AP-2.0 is an explainable description of available research evidence. It has not been calibrated as a prediction of returns, a success probability or a personal recommendation. The score’s confidence describes evidence reliability and agreement.
The scale
90–100 Exceptional; 80–89 Strong; 70–79 Positive; 50–69 Neutral / Mixed; 30–49 Weak; 0–29 Very Weak. Text labels accompany the number. Higher Risk Score means higher observed risk: 0–24 low, 25–49 moderate, 50–74 elevated, 75–100 high. In the overall calculation, risk contributes as 100 minus Risk Score.
Asset and timeframe weights
Stocks
| Category | Intraday | Swing | Long-Term |
|---|---|---|---|
| Technical | 40% | 25% | 10% |
| Momentum | 20% | 10% | 5% |
| Fundamental | 0% | 20% | 35% |
| Valuation | 0% | 10% | 20% |
| Catalyst | 5% | 5% | 5% |
| Sentiment | 10% | 10% | 5% |
| Earnings | 5% | 5% | 5% |
| Ownership | 0% | 5% | 5% |
| Risk | 20% | 10% | 10% |
| Quality | 0% | 0% | 0% |
| Options | 0% | 0% | 0% |
Gold
| Category | Intraday | Swing | Long-Term |
|---|---|---|---|
| Technical | 35% | 25% | 15% |
| Macro | 10% | 20% | 30% |
| Momentum | 20% | 15% | 10% |
| Volatility | 15% | 10% | 5% |
| News | 5% | 10% | 10% |
| Positioning | 5% | 10% | 20% |
| Risk | 10% | 10% | 10% |
Forex
| Category | Intraday | Swing | Long-Term |
|---|---|---|---|
| Technical | 35% | 25% | 15% |
| Macro | 5% | 20% | 30% |
| Yield differentials | 5% | 15% | 20% |
| Momentum | 20% | 10% | 5% |
| Sentiment | 10% | 10% | 10% |
| Session context | 15% | 10% | 5% |
| Risk | 10% | 10% | 15% |
Other market references
| Category | Intraday | Swing | Long-Term |
|---|---|---|---|
| Technical | 55% | 55% | 55% |
| Momentum | 25% | 25% | 25% |
| Risk | 20% | 20% | 20% |
Stock Quality is a diagnostic within Fundamental, so it receives no extra overall weight. Options direction is unavailable from the current chain feed and also receives zero overall weight. Other market references use price and risk evidence only and are labelled as such.
How a score is calculated
Each factor has a disclosed 0–100 rule. Available factor weights are averaged within its category. Effective category weight = base weight × available-input fraction × freshness × source-quality factor. The remaining effective weights are normalised to 100%, and the overall score is their weighted average. Risk is inverted only for this overall contribution. Missing factors never receive a fabricated 50.
Source-quality factors are conservative heuristics: Technical/Momentum/Volatility 0.85; reported Fundamental/Quality 0.90; relative Valuation 0.75; Earnings and Risk 0.80; delayed Ownership 0.65. News uses its attributed publisher-quality average, with the current value shown. These are disclosed weighting adjustments, not audited statistical accuracy estimates.
Price scores require at least 50 completed candles. Technical uses structure (40%), moving-average comparisons (40%), established-level reactions (10%) and directional volume (10%). Momentum uses RSI (40%), MACD histogram (35%) and 20-candle return (25%); RSI is scaled 30–70. MACD positive/flat/negative scores 75/50/25. Return scales are −3% to +3% intraday, −10% to +10% swing and −30% to +30% long-term.
Fundamental reuses the disclosed reported-period financial model: revenue growth 15%, EPS growth 10%, profitability 20%, cash flow 20%, balance sheet 15%, returns 10%, dilution 5%, consistency 5%. Financial and real-estate sectors remain unscored where specialist rules are needed. Quality shows profitability, cash flow, returns and consistency without adding their influence twice.
Valuation compares at least two independent multiple groups, each with at least three same-industry, same-sector, same-currency, similarly sized peers. Group score = 50 − 50 × log2(company multiple / peer median), capped 0–100. Trailing and forward P/E share one group; sales and book are separate groups. Related securities alone do not qualify.
Earnings uses comparable EPS surprise (50%), revenue surprise (25%), at least three reports of EPS consistency (15%) and completed first-session reaction (10%). Ownership uses dated holder-change breadth (60%), verified discretionary insider trades (30%) and comparable short-share changes (10%). Routine/unverified sales are excluded; short crowding belongs to Risk. Sentiment uses recent attributable company news, excludes duplicate coverage and analyst/insider headlines, and requires at least three eligible stories.
Stock Risk uses ATR/price (40%), maximum close drawdown over 21 candles (20%), reported absolute beta (15%), debt/revenue (15%) and dated short float (10%). ATR/price scales to 100 risk at 1% intraday, 8% swing and 20% weekly. Beta 1–3, debt/revenue 0–3 and short float 0–30% are capped linear scales. Other assets use drawdown (60%) and the largest absolute open/preceding-close gap (40%); Gold Volatility is separate and uses 100 minus ATR/price percentile over up to 100 candles. Drawdown caps are 3%/25%/40%; gap caps are 1%/8%/15% for intraday/swing/long-term. Broker leverage, liquidity, spread and all possible future gaps are not modelled.
Eligibility and missing inputs
Overall scores require at least 65% available weighted coverage intraday or 60% otherwise, with current Technical and Momentum evidence. Swing stocks additionally require company or attributable news evidence; long-term stocks require Fundamental. Long-term Gold requires Macro; FX requires Macro or Yield differentials. If these gates fail, sub-scores can still be shown and Overall stays Unrated.
Directional catalysts, Gold macro/positioning/news, and pair-specific FX macro/yields/sentiment/session-liquidity feeds are not connected in this release. Their categories stay unavailable. A release calendar or a session clock never provides directional points. Spot XAUUSD is not connected; GC=F is explicitly separate futures research. Stock fundamentals are never applied to either asset.
Confidence
High: at least 85% coverage, four contributing categories, 80% weighted source quality, 80% directional agreement and no significant conflict. Medium: at least 70% coverage, 65% quality, 65% agreement and no significant conflict. Otherwise Low. Agreement = 1 − 2 × the smaller bullish/bearish contributing weight / directional weight. Category scores ≥60 and ≤40 establish those sides. Risk and relative volatility do not vote on price direction. Substantial two-sided weight or conflicting internal factors lowers confidence.
Freshness and update frequency
Scoring uses completed 15-minute, daily and weekly candles for Intraday, Swing and Long-Term respectively. Retrieved price history must be within 15 minutes and pass the existing session/timeframe freshness checks. Company snapshots, earnings and ownership retrievals must be within one day; financial reporting and ownership comparison periods retain their own age gates. Short interest must be dated within 45 days; fundamental quarterly periods within 200 days; comparable annual periods within 550 days. Missing or future timestamps are ineligible.
Intraday summaries are cached for one minute; Swing and Long-Term for five minutes. Underlying price/company/news/statement feeds have their existing caches. Scores update when research is requested or refreshed, not continuously in the background. New completed candles, dated reported results or new source evidence can change them.
Stability and history
For a comparable prior observation, routine updates move 35% towards the raw weighted score and are capped at three points per changed observation. The unsmoothed raw score is visible. New earnings or financial-reporting periods, a completed-candle move of at least 2% intraday or 5% otherwise, or at least two 20-point category changes bypass smoothing. Changes in evidence coverage, methodology or an old comparison start a new baseline. Unchanged evidence is not repeatedly smoothed.
History stores immutable server-calculated public observations, separated by symbol, asset class, timeframe and method. It records previous/current score, comparable delta, time and leading factor changes. History begins with actual observations after this release; there is no invented backfill or retrospective backtest. An observation is not a personal trade, and no visitor identifiers, watchlists or positions enter it. Storage failure leaves current research available and marks history unavailable.
Reading explanations
“Why this score?” shows factual inputs separately from interpretation rules, base and effective weights, category confidence, sources and dates. Impact points are the contribution relative to a neutral 50 before smoothing; positive Risk impact means lower observed risk reduced the penalty. A strong intraday setup does not establish a strong long-term investment.