Research,
not signals.

Measures the underlying business — profitability, financial health, and consistency.

For stocks, the score reflects revenue growth, EPS, profit margin, debt-to-equity, free cash flow, return on equity, and a market-cap stability proxy. For ETFs, the inputs collapse to expense ratio, AUM, and tracking quality. Each input is bucketed into discrete points (e.g. revenue growth above 10% earns more than 5–10%), then summed and normalised to 0–100.

Measures the rate and consistency of business expansion.

Growth combines historical revenue growth, EPS growth, forward revenue guidance, and the trend in analyst estimate revisions. The score is asymmetric: an estimate revision moving down subtracts points, while a revision moving up adds them. The cap of 100 ensures a single explosive metric cannot dominate.

Measures whether the current price offers a reasonable entry relative to fundamentals.

A high valuation does not automatically remove a candidate from research. It lowers the valuation score and lifts the risk score; a high-quality compounder with a stretched P/E may still rank well overall. Inputs include trailing P/E, forward P/E, price/sales, PEG, sector median P/E, and dividend yield. ETFs use expense ratio as the primary valuation signal.

Measures whether current price action and trend support a reasonable entry.

Technical setup uses moving-average alignment (price vs SMA 50 / SMA 200 / golden cross), RSI in a healthy 45–65 band, MACD direction, volume trend, and distance from the 52-week high. Extreme RSI readings score poorly in either direction — overbought conditions earn fewer points than balanced trends.

Measures recent news tone and proximity to material event risk.

The v1 scorer uses a simplified bucket — predominantly positive recent news (80), mixed or neutral (50), predominantly negative (25). Subtractions apply for upcoming earnings within 48 hours (−20) and within seven days (−10). Regulatory or legal flags subtract 15. The final value is clamped to 0–100.

Measures downside risk. Higher score = lower risk.

The risk score blends 30-day annualised volatility, 90-day max drawdown, debt-to-equity, earnings event proximity, average daily liquidity, asset type (ETFs receive a diversification bonus), and a data-quality completeness bonus. It is the dominant input for conservative and dividend-focused weight presets.

Measures how the candidate complements the active strategy template, not the user as a person.

Portfolio fit checks whether the asset type matches the strategy need (e.g. an ETF for an ETF-first template), whether the candidate fills an uncovered sector gap, whether it duplicates an already-high-scoring candidate, and whether it lines up with the core/satellite role described by the strategy. The score is clamped at zero when speculative assets appear in templates that exclude them.

Reads the ETF in context of the strategy and asks whether it can carry the backbone of the portfolio.

Looks at the business — durability, growth, margins, debt — and writes a candid summary, weaknesses included.

Reads the trend, momentum, and overbought / oversold posture — and is honest when the data does not support a clear read.

Frames the candidate against the active strategy — role, allocation band, duplication risk — without ever framing the candidate as personally matched to the user.

Takes the opposing view. Required to cite specific data, never to invent generic risks.