Help / Data, sources and evidence

What is point-in-time data, and why does it matter?

Data as it was known on each past date, so a test of the past does not use hindsight.

The idea

A test of how a measure behaved in 2020 may use only what was known in 2020. Point-in-time data keeps each past date’s view: the company filings filed by then, the size list in force then, the funds that existed then.

Where Anveshan applies it

Every input of the fund rank uses the history up to that month and nothing later. Book quality reads company filings as filed by each month-end. Mandate adherence and size bands use the half-yearly size list in force at the time.

The biases it avoids

Hindsight bias: using a restated figure or a later list to judge an earlier decision. Survivorship bias: testing only the funds alive today, which leaves out the ones that closed or merged. Absent figures are never treated as zero; a missing value stays missing.

What it does not do

Point-in-time care removes one source of error. It does not make a past result repeat, and historical findings are not forecasts.

In Anveshan

Related help

A description of published information, not advice. Nothing on Anveshan is a recommendation to buy, sell, switch or hold an investment. All help · How the evidence is read