Walk-forward tracking

A backtest is hindsight; this page is the opposite. Every list this site has published is measured after the fact: an equal-weight basket of the names at their publication-day closes, held untouched, compared with the segment's benchmark index over the identical window. Nothing here can be tuned or back-filled — the record simply accumulates with each refresh, every trading day.

Published lists — Nifty 500

30 lists shown28 currently ahead of the benchmark2 currently behind— a snapshot that changes with every refresh.

Published	Tracked through	Days	Stocks	List	Nifty 500	Difference	Worst dip
25 Aug 2026	26 Aug 2026	1	20	-0.1%	-0.2%	+0.1%	-0.1%
24 Aug 2026	26 Aug 2026	2	20	+0.0%	+0.2%	-0.2%	-0.1%
21 Aug 2026	26 Aug 2026	5	20	-0.0%	+0.1%	-0.2%	-0.2%
20 Aug 2026	26 Aug 2026	6	20	+1.3%	+0.2%	+1.1%	-0.2%
19 Aug 2026	26 Aug 2026	7	20	+1.2%	+0.7%	+0.5%	-0.1%
18 Aug 2026	26 Aug 2026	8	20	+0.8%	+0.4%	+0.4%

“Worst dip” is the basket's deepest peak-to-trough fall since publication — the part of the ride a return figure hides.

The list published on 26 Aug 2026 is on record but is not in the table yet: a list can only be measured once a later session has closed, so it joins the record at the next refresh.

How to read this record

Tracking began in July 2026. A record this young says very little either way — short windows are mostly noise. It is published from day one so that it cannot be curated later.
Figures are price returns on adjusted closes with no transaction costs and no taxes; real-world results would be lower.
Each segment is compared against its own benchmark price index where a free daily series exists (Nifty 50, Next 50, Midcap 150, Nifty 500). No free series exists for the Smallcap 250 index, so that segment compares against the Nifty 50 — a substitution stated rather than hidden. All benchmarks are price indices and exclude dividends.
Note (13 Aug 2026) — two gaps in this record, stated rather than hidden. Before 16 Jul 2026 the refresh ran weekly, so 13–15 Jul carry no list; the daily lists on 6–10 Jul came from manual runs while the site was being built. From 16 Jul the job ran later in the day and published the current session instead of the previous one, so the Mondays of 20 and 27 Jul and 3 and 10 Aug were never published. Fixed on 13 Aug 2026: lists are computed from the latest completed session and the job now runs twice a day. Two lists, 11 and 12 Aug, were computed from mid-session prices and are left as published — re-ranking a list after the fact is precisely what this record must never do. Nothing was deleted or back-filled; the gaps stay on record.
This is a descriptive record of a published screen, not a recommendation, a rating, or advice of any kind. How the formula works is on the methodology page.