TREVOR · GHOSTNET LABS

The math behind the numbers

Textbook definitions. They describe how the figures on this page are computed, not how the engine decides anything.

ROE % on margin

ROE = \frac{realized result}{collateral posted at entry} \times 100

The return on the collateral a leveraged position ties up, as a percent.

R multiple

R = \frac{realized result}{risk planned at entry}

A result measured in units of what the trade stood to lose at its stop.

Win rate (decisive trades)

win rate = \frac{W}{W + L} \times 100

Wins over wins plus losses; flat trades count as neither.

Expectancy in R

E = \frac{1}{n} \sum_{i = 1}^{n} R_{i}

The average R per decisive trade: what one more trade is worth on average.

Profit factor (R)

PF = \frac{\sum_{R_{i} > 0} R_{i}}{| \sum_{R_{i} < 0} R_{i} |}

Total R won divided by the size of total R lost. Above one means the wins outweigh the losses.

Cumulative R

C_{t} = \sum_{i \leq t} R_{i}

The running total of R through each close date: the shape of the record, independent of position size.

Ranking setups: the Beta-posterior idea

θ_{k} \sim Beta (α_{0} + W_{k}, β_{0} + L_{k})

k^{*} = \underset{k}{arg max} {\tilde{θ}}_{k}

Each setup keeps a belief about its chance of winning, updated by its wins and losses. Thompson sampling draws one plausible value from each belief and tries the setup with the highest draw, so promising but little-tested setups still get chances while proven ones get most of them.