AInvestor Recommendation: Definition and How It Is Calculated
The AInvestor Recommendation is the actionable label (e.g., Strong Buy, Buy, Hold, Sell, Strong Sell) produced by combining model scores, fair value comparisons, and confidence/data checks to guide investors.
Definition
An AInvestor Recommendation summarizes multiple analytic signals into a single buyer/seller stance. It is intended to help prioritize further research and is not investment advice.
How the recommendation is calculated (step-by-step)
- Inputs: primary inputs are the AInvestor Score (0–100), one or more fair value estimates (weighted fair value), current market price, and a confidence level produced by data validation checks.
- Compute Fair Value Ratio (FVR): FVR = weightedFairValue / currentPrice. FVR > 1 implies the stock may be undervalued, FVR < 1 implies overvaluation.
- Data quality & confidence: verify required data (price, fair value, and score) and use the confidence level. If confidence is low or key inputs missing, the recommendation is downgraded or marked as not actionable.
- Decision rules: combine score and FVR using threshold logic and tie-breakers. Example simplified logic:
- Strong Buy: Score ≥ 80 and FVR > 1.25
- Buy: Score ≥ 65 and FVR > 1.05
- Hold: Score between 50–64 or FVR between 0.9–1.05
- Sell: Score between 35–49 or FVR < 0.9
- Strong Sell: Score < 35 and FVR < 0.75
- Weighting and tie-breakers: when score and FVR disagree, the system may prefer the higher-confidence signal (e.g., a very high FVR with low score may be downgraded). Additional heuristics include capping recommendation changes if recent volatility or data revisions exist.
- Output: the produced recommendation includes: label (Strong Buy/Buy/Hold/Sell/Strong Sell), reason string (score and FVR summary), the fair value ratio, confidence level, and any data-quality flags for downstream UI use.
Practical notes
- Recommendations are automated and should be used as a starting point for research, not as sole decision criteria.
- Different products or viewers may vary thresholds or weighting; always check the component breakdown.
- The UI surfaces confidence and breakdowns so users can understand which inputs drove the recommendation.