How these numbers are produced

A collector runs on the same machine as each strategy's MetaTrader 5 terminal. It reads the terminal's own deal history through the MT5 API — the same records the broker statement is built from — and pushes closed positions and end-of-day balances to this site over an authenticated channel.

Nothing is entered by hand, and the collector has no write access to the trading account. If the terminal is not running, the strategy is marked offline rather than silently frozen on its last good figure.

Each MetaTrader login produces a new session key. When the collector reports a key we have not seen before — a restart, a migration, a change of account — the strategy's entire history is rebuilt from the terminal rather than appended to.

That rule is deliberately blunt. It makes it impossible to end up with two accounts' trades stitched into one track record, which is the most common way published performance quietly becomes fiction.

A rebuild is assembled out of sight and published only once it is complete, so the figures on this page are never a partially reloaded history. If a rebuild fails part way through, nothing changes at all.

Where daily balance snapshots are available, return is time-weighted: each day's return is measured against the previous close adjusted for that day's deposits and withdrawals, and the daily figures are then compounded. Adding capital to an account does not flatter its percentage return.

Where snapshots are not yet available, the equity curve is reconstructed from the strategy's baseline balance plus cumulative net profit. Net profit is always after commission and swap.

Maximum drawdown is the largest peak-to-trough decline of the equity curve inside the selected window, expressed as a percentage of the peak. It is measured on closed-position equity, so it reflects realised pain rather than the worst intraday tick.

The risk band blends annualised volatility of daily returns with peak drawdown onto a 1–10 scale, where 1 is the calmest. It is intentionally coarse: a single-digit band is an honest summary, whereas a decimal would imply precision that does not exist.

The six radar axes — win rate, return, stability, drawdown control, active time and consistency — are each scored from 0 to 100 off the closed-trade record for the selected window.

Every axis is scaled down when the sample is thin. A strategy with twenty trades cannot paint a full hexagon no matter how well those twenty went.

Back-tests are not published here. Neither are demo accounts, hypothetical fills, or windows chosen after the fact — the window selector moves the view, not the underlying record.

Copier counts, assets under management and fee terms are published by MMT rather than measured from the terminal, and are labelled as such. Where a figure has not been published it reads NIL rather than zero.