Direct answers on model behaviour, security boundaries and configuration. Further detail is available in an executive walkthrough.
How does the Smart Stop-Loss avoid premature whipsaw exits?
Fixed stops trigger on price alone. Redphonix requires confirmation across depth, volatility regime and correlation before signalling. Brief price wicks without liquidity deterioration generally leave corridors unchanged, while sustained thinning tightens protection. Sensitivity remains adjustable to your mandate.
Is the platform custodial? How are API security boundaries handled?
Redphonix is non-custodial. Assets remain with your chosen custodian or exchange. Where execution connectivity is enabled, we recommend read-only or trade-only keys without withdrawal permissions, IP allow-listing and regular key rotation. You can operate in signal-only mode with no trading access at all.
What data inputs are used, and how explainable are the models?
Inputs include order-book depth and spread, trade-flow imbalance, realised and implied volatility measures, cross-asset correlations and selected macro liquidity indicators. Each signal references the dominant inputs, threshold values and model version, so risk teams can review why an adjustment was proposed.
Can we customise risk tolerance and drawdown ceilings?
Yes. Drawdown ceilings, corridor width, alert staging and reserve levels are set per portfolio. Conservative mandates can require earlier de-risking and staged re-entry; more flexible mandates can allow wider corridors. All changes are logged with author, time and prior values.