Our “weekly trend” is a 150-day average, not a weekly scan
Thirty weeks is about 150 sessions. We compute it the daily way, and the difference is not cosmetic.
Several StockLearn pages and guides refer to a stock's "weekly trend". That phrase is easy to misread, and the misreading matters. There is no separate weekly scan. What we call the weekly trend is a 30-week moving average computed from daily closing prices — a 150-session average, calculated in the same daily pipeline as everything else.
This piece explains what the distinction is, why we settled on the daily-derived version, and what it means if you've read older material that implied otherwise.
Two ways to build the same idea
Weinstein's stage analysis, which our stage classification draws on, is conventionally described using a 30-week moving average on a weekly chart. There are two ways to compute that.
The first is to build genuine weekly bars — group the daily data into weeks, take each week's closing price, then average the last thirty of those. The second is to stay in daily data and take a 150-session moving average, since thirty weeks is about 150 trading sessions.
The two produce very similar lines and near-identical stage classifications most of the time. But the first requires a whole parallel data structure, and that structure creates problems.
Why the weekly-bar version causes trouble
The awkwardness is the current, incomplete week. On a Wednesday, this week has three sessions in it. Do you treat that as a bar? If you do, the most recent point in your average is built from partial data and shifts every day until Friday, so a stage can flip mid-week and flip back. If you don't, your trend line is up to four sessions stale, which is worse on a fast move.
Neither answer is satisfying, and both create a category of support question that boils down to "why did the label change when the price didn't do much?" Holiday-shortened weeks make it worse. Indian markets have plenty of those.
A 150-day average has no such ambiguity. It updates once per session, uses only completed sessions, and every point on it is built the same way as every other point.
Method
The trend line is a simple moving average of the last 150 daily closing prices, computed from NSE end-of-day bhavcopy data on adjusted closes. It updates once per trading session. There is no weekly bar construction anywhere in the pipeline.
The gate that used to exist
Earlier versions of the scanner did carry a separate weekly gate — a condition evaluated against weekly-constructed data, layered on top of the daily checks. It was retired at gate version 4.
It was removed for the reasons above, plus one more: it was a second source of truth. When the daily and weekly views disagreed at the edges, we had to decide which one won, and any answer to that was arbitrary. Collapsing to a single daily-derived measure removed a whole class of inconsistency.
If you've read older material
Anything that describes StockLearn running a distinct weekly scan is out of date. Where our current guides say "weekly trend" or "30-week average", they mean the 150-day daily-derived line. If you find a page that still implies a separate weekly pass, that's a leftover and we'd like to know.
What this changes in practice
For a reader, very little — the line looks the same and the stage labels are the same most days. What changes is behaviour at the margins: labels no longer wobble mid-week, and they don't lag by several sessions after a sharp move either.
What it changes for us is bigger. One data path means one place for a bug to live, one definition to document, and one thing to test. The pipeline already handles enough special cases — symbol changes, corporate actions, delistings — without a parallel timeframe multiplying each of them.
Does the choice affect the measurement?
It could, in principle. If the weekly-bar version produced meaningfully different stage classifications, then any research built on those stages would differ too. In our data the two constructions agree on the large majority of stock-days, and the disagreements cluster exactly where you'd expect — stocks sitting right on the line, where the classification is marginal anyway.
Those marginal cases are also the ones where the label carries least information, which is worth remembering generally. A stock 0.2% above its 150-day average and one 15% above it both read as "above the trend line", and they are not the same situation.
The naming problem
Honestly, "weekly trend" is a slightly misleading label for a daily-computed number, and we've kept it mainly because thirty weeks is how the concept is known. The precise phrasing would be "the 150-session moving average, which approximates the conventional 30-week line."
Where it matters — in methodology guides and anywhere we report a measurement — we try to use the precise version. This page exists so the shorthand elsewhere doesn't mislead anyone.
Key takeaways
- There is no separate weekly scan — the “weekly trend” is a 150-session moving average of daily closes.
- Weekly-bar construction forces a bad choice about the current incomplete week; a 150-day average avoids it.
- A separate weekly gate existed in earlier versions and was retired at gate version 4.
- The two constructions agree on most stock-days; disagreements cluster at the margin, where the label means least.
- Older material implying a distinct weekly pass is out of date.
See these ideas on real stocks
StockLearn runs this read on ~2,000 NSE stocks every evening. Nifty 50 is free, no login.
Browse today's scan →This guide is educational and explains how StockLearn interprets common technical indicators, using illustrative examples. It is not investment advice or a recommendation to buy or sell any security.