Learn
How to read a stock forecast: dates, ranges and horizons
Read Vaniras stock forecasts by separating observed prices, starting prices, lower, median and upper model estimates, and realized outcomes.
Published by Vaniras · 2026-09-13 · Updated 2026-10-04
A stock forecast estimates a future outcome from a particular set of information. It is not a live quote or an instruction to trade. Start with a company in the stock directory, then open its forecast and read the dates before comparing numbers.
Separate the prices
| Value | What it means |
|---|---|
| Observed market price | A recorded market observation with its own timestamp; it can be delayed, stale or missing. |
| Forecast starting price | The price frozen when the forecast was issued. Later market moves do not change it. |
| Median Estimate | The model's central estimate at the selected target date. It is not an analyst consensus or a guaranteed future price. |
| Analyst estimate | A separate third-party estimate with its own source, date and method. |
| Descriptive score | A summary of available company evidence and peer context. It is not a forecasted return or investment recommendation. |
The displayed estimated percentage change is measured from the forecast's starting price. It does not necessarily represent the change from the latest market observation.
An illustrative reading
This is a hypothetical illustration, not a forecast for a real stock. Assume one issued forecast has the following values for a single target date:
| Value | Illustrative amount |
|---|---|
| Frozen starting price | USD 100 |
| Lower Estimate | USD 80 |
| Median Estimate | USD 110 |
| Upper Estimate | USD 140 |
| Later observed market price | USD 120 |
The median's estimated change from issuance is 10%: (110 − 100) / 100. The later observation of USD 120 does not rebase that issued estimate. USD 80 and USD 140 are model estimates for the target date, not promises that the price will remain within that range before or at that date.
Read the dates independently
The data-through date identifies the information cutoff. The generation time records when the forecast was produced. The starting-price date belongs to the frozen reference price. The market quote carries a separate observation date.
Refreshing a page or checking a cached response does not make the forecast newer. A stale label is useful information. Missing data is unavailable, not zero; a supported company can have no forecast.
What lower and upper estimates do and do not say
Lower and upper estimates apply to a target date. They are not guaranteed price bounds, and prices can move outside them. The displayed range does not establish that it contains 80% of future realized observations.
That coverage claim needs independent evaluation on later, untouched outcomes. A narrow range does not prove accuracy, and a median-only result does not justify inventing lower and upper estimates. Read the methodology and calibration limits.
Four horizons are four target dates
The model uses 3, 6, 9 and 12-month horizons. Free horizon access follows the current membership policy; additional horizons require eligible Plus or Pro membership. Purchase availability follows the payment provider and supported territory. Each horizon is a separate target-date estimate, not a predicted daily path. Target sessions can remain unavailable until the required market-calendar evidence exists.
Use the same horizon when comparing stocks, and check whether their snapshots and starting-price dates match. The forecast directory keeps horizon selection separate from observed prices.
Backtests and live outcomes answer different questions
A backtest examines historical forecast origins under specified data and validation rules. Live forward results use forecasts recorded before their outcomes became known. Neither is a guarantee of investment returns.
A live outcome becomes measurable only when its exact target-date observation can settle under the stated rules. Pending outcomes are not successful predictions. The methodology explains these distinctions; public track-record results are not currently available.
Continue your research
Before comparing estimates, identify the horizon, target date, frozen starting price and information cutoff. Keep absent values unavailable, and distinguish a pending outcome from a measured one. The missing and stale data guide explains those states in more detail.
Explore Apple or Microsoft for dated examples, read the forecast methodology, and check the financial disclaimer. Forecasts may be wrong and can never remove investment risk.