VĀRĀHĪ · Natural Time Intelligence

Vārāhī studies when natural events become more or less likely by modelling nature in its own temporal coordinates rather than treating every day as equivalent.

Nature has its own rhythm. Vārāhī makes that rhythm visible.

Kerala 2026 · flagship case

Historical compression demonstrates historical structure. It does not establish prospective forecasting ability. Independent replication has not yet been established.

Evidence classes

Studied · Historical pattern found · Modelled · Retrospectively tested · Out-of-sample · Prospectively frozen · Prospectively validated · Independently replicated · Not yet sufficient evidence.

Every result is labelled by evidence class. Sources, assumptions and transformations are inspectable.

Approved historical replay

126 years of Kerala monsoon replay (1901–2026). 107 phenomena studied. 100 computational models. One frozen prospective experiment.

PROVE IT. Inspect the evidence behind Vārāhī. · Nine historical case studies · Year explorer · 2026 default · Methodology · Research Library →

VĀRĀHĪNatural Time Intelligenceप्राकृतिक काल-बुद्धि

PROVE IT.

Inspect the evidence behind Vārāhī.

VĀRĀHĪ IS IN BETA

126 YEARS OF KERALA MONSOON EVIDENCE.
ONE FROZEN PRE-EVENT EXPERIMENT.
INSPECT THE ENGINE YOURSELF.

1901–2026Historical replay
126Seasons processed
125Complete three-band years
3Vārāhī natural-time phases
IMDExternal comparison
18 APR 2026Frozen pre-event artefact
CURRENT SEASON →HISTORICAL CASES →VALIDATION →METHODOLOGY →EVIDENCE CLASSES →SOURCES →LIBRARY →
RETROSPECTIVE RECONSTRUCTION

Historical observations are re-expressed after the events occurred.

OUT-OF-SAMPLE VALIDATION

A held-out or walk-forward observation is assessed without using its outcome to select the tested state.

PRE-EVENT FROZEN PREDICTION

A dated prediction artefact exists before the target event. This does not by itself establish a successful outcome.

PROSPECTIVE VALIDATION

A forward prediction is later scored against the same prespecified product and observation definition.

INDEPENDENT REPLICATION

An independent party repeats the method and reproduces the result. No case on this page is currently assigned this class.

Retrospective replay, prospective artefact and external reporting remain separate evidence classes.