For students and curious investors

Ask a portfolio question. See what the evidence can show.

SumWorth is a research workbench for students and curious investors. Compare a portfolio idea with available history, record the assumptions, and keep source, coverage, and limitations beside the result.

Open research workspace I already have an account

No account needed to explore · sign in only when you want to save

Take the guided example

• Reported sources and coverage gaps
• Assumptions and limits
• Observed and modelled kept separate

A public instrument, in miniature

See what a result can—and cannot—say.

Inspect a fictional two-sleeve fixture: DEMO-A at 60% moves from 100 to 112; DEMO-B at 40% moves from 100 to 106. Fixed arithmetic gives 109.6 normalized ending units from 100 starting units. The dates 2020-01-02 to 2023-12-29 are labels only, not observed history; no currency is represented.

No real securities or provider observations exist here. No analysis or AI call runs in the showcase. This is not a real portfolio or investment result, and no sample data is stored or carried into research.

1. Inspect the inputs. See the fictional allocations, levels, dates, and arithmetic before interpreting an output.
2. Ground the question. No historical answer is supported because the fixture has no provider observations. Live questions can be asked as a guest after a supported historical-stress result; sign-in is needed only to save an answer.
3. Read the evidence. Inspect provenance, date, coverage, assumptions, and limitations in the existing illustrative evidence passport and brief. Open the fixed illustrative evidence brief.

If evidence is unsupported or live AI is unavailable, SumWorth does not invent an answer.

Ask after supported evidence Preview the illustrative evidence brief Open the guest research lab

Illustrative UI · static visual fixture

Evidence before output

No live values, provider call, or predictive claim is included in this preview.

Illustrative evidence passport: no observations, model run, or forecast are attached. Real results show available source, observation date, coverage, and caveats; missing fields stay marked as not reported. Provider metadata is shown as reported; SumWorth does not independently verify source accuracy.

Observed

Real results show the provider, observation date, and price basis when supplied. This public illustration contains no market observations.

Modeled

Assumptions and modeled overlays are identified separately from observations.

Forward-looking

No forecast is shown here. Forecasts are model output, not provider data or a guarantee.

Coverage and caveats

Real results show common coverage and recorded caveats; missing fields remain marked as not reported.

A result is a chain of custody.

1. Start with observed history. Check what was supplied, its source, date, price basis, and common coverage.
2. Separate modeled overlays. Keep rebalancing, stress, resampling, and diagnostics distinct from observations.
3. Check the boundary. Label forecasts and possible paths as model output, with uncertainty and limitations visible.

The capability map

Follow the question, not the hype.

Guest research can run the analysis. An account is needed only for gated health and saved-portfolio areas, or to save reports and other work.

• News & Context — Guest ready. Provider headlines stay separate from historical evidence.
• Backtest — Guest research. Replay selected instruments, dates, and assumptions.
• Allocations — Guest research. Compare weights and modeled exposures.
• Monte Carlo — Guest research. Explore resampled paths derived from observed history and chosen assumptions.
• Correlation — Guest research. Inspect co-movement over a stated historical window.
• Portfolio Health — Account required.
• Risk & Exposure — Guest research. Inspect historical downside and exposure diagnostics.
• Data Trust — Guest ready. Check provider, dates, coverage, missing-data handling, and limitations.
• Famous Portfolios — Guest research. Study public allocation examples and their assumptions.
• Saved Portfolios — Account required.
• Forecast — Guest research. Forward-looking estimates remain model output, not guarantees.
• Quant Lab — Guest research. Inspect sampled allocation diagnostics.
• Scenarios — Guest research. Apply historical stress and diagnostic lenses.
• Reports — Guest reports are available; sign in to save work to an account.

Public field notes

Learn the instrument before you trust the reading.

Search one learning library for quick explainers, research concepts, practical guides, and applied strategy profiles.

Search the learning library

Explore first. Save only what you can explain later.

Run analyses without an account. Create one when you want portfolios, reports, and descriptive monitoring to follow you.