How Forecast Logic Is Generated
This article describes how Model Reef uses imported ticker fundamentals to automatically generate baseline forecast logic and how you can refine it for your own scenarios.
You will learn:
How default forecast rules are created.
How growth and margin assumptions are derived.
How timing and cash behaviour is handled.
How to override or extend the defaults.
Default forecast horizon
When you import fundamentals for a ticker, Model Reef:
Sets a default forecast horizon beyond the last historical period.
Creates base case forecast values for revenue, costs, operating expenses and other key lines.
Ensures that P&L, Balance Sheet and Cashflow extend through that horizon.
You can change the model start date, end date and periodicity at any time. Model Reef will regenerate forecasts and statements accordingly.
Default growth and margin logic
For many lines, the default forecast logic is derived from historical fundamentals. Typical patterns include:
Using average historical growth rates to project revenue and some cost lines.
Using recent margin relationships between revenue and COGS or operating expenses to estimate future values.
Applying simple, explainable rules like:
Revenue grows at an average of the last few years.
COGS scales with revenue to preserve or gradually adjust gross margin.
Opex grows at a smoothed rate based on history.
These rules are intentionally simple and transparent so that you can understand and override them easily.
Default timing and cash behaviour
By default, Model Reef applies timing rules that are consistent with typical listed company behaviour, for example:
Most revenue and cost lines are assumed to have short or zero delays between accrual and cash.
Depreciation and loan interest are scheduled based on asset and liability patterns in the imported data.
Tax and dividends follow approximate patterns based on historical ratios and payout behaviour.
You can refine these assumptions by editing timing settings on each variable if you need more detailed cash modelling.
How ML or regression presets may be used
For some series, Model Reef may use simple regression or pattern based methods to generate forecasts, for example:
Fitting a trend to historical revenue and extending it forward.
Using relationships between key drivers (for example revenue vs gross profit) to stabilise margins.
Smoothing noisy series before projecting them.
These methods are designed to be conservative and interpretable rather than opaque black box forecasts.
Overriding and refining forecast logic
You can override the baseline logic at any time by:
Replacing default formulas with custom formulas that reference drivers and other variables.
Introducing explicit drivers (units, prices, usage metrics) and aligning variables to those drivers.
Editing growth rates, escalation patterns and seasonality profiles.
Changing timing and delays to better reflect receivables, payables and other cash dynamics.
The imported fundamentals give you a starting point, but the model is fully editable.
Scenarios and alternative forecast cases
To build multiple forecast cases for the same ticker, you can:
Duplicate the model to create separate scenarios (for example Base, Upside, Downside).
Keep the same historical fundamentals in each scenario.
Change forecast logic, discount rates or terminal value settings independently.
Compare outcomes across scenarios for KPIs and valuations.
Because each scenario is its own model, changes remain isolated and easy to manage.
Related articles
Last updated