Data quality
Completeness and reliability of the inputs.
Forecasting MethodologyComing Soon
Our approach to probabilistic forecasting.
Amonhen Forecasting is being developed to transform market and onchain information into model-generated probabilities, expected ranges, confidence estimates, and forward-looking risk assessments.
Amonhen Forecasting is in development. This page describes the planned approach. It does not describe functionality or results that exist today.
A
Probabilistic Forecasting
The planned approach is to estimate future outcomes with probabilities and expected ranges rather than single price targets. A forecast is intended to express how likely different outcomes are, and how wide the plausible range is.
B
Forecast Confidence
Confidence is intended to reflect how much weight a forecast deserves. The factors planned to inform it are:
Asset history
How much history exists for the asset.
Liquidity
Depth and activity of the underlying markets.
Model calibration
How well past probabilities matched outcomes.
Data freshness
How current the inputs are.
Forecast horizon
Longer horizons carry more uncertainty.
C
Forecast Horizons
Forecast outputs may be offered across different time horizons, depending on model and asset coverage. Not every horizon is expected to be available for every asset, and no horizon is supported today.
D
Calibration & Performance Evaluation
The intended evaluation approach is to:
- Compare forecast probabilities with realized outcomes
- Track historical forecast errors
- Evaluate calibration, meaning whether stated probabilities behave like probabilities
- Assess accuracy across asset classes and horizons
No accuracy, calibration or historical performance results are published here, because none have been validated for release.
E
Model Limitations
Market forecasting is inherently uncertain. Limitations include:
- Changing market regimes
- Liquidity limitations
- Incomplete or delayed data
- Unexpected events
- Model uncertainty
- A forecast probability is not a guarantee
Forecasting is probabilistic. Model-generated estimates represent uncertainty and should not be interpreted as guaranteed outcomes or investment recommendations.
Interested in Amonhen Forecasting?
Register your interest to be considered for early access.
Amonhen is an independent data and forecasting platform and is not affiliated with, endorsed by, or sponsored by Robinhood Markets, Inc.