Glossary
Every term used across the app, defined.
Produced by fixed rules over stored data, so the same inputs always give the same output. AlphaForgeReco's scores, rankings, signals and performance figures are deterministic and auditable — no randomness and no model guessing. AI only writes explanations after the numbers are final.
A structured, rules-detected event worth attention — a breakout, an oversold RSI, a volume spike, a moving-average cross or a rating change — that has passed a minimum importance threshold. Signals are deterministic: they fire only when stored data meets the rule.
A 0–100 composite that blends the factor scores (momentum, quality, growth, value, risk) under a chosen style. Higher means a stronger relative standing. The score drives ranking order — it is not a price target or a probability.
How much corroborating evidence supports a recommendation — e.g. agreement across factors, data completeness and signal strength. Higher confidence means the deterministic case is better supported; it is not a guarantee of the outcome.
The highest conviction rating: the asset scores in the top band and multiple factors align. A strong-but-still-deterministic case, not a promise.
A favorable rating: the asset scores above the buy threshold with a generally supportive factor mix.
A neutral rating: the score sits in the middle band, or signals are mixed. No strong deterministic edge either way.
An unfavorable rating: the score is below threshold and/or risk factors dominate. For tracking, a correct Sell is one where the name subsequently underperforms.
The lowest rating: weak score plus elevated risk (e.g. deep drawdown, high volatility). Outcome tracking credits Avoid with the loss it helped sidestep.
Return in excess of a benchmark. If a position returns +12% while the benchmark returns +5%, its alpha is +7 percentage points. AlphaForgeReco measures alpha everywhere it claims an edge.
The reference return an idea is compared against (e.g. a broad index or QQQ). Returns are only meaningful relative to a benchmark; excess return over it is alpha.
The peak-to-trough decline over a period, expressed as a percentage. Max drawdown captures the worst loss an investor would have endured — a core risk measure.
How much a price swings, usually the annualized standard deviation of returns. Higher volatility means larger swings and more risk; it feeds the deterministic risk factor.
The tendency of recent performance to persist. AlphaForgeReco measures momentum from multi-horizon price returns and trend; strong momentum lifts the score.
How a security is performing relative to the broad market or its peers, rather than in absolute terms. Leaders show positive relative strength.
Relative Strength Index — a 0–100 momentum oscillator. Below ~30 is often called oversold, above ~70 overbought. See the Technical Indicator reference for detail.
Moving Average Convergence Divergence — a trend/momentum indicator built from the difference of two EMAs and a signal line. Crossovers flag shifts in momentum.
The average price over a trailing window (e.g. 50 or 200 days) that smooths noise to reveal trend. Price above a rising average is a constructive backdrop.
Replaying a strategy over historical prices to see how it would have performed. Useful but not a guarantee — real results differ, and overfitting is a risk.
A tracked 'what if you had kept it' position. When you sell a holding, AlphaForgeReco keeps a shadow of the old position so it can compare the road not taken against what you actually did.
The performance difference between a decision you made and the counterfactual. Example: you sold AMD (+8.4% if kept) and bought ALAB (+18.7%) → decision alpha of +10.3 points. It answers 'did switching help?'.
An immutable daily record of the market state, recommendations, signals, opportunities and risks. Because it is frozen, later outcome tracking can honestly grade what was said that morning.
A transparent 0-100 rating of one portfolio decision across seven weighted components: outcome, position sizing, risk-adjusted return, timing, agreement with the recommendation at the time, process consistency, and confidence calibration. Components without enough history are left blank and the score is re-weighted over what's available -- never invented.
A 0-100 grade (High/Medium/Low) of how much data-quality confidence backs a rating -- evidence diversity, coverage and freshness. It is informational only and never influences the rating itself; a weak Trust Score means 'treat this rating cautiously,' not 'the rating is wrong.'
The written reason you hold a position, broken into pillars -- specific, checkable conditions (e.g. 'revenue growth stays above 20%'). Thesis Monitor tracks supporting and contradicting evidence per pillar over time and flags when a thesis has strengthened, weakened or broken.
A system-detected change -- a rating change, technical trigger, catalyst or portfolio-risk breach -- surfaced automatically in Notifications as it's detected. Distinct from a user-defined price/RSI Alert, which only fires when a rule you configured is met.