Make sense of backtests, live records and drawdown.
12 articles
A backtest is a simulation on past data; a live record is activity on a funded account. See why EA results differ and how to investigate a mismatch.
Since mid-2022 a Saturday moved 1.6% from high to low in the median, a weekday 3.4–3.7%. Bitcoin weekend data by day, by hour and for the CME gap.
Can a trading indicator make you rich? One 100% month is not proof: repeated for a year, $1,000 would become $4,096,000. Seven questions before buying.
Break-even trading costs for a backtest: in our 23,899-event sample, the mean reached zero at 7.7% of the assumed costs. Method, 12 datasets and code.
Equity drawdown includes open-position profit and loss; balance drawdown does not. Hypothetical example: equity falls 27.27% from its peak, balance 2.86%.
Fair value gap backtest: 23,899 FVG retest events averaged +0.020R before and −0.240R after assumed costs. Rules, 12 datasets, controls and Python code.
FX Blue grades where the data came from: verified, semi-verified or unverified. Myfxbook checks history and account control separately. Neither names the EA.
A coin-flip EA on XAUUSD made 18,393 USD on a 10,000 USD deposit in our backtest. Three benchmarks to run before you believe a gold EA result.
MT5 can backtest on a broker's real ticks with a changing spread; MT4 builds ticks from bars at one fixed spread. What that changes, and what 99% quality means.
We tested the previous day high/low breakout on gold, 2016–2026, with costs. It lost in every flat gold year and earned mainly with buys in the rally.
We rebuilt 33 figures of a real MT5 Strategy Tester report from its deal list. All match: Gross Profit contains commission and the balance graph counts deals.
Myfxbook verification covers trade history and account control, not which EA made the trades. Seven checks, a dated account example and a free worksheet.
