One map of the whole platform.

1. 1Search Signal
   Finds the moment

   Decides buy or sell on every bar. Its exit takes two booleans — raising both emits two independent signals.

2. 2Filters
   Guard the gate

   A chain of filter documents interrogates each signal. Every filter answers with one boolean: pass, or stop it here.

3. 3Create Order
   Shapes the trade

   Turns a surviving signal into price levels: one or more take-profits, one or more stop-losses, and an optional confidence that scales position size.

4. 4Deal Stage
   Manages the position

   After execution: qualification models decide whether a management suggestion is allowed, sizing models decide how big. Prop-firm presets included.

Platform terms

• Tool

  The smallest unit of computation — one canvas node type from the catalog.

• Mode

  A named variant of one tool, chosen in its configuration — one node, several behaviours.

• Wire

  A typed connection between two ports. Types must agree, or an adapter bridges them.

• Interpretor

  A tool that reads raw indicator output and answers a semantic question with a verdict.

• Document

  One authored artifact — a search signal, filter, create order, deal stage or indicator — stored as a graph or typed configuration.

• Draft

  Your single mutable working copy of a document. Publishing turns it into a version.

• Version

  An immutable published snapshot of a document. Runs and compiles always execute versions.

• Search signal

  The stage document that decides buy or sell on every bar.

• Filter

  A stage document that gates signals with one boolean answer; its category decides cancel versus hold.

• Create order

  The stage document that shapes a signal into take-profit, stop-loss and confidence.

• Deal stage

  The post-execution stage: qualification and sizing models managing the open position.

• Strategy

  The composed assembly of the four stages, published as a version.

• Run

  One backtest or optimization executing a published version against history.

• Robustness score

  A 0–100 composite from six validation analyzers, computed after every optimization.

• Robot

  A compiled Expert Advisor file produced from a strategy version.

• Indicator document

  The document family that compiles to a custom MetaTrader indicator; its graph ends in sinks.

• Sink

  A terminal indicator-lane node: a plot, an alert, or a chart object.

• Cell

  A reusable sub-graph, shared by link and referenced inside filters.

• Journey data

  The per-signal data bus stages use to pass values forward.

• Credits

  The prepaid balance that meters compute. Runs hold a worst-case amount, then settle to actual.

• Entitlement

  What a marketplace purchase grants: run access to a sealed product, or a private clone of an open one.

• Repainting

  When a tool's past values change as new bars arrive. The engine re-evaluates such tools bar by bar so backtests stay honest.

Report metrics

• Initial deposit

  Starting equity used to open the backtest. Every profit-and-loss figure on the report is measured relative to this number.

• Final balance

  Where the account ended up at the last bar of the run — starting equity plus the sum of every closed trade.

• Total net profit

  Sum of every closed trade over the whole run. Positive means the strategy made money; negative means it lost money.

• Gross profit

  How much the winning trades alone made, before any losers are netted out. Always non-negative.

• Gross loss

  How much the losing trades alone gave back, shown as a positive magnitude so it lines up with gross profit side-by-side.

• Profit factor

  Dollars earned per dollar lost. Values above 1 are profitable; a value of 2 means the strategy makes two dollars for every one it gives back.

• Expected payoff

  Statistically expected dollar outcome of the next trade — literally the average result per trade so far. Think of it as the house edge of the strategy.

• Recovery factor

  How much profit was produced for every dollar of the worst drawdown. High numbers mean the strategy digs itself out of holes quickly relative to the money it makes.

• Calmar ratio

  Return versus the largest peak-to-trough hit — a risk-adjusted return measure that uses worst-case drawdown instead of volatility. In this implementation it shares the formula of recovery factor.

• Sharpe ratio

  Classic risk-adjusted return — average return divided by its standard deviation, annualised. It assumes a zero risk-free rate. Higher values mean the strategy earns more per unit of volatility.

• Sortino ratio

  Like Sharpe but only punishes downside moves — upside volatility is treated as a feature, not a bug. Fairer for strategies with asymmetric, skewed payoffs.

• Ulcer index

  Downside-only risk score — the root-mean-square of the percentage drawdowns from every running peak. Large drawdowns hurt the score proportionally more than small ones, so it reflects the pain of holding the strategy through bad stretches (Peter Martin, 1987).

• Balance drawdown absolute

  Absolute drawdown — how far below the starting equity the account ever dipped, measured in account currency. It only counts time spent underwater relative to day one.

• Balance drawdown maximal

  Maximal drawdown — the biggest peak-to-trough drop on the closed-trade balance curve, in account currency. Captures the worst streak of losses regardless of where in the run it occurred.

• Balance drawdown relative

  Worst percentage drawdown at any peak on the balance curve. Unlike the maximal dollar version, this picks out the steepest relative hit even if it happened early when the account was small.

• Equity drawdown absolute

  How far below the starting equity the account ever went when floating open-position P&L is included. It can be worse than the balance-curve version because unrealised losses count.

• Equity drawdown maximal

  The largest peak-to-trough drop measured on the equity curve (balance plus open-position P&L). Matters more than the balance version for strategies that hold positions through storms.

• Equity drawdown relative

  Worst percentage drawdown measured against any running peak on the equity curve — the single worst bar-to-bar equity pullback in the whole run, in percentage terms.

• Max drawdown

  The headline worst-case drawdown percentage — how much of the peak equity was given back at the worst point during the run. Prop-firm evaluators care about this number a lot.

• Max drawdown duration

  Longest stretch, measured in bars, during which the equity curve stayed below its most recent peak without making a new high — i.e. the longest time spent underwater.

• Total trades

  How many trades closed during the backtest window. Statistical significance of the other metrics rises with this number.

• Winning trades

  Number of trades that closed with a profit strictly greater than zero.

• Losing trades

  Number of trades that closed flat or at a loss. Break-even trades are counted on this side, not the winning side.

• Win rate

  Share of trades that closed profitably. On its own it tells you almost nothing — a 90% win-rate can still lose money if the occasional loser is huge.

• Long trades (won %)

  How many buy-side trades were taken, with the share of those that were profitable shown in parentheses.

• Short trades (won %)

  How many sell-side trades were taken, with the share of those that were profitable shown in parentheses.

• Largest profit

  Biggest single profitable trade of the whole run. Useful for spotting whether the net result is driven by one lucky outlier.

• Largest loss

  Biggest single losing trade of the run, shown as a negative value. A realistic stop-loss should keep this within your risk budget.

• Average profit

  Mean dollar size of a winning trade. Together with the losing counterpart it tells you the payoff ratio of the strategy.

• Average loss

  Mean dollar size of a losing trade (negative magnitude). Compare with the average winner to see whether your winners are bigger than your losers.

• Max consecutive wins

  Longest run of back-to-back profitable trades. Long winning streaks can signal trend-following behaviour — or, in a thin sample, just luck.

• Max consecutive losses

  Longest run of back-to-back losing trades. Matters a lot for position sizing — if this run is 8, could the account survive 8 losers at your risk-per-trade setting?

• Max consecutive profit

  Dollar sum of the best winning streak. Useful for calibrating realistic upside: this is what a run of good luck actually felt like.

• Max consecutive loss

  Dollar sum of the worst losing streak (negative). A key stress figure for risk-of-ruin: can the account absorb this and keep going?

• Avg consecutive wins

  How long, on average, winning streaks last. Higher numbers hint at momentum/trend behaviour in the edge.

• Avg consecutive losses

  How long, on average, losing streaks last. Should shape how aggressively you scale position size after a loss.

• Average holding

  Mean number of bars a position stayed open. Combined with the timeframe it tells you the typical holding horizon of the strategy.

• Min holding

  Shortest position duration (in bars). Very small values often mean SL/TP levels are tight relative to bar size or the logic is scalping.

• Max holding

  Longest position duration (in bars). Useful for catching trades that overstayed — might flag a missing exit rule or a timeout opportunity.

• Z-score

  Run-test on the win/loss sequence (MetaTrader-5 definition). Positive z means wins tend to follow wins and losses follow losses — the sequence has momentum. Negative z means the sequence is more alternating than random. Values with magnitude greater than 1.96 are statistically significant at the 5% level.

• LR correlation

  Pearson correlation between the balance curve and a straight line drawn against bar index. Values near 1.0 mean a remarkably steady, straight-line equity curve; values near 0 mean the curve is jagged and uneven.

• LR standard error

  Standard deviation of the balance curve around its straight-line fit. Smaller means the account grew in a tighter, more predictable corridor around the trendline.

• Average MAE

  Average Maximum Adverse Excursion — the typical worst unrealised loss a trade had to sit through before closing, converted to account currency (John Sweeney, "Maximum Adverse Excursion").

• Average MFE

  Average Maximum Favourable Excursion — the typical best unrealised profit a trade reached before closing, in account currency. If MFE is much bigger than the actual win, trade management is leaving money on the table.

• Correlation (profit, MAE)

  Pearson correlation between the final profit of a trade and how far the trade went against it first. A strongly negative value means trades that dipped deep tended to close as losers — stops are probably too tight.

• Correlation (profit, MFE)

  Pearson correlation between the final profit of a trade and how far it ran in its favour. Strongly positive means the big winners did, in fact, run far — suggests take-profits could be wider or replaced with a trail.

• Correlation (MAE, MFE)

  Pearson correlation between how far a trade went against it and how far it ran in its favour. High values mean volatile trades just move a lot in both directions — the strategy is exposed to noise as much as edge.

Validation metrics

• PBO

  Probability of Backtest Overfitting. The fraction of in-sample/out-of-sample splits where the in-sample winner underperforms out of sample. Below 50% is the acceptable range.

• DSR p-value

  Deflated Sharpe Ratio significance, corrected for multiple testing and higher moments. Above 0.95 means the strategy beats the null at the 5% level.

• MC p-value

  Monte Carlo trade-shuffle test. How often a random permutation of the trades produces a Sharpe ratio as good as the real sequence.

• SPA p-value

  Hansen’s Superior Predictive Ability test. Whether the strategy beats a benchmark once data-snooping bias is accounted for.

• WFE mean

  Walk-Forward Efficiency. The ratio of out-of-sample to in-sample performance, averaged across windows. Values near 1.0 mean no degradation.

• Param drift

  The mean coefficient of variation of the best parameters across walk-forward windows. High drift means the strategy has no stable optimum.