Blueprint Engines Explained

Blueprint's power comes from five specialized analysis engines that work together to produce a complete picture of your trading edge. Each engine focuses on a different aspect of your performance and uses proven statistical methods to generate reliable insights.

Engine 1: Edge Stability

The Edge Stability engine measures how consistent your edge is. It computes a Sharpe-like stability ratio (your average R divided by the standard deviation of R) and classifies your edge as Stable or Volatile -- the result is shown under Analytics > Smart Analysis. Your core risk profile (the list below) is calculated by Blueprint's separate Tier-1 risk formulas, which were developed through consultation with three independent AI systems (ChatGPT, Gemini, and DeepSeek) and use a hybrid drawdown-aware approach rather than the Kelly Criterion, which all three AIs agreed is dangerous for retail traders.

• Risk per trade percentage (clamped between 0.25% and 2.0%)
• Fixed vs. compounding decision tree based on account type, profit factor, drawdown, and consistency
• Daily loss limit (2R to 5R based on your trading frequency and worst-case scenarios)
• Weekly loss limit (4R to 10R, scaled from daily limit)
• Breakeven tolerance recommendation based on your average fee impact

Engine 2: Interaction Discovery (Bayesian Synergy)

This is the heart of Blueprint. The Interaction Discovery engine groups your trades by core combinations (symbol + direction + timeframe) and calculates cross-dimensional metrics for each combo (a combo needs at least 8 trades to appear). It uses a Bayesian win rate adjustment -- a Beta(1,1) prior, which is the same as adding one phantom win and one phantom loss -- inside the synergy and grading math to reduce small-sample bias, so a 3-for-3 combo is not graded as if a 100% win rate were proven. The Win Rate displayed on each setup card is your actual raw win rate.

Each qualifying combo becomes a graded setup card showing the combination's own dimensions (symbol, direction, timeframe, plus session or strategy when they are part of the combo) together with its key metrics: Win Rate, Expectancy, Profit Factor, Synergy, the trade count behind it, and a credibility interval for its win rate.

Engine 3: Behavioral Diagnostics (Markov)

The Behavioral Diagnostics engine applies Markov chain analysis to your win/loss sequences: it builds a transition matrix (for example, your probability of winning after a loss), detects edge decay, and tests whether your streaks are random. It unlocks at Tier 3 (60+ trades) and its results are shown under Analytics > Smart Analysis. Separately, a psychological-patterns module unlocks at Tier 5 (200+ trades) and flags revenge trading (re-entering within 15 minutes after a loss), tilt (continuing to trade after 3+ consecutive losses — flagged as harmful only when the win rate after such streaks measurably drops below your baseline), and overtrading (days with double your average trade count).

Engine 4: Monte Carlo Risk Simulator

The Monte Carlo engine runs simulations on your trade history to estimate the probability of various drawdown scenarios. It sweeps a range of risk levels -- estimating median growth, worst-case (5th percentile) outcomes, and the probability of exceeding drawdown thresholds -- and computes the analytic estimate of the probability that your drawdown exceeds its limit, plus the optimal risk level, validating the recommended risk profile. This provides a reality check: even your best setups will have losing streaks, and this engine shows you the likely range.

Engine 5: Edge Evolution (Change-Point Detection)

The Edge Evolution engine applies change-point detection to your performance over time. It identifies when your edge improved or deteriorated, helping you understand whether a recent winning streak is a genuine improvement or just variance. Recency weighting is applied to the core risk-profile metrics: trades from the last 90 days carry full weight (a factor of 1.0), while older trades are down-weighted (a factor of 0.43). The analysis engines instead focus on your most recent performance regime via change-point filtering, so your recent behavior counts for more than stale history either way.