Skip to main content

Roadmap

Page role

Future-facing lane. Use this page for planned or experimental work. If you want the shipped path today, start with Getting started. If you want near-term unreleased work with explicit graduation status, use Next release track.

Goal: Make XRTM the standard local workbench for forecasting and model evaluation with scientific rigor.


Strategic pillars

PillarFocusAudienceStatus lens
Core physicsValidity, math, timeResearchers, model-eval teamsShipped foundation first
Research gradeAuditability, advanced workflowsOperators, labs, teamsMixed shipped + in progress
ExperimentalFrontier capabilitiesExploratory usersNot a homepage claim

Released and shipping now

FeatureDescriptionStatus
Chronos ProtocolEnforced time cutoffs to prevent look-ahead bias during backtests.DONE
Calibration EnginePlatt scaling and Brier-score-oriented evaluation.DONE
Sentinel ProtocolForecast trajectory tracking over time.DONE
Recursive ConsensusSupervisor-led disagreement refinement loops.DONE
Truth ProtocolTamper-evident reasoning traces and audit packaging.DONE
Dossier GeneratorAudit-ready research reporting outputs.DONE

In progress

FeatureDescriptionStatus
Epistemic SecuritySource verification and anti-manipulation work.In Progress

Experimental or future

FeatureDescriptionStatus
Causal ReasoningFormal DAG generation beyond current text-first reasoning traces.Experimental
Prompt CompilerAutomated Brier-driven prompt optimization.Experimental
Centaur WorkbenchHuman-in-the-loop collaboration layer.Concept
ABM SimulationAgent-based market dynamics and liquidity analysis.Experimental

Contribution guidance

  1. Highest impact: harden time-safe evaluation and adapter coverage
  2. Next: improve shipped researcher and operator workflows
  3. Later: experimental work only if you are prepared to maintain it