How the research works
NEPSE has no shortage of opinions. The methodology here exists to answer a narrower question: did this idea actually work, tested honestly, against data it couldn't have seen coming?
Most market analysis is built and judged after the fact — an idea gets tested against data the analyst already knows the ending of. That's not a small technicality: it's the difference between finding a real pattern and finding a story that fits what already happened.
Every research call made here is tested against only the data that existed at the moment the call would have been made. Corporate actions, later revisions, and anything discovered afterward are excluded by construction, not by discipline alone.
The five-step process
Data
NEPSE trade & floorsheet data, captured and validated — point-in-time. Nothing is back-filled with information that wasn't yet public.
→
Test
An idea is researched against only what was knowable at the time it would have been made — no lookahead into future prices, news, or corporate actions.
→
Record
The call is logged and frozen before its outcome is known: the entry conditions, and the reasoning behind them, can't be revised after the fact.
→
Forward result
The market plays out. The outcome is observed as it happens — never edited, never quietly dropped if it goes the wrong way.
→
Evidence
Once a call's forward window matures, it's graded automatically — against a random baseline, not a cherry-picked comparison — and the grade becomes part of the permanent record.
A call isn't judged as "right" or "wrong" in isolation — it's compared to what a random pick would have done over the same window. Beating the market's own noise is a much higher bar than beating nothing, and it's the bar every strategy here is actually held to.
Grading happens once, automatically, when a call's forward window matures. It is never re-run with hindsight, never re-graded after the fact, and never quietly removed from the record because it didn't work out.
Most ideas don't survive contact with real data
The honest majority of approaches tested against this discipline fail to clear real trading costs, or turn out to be an artifact of one regime rather than a durable pattern. That record is kept, not discarded — a research process that only remembers its wins isn't a research process, it's marketing.
See the Journal for specific, dated findings — including the ones that didn't hold up.
This page describes a process, not a promise. Nothing here is investment advice — see the FAQ.