Twywell applies artificial intelligence to live market data and more than a decade of historical cycles, giving location-independent investors the analytical depth once reserved for trading floors.
Dashboard view shown above: portfolio risk exposure, model confidence intervals, and recommended rebalancing actions, refreshed continuously as new data arrives — accessible from any location with no on-site infrastructure required.
How It Works
Each recommendation Twywell produces is the output of a defined pipeline, not a single algorithmic guess. The platform is built around three functions that operate continuously rather than on a scheduled review cycle.
Market feeds, economic indicators, and asset-specific signals are consolidated as they arrive, removing the delay between data publication and portfolio review.
Forward-looking exposure is recalculated against current volatility and correlation patterns, rather than relying on a static risk rating set at the time of purchase.
Outputs are ranked by projected risk-adjusted return and presented with the underlying rationale, so the reasoning behind each suggestion can be checked before acting on it.
Time-to-decision: Positions that previously required manual spreadsheet consolidation and a scheduled weekly review are now reassessed every time new data is ingested, allowing decisions to be made between model updates rather than waiting for the next review cycle.
About Twywell
Twywell was designed around a single constraint: the user is rarely sitting at a desk on a trading floor. Remote professionals, independent investors, and digital entrepreneurs need the same analytical rigour as an institutional desk, delivered through a browser rather than a terminal room.
The platform does not replace judgement. It narrows the field of decisions worth reviewing, flags the ones with changing risk profiles, and keeps a record of how each recommendation performed against what the models projected, so accuracy can be checked rather than assumed.
Methodology
Every recommendation can be traced back through four defined stages. None of them are skipped, and none are presented without the data that supports them.
Pricing, volume, macroeconomic releases, and sector-level indicators are pulled from source feeds and normalised into a single schema.
Models identify recurring structures in price behaviour and correlation shifts, distinguishing short-term noise from sustained movement.
Every candidate pattern is cross-referenced against more than ten years of market cycles, including downturns, before it is permitted to influence a live recommendation.
Surviving signals are translated into a ranked recommendation with a stated confidence level and the historical basis behind it.
Applications
The underlying models are the same across each scenario. What changes is the weighting applied to risk tolerance, time horizon, and liquidity requirements.
When correlations between holdings shift, the model recalculates optimal weighting and surfaces the trades that reduce concentration risk without requiring a full manual review of the portfolio.
Before capital is allocated to an unfamiliar asset or sector, the backtesting layer runs the proposed entry against comparable historical conditions, returning a range of probable outcomes rather than a single projection.
Positions with increasing downside correlation are flagged as risk thresholds are approached, giving remote operators time to act before a single market event affects a disproportionate share of the portfolio.
System Performance
Twywell publishes how the system performs against its own backtested projections, rather than relying on third-party endorsement.
Illustrative comparison: backtested model output against subsequent live performance over a representative holding period. Full historical charts by asset class are available within the platform.
Infrastructure uptime is monitored on a rolling basis, with any service interruption logged and reported to subscribers through the status page.
Model accuracy is reassessed on a recurring schedule against live outcomes, and variance against the original backtest is documented rather than withheld.
Frequently Asked
Account data and portfolio information are encrypted in transit and at rest. Access to the platform requires authenticated login, and no portfolio data is shared with third parties for marketing purposes.
Models are retrained on a recurring schedule as new market data becomes available. Backtesting is re-run whenever a model update is deployed, so historical validation stays aligned with the live version in use.
Yes. The platform is browser-based and does not require local software installation, on-site hardware, or a fixed location. Dashboards and alerts are accessible from any internet-connected device.
A subscription provides access to the live dashboard, automated recommendations, historical backtesting records, and the system status reporting referenced on this page. Specific plan details are confirmed at sign-up.
Each recommendation is presented with the backtested basis and current confidence level that produced it, so the underlying reasoning can be reviewed before any decision is made.
Twywell is built for investors and operators who make decisions away from a trading desk, and who want the reasoning behind each recommendation available before acting on it.
Launch Twywell Dashboard Questions about setup or your subscription? Contact the support team.