Table requests
See how each model compares—suggest your own scenario
Side-by-side tables let you weigh methods quickly. If you want a more detailed review,
or your use case isn’t listed, ask us for a comparison. We welcome requests for new
datasets or techniques. Results may vary; past performance does not guarantee future
results.
Request table
Explore the tables
1
Classic vs neural
Direct table comparison of classic statistical models and neural-based methods, updated for recent datasets.
2
Metrics shown
Each table includes a metric summary, so you can see accuracy and error at a glance.
Assumptions matter
Assumptions are listed for every model, making it easier to spot potential mismatches with your data.
Suggest new table
Request a new scenario for comparison. We add relevant tables as feedback comes in.
Table updates and feedback
Feedback applied
We revisit each table after feedback, clarifying when a model’s complexity is justified. Trade-off notes are published with every revision.
Change logs
All comparison tables are versioned, with logs showing what’s new or changed. You always know what’s been updated.
When scores match
Where models produce similar scores, we highlight any differences in setup, assumptions, or robustness.
Simplicity focus
Tables are designed for clarity, not just completeness. We avoid jargon where possible.
Comparison at a glance
You know the moment when two models score nearly the same, but behave completely
differently on unseen data. Tables here let you see those moments: every comparison
includes not just the score, but why the difference matters.
Every table includes model assumptions, setup notes, and contextual warnings. Past performance does not guarantee future results; always consider model suitability for your own data.
Results may vary depending on dataset characteristics. If you have a dataset or scenario
not shown, let us know. We add new tables based on user feedback and emerging research.
How our tables help you compare models with fewer surprises
1
Clarity-focused comparison with context
Each table is designed to be immediately readable: key metrics, main assumptions, and limitations are side by side. You see the context for each score, not just the number.
3
Explicit callouts of common pitfalls
Assumptions and pitfalls are called out, so surprises are minimized. We note where a model may underperform due to structural issues.
Model comparison tables
Tables here summarize model performance, trade-offs, and context, so you can make informed decisions without wading through jargon.
This page focuses on direct, table-based comparisons of time series forecasting models. Each table highlights differences in performance, assumptions, and dataset suitability. The goal: provide a clear, at-a-glance reference for technical readers.
Trend vs seasonal strengths
Understand which models excel in trend-dominated datasets versus those with strong seasonality.
Setup and tuning needs
Quickly see which models require the most preprocessing or tuning, thanks to side-by-side notes.
Comparing error metrics
Error metric columns show how each model’s reported accuracy changes with data characteristics.
Weighing complexity
Trade-off summaries indicate when a model’s added complexity is justified by real-world results.
Table gallery
Sample images from our comparison tables and updates
