Why You Can Trust Our Rankings
Our rankings are based on objective historical performance, transparent methodologies, and consistent evaluation criteria—not opinions, popularity, or marketing.
Introduction: The sole objective of Analyst Performance Tracker is to help investors evaluate the historical performance of financial analysts and research organizations. We achieve this by using measurable, repeatable, and audited mathematical models that remove guesswork, reputation bias, and promotional noise from financial forecasting.
Our Philosophy
Historical Performance Matters
We believe that past performance, when measured systematically over significant time horizons, provides critical context. It helps investors identify genuine analytical edge versus temporary market luck.
Consistency Over Anomalies
A single extreme prediction that happens to succeed should not define an institution's rank. Consistent accuracy and repeatable success across numerous market cycles are far more valuable than isolated, lucky outliers.
Absolute Objectivity
Ratings must be free of conflicts. We do not accept sponsorship or placement fees from analyst organizations to boost their scores. Every institution is subjected to the exact same rigorous rules.
Standardized Methodology
No exceptions or subjective adjustments. Whether evaluating a global investment bank or an independent research shop, we apply the exact same mathematical parameters and grading schemas.
Data Sources & Flow
Our platform functions as an independent analytics layer. We ingest market information from professional-grade feeds and compile them into standardized database registries.
Aggregated recommendations (BUY, SELL, HOLD) published by research institutions, capturing the start date, target duration, and firm identifiers.
Historical end-of-day (EOD) and intraday market pricing datasets to verify the exact asset trajectory following each prediction.
Supporting corporate metrics such as sectors, industries, market caps, and earnings calendars to categorize peer performance.
Analyst Performance Tracker is not a primary exchange, broker, or market data owner. Raw prediction metadata and pricing datasets are licensed from third-party provider feeds, including the Financial Modeling Prep (FMP) API and verified public market price archives. We do not claim proprietary ownership of the underlying historical raw tape pricing; instead, we own the downstream analytical models, calculated performance indices, and leaderboard metrics generated from this data.
How We Evaluate Predictions
Unlike static trackers that look solely at arbitrary annual windows, we evaluate each recommendation on its own timeline using three distinct mathematical options. This gives investors the flexibility to measure return-on-investment (ROI) based on their specific trading style.
1. Peak System (Default)
Measures the maximum favorable price deviation reached at any point during the evaluation timeframe. This credits analysts for correctly identifying directional entry trends, even if macro shifts drag the asset back down before expiration.
2. End ROI
Evaluates performance strictly based on the market price recorded on the final date of the prediction's timeframe. It ignores intermediate fluctuations, focusing purely on long-term terminal return.
3. Symmetric Peak ROI
Factors both the maximum upside peak and the maximum downside drawdown. It subtracts the downside risk from the upside gain, rewarding analysts who achieve returns with low volatility.
Incorrect Prediction Punishment Options
When a prediction moves in the wrong direction (e.g., a BUY call where the stock declines), users can toggle how this error is recorded:
Unsuccessful calls are capped at 0.0% ROI. This isolates the average return of successful ideas without penalizing the overall average with failed positions.
The actual negative return of the failed prediction is factored in. A BUY call that fell to -25% records a -25% return, dragging down the average.
Why are HOLD Ratings Treated Separately?
A HOLD rating represents a call for stability or range-bound consolidation rather than directional gain. Therefore, HOLD recommendations are excluded from ROI pools. It would skew performance statistics to average directional returns (BUY/SELL) with non-directional calls. Instead, HOLD predictions are evaluated solely on stability success metrics (whether the stock price stayed within a ±5% or ±10% range).
How Our Rankings Are Calculated
Our platform evaluates institutions using five core statistics. Here is exactly what they measure, why they are useful, and how to interpret their mathematical definitions.
Prediction Success Rate (PSR)
Measures directional accuracyIdentifies whether the analyst consistently predicts the right direction of the asset.
Expressed as a decimal or percentage. A score of 0.75 means the analyst achieved full or partial success targets on 75% of their total prediction volume.
Average ROI
Measures simple return magnitudeCalculates the arithmetic average return across all BUY and SELL recommendations.
Useful for checking the absolute return generated by calls, but can be highly skewed by single extreme stock movements.
Median ROI
Measures statistical robustnessIdentifies the middle return value. Protects against outlier predictions skewing the data.
A true representation of what a typical recommendation yields. If Median ROI is 15%, a random recommendation from this firm is most likely to yield close to 15%.
Index Score
Combines return magnitude with consistencyA dimensionless score that prevents high-return/low-accuracy players from ranking first.
A composite rating. An analyst with 40% Median ROI and 60% PSR achieves an Index of 24.0 (0.40 * 0.60 * 100), outranking an analyst with 100% Median ROI but only 10% PSR (Index = 10.0).
Confidence Weighted Index
Rewards statistical sample sizeProvides statistical confidence. A firm with 100 predictions is more reliable than a firm with 5.
Firms with a larger volume of predictions receive a confidence multiplier. Exponent defaults to 0.50 (square root scaling), balancing the leaderboard.
Step-by-Step Worked Example
Let's walk through how Institution A's standard Index Score is calculated using their historical prediction datasets.
Calculation Breakdown:
- Step 1: Convert Median ROI to Decimal CoefficientMedian ROI of 18.0% is represented as a decimal coefficient:
0.18. - Step 2: Multiply by PSR BaseMultiply the ROI coefficient by the Prediction Success Rate (PSR = 0.82):
0.18 × 0.82 = 0.1476 - Step 3: Scale Score to Leaderboard IntegerScale the result by 100 to yield the final Index Score displayed on the dashboard:
0.1476 × 100 = 14.76
Why We Use Median ROI
In financial reporting, average figures are easily manipulated. An analyst firm can publish 10 predictions, 9 of which result in heavy losses or flatlines, while a single micro-cap biotech stock rallies 400% on FDA approval.
If we look solely at the Average ROI, this firm appears to have achieved an outstanding +35% return across all ideas. However, any investor copying their average trade would have lost money on 90% of their trades.
By using Median ROI (the middle value of sorted returns), we bypass this mathematical distortion. Outliers and extreme stock movements cannot skew the median, making it a much more robust and honest indicator of typical performance.
* Note: The average (+56%) claims stellar performance, while the median (+15%) shows what the typical investor actually experienced.
Why We Offer Multiple Methodologies
In quantitative finance, there is no single "correct" definition of success. Different investors operate under different mandates and risk tolerances:
Growth-oriented investors prioritize Peak ROI to locate analysts who can spot momentum spikes.
Patient, buy-and-hold allocators look at End ROI to focus strictly on terminal settlement.
Risk-conscious managers prefer Symmetric Peak ROI to penalize high-drawdown recommendations.
High-frequency strategies look at PSR Base to maximize the hit rate of their operations.
By allowing users to customize these methodologies, we enable every investor to build a personalized definition of success instead of forcing a singular, opinionated rating index on them.
System Limitations & Risks
Transparency is not complete without disclosing limits. No quantitative platform is bulletproof, and investors should review the following realities:
- ·Past Performance Limits: Historical success scores do not guarantee future accuracy. Market regimes change, and analysts can lose their analytical edge.
- ·External Provider Inaccuracies: While we verify outcomes, raw analyst recommendation dates and values can contain inherited errors from feed sources.
- ·Mathematical Multiplicity: Different methodologies can generate significantly different leaderboards. No ranking is "the absolute truth."
- ·Due Diligence Requirement: Our rankings are analytical tools, not investment advice. Users must conduct independent research before making capital commitments.
Frequently Asked Questions
Why are there multiple ranking methods?
Why did a firm's rankings change suddenly?
Why is Median ROI different from Average ROI?
Why isn't the institution with the biggest winning prediction ranked first?
Why are HOLD ratings excluded from ROI pools?
How often is the database updated?
Our Institutional Commitment
We are dedicated to building the most transparent, objective, and conflict-free analyst tracking engine on the web. We commit to keeping our methodologies open, updating data regularly, and actively refining our platform based on user feedback.