This article explains how a systematic pre-market scan works. It is educational, not personalised investment advice, and the ranked output it describes is research, not a recommendation to buy or sell any security.
Most people picture a stock scanner as a filter: set a few conditions, price above the 50-day, volume up, RSI under 70, and get back the names that match. That is a screener, and it answers a yes/no question. A pre-market scan that produces a daily ranked list is doing something harder: computing a broad set of measurements for every name in a universe, then ranking the whole field by how strongly each one fits, before the market opens. The difference between a filter and a ranked scan is the difference between “which names qualify” and “of the names that qualify, which are the strongest, and why.”
What a real scan measures
A serious scan doesn’t look at one or two conditions. It computes dozens of measurements per name from its price and volume history, not because more is automatically better, but because a single indicator is too easy to fool. The measurements fall into a handful of families:
- Trend & position. Where price sits relative to its key moving averages, and whether those averages are stacked in order, a read on whether the name is in an established uptrend or just bouncing.
- Volatility. How much the name typically moves in a day (average daily range, ATR-style measures). This sizes everything downstream: the same dollar stop means something different on a quiet name than a wild one.
- Momentum & thrust. Recent return over several horizons, and whether the move is accelerating or stalling, the difference between a name that’s already run and one just getting going.
- Structure. How tight or wide the recent base is, how far price is from its highs, how it sits within its range, the shape of the setup, not just its direction.
- Liquidity. Dollar volume, is the name actually tradeable at size, or is it a thin ticker that looks great on a chart and can’t be entered without moving the price?
None of this requires an opinion. Every figure is arithmetic computed the same way for every name, every day, the enrichment step of the pipeline in how a stock signal is made. The breadth matters because it lets the scan see a name in the round: strong trend but no liquidity, or great structure but already extended, are different propositions, and one or two indicators can’t tell them apart.
From measurements to a single ranked list
Dozens of numbers per name is data, not a decision. The scan’s real output is a ranking, which means collapsing all those measurements into one comparable score so the entire field can be sorted. The discipline here is not to let any single measurement dominate, a name shouldn’t top the list purely because it’s the most volatile, or purely because it’s closest to its highs. The measurements are weighted so that a high rank requires several things to be right at once. How that composite is built, and why several weighted features beat any single indicator, is the subject of composite scoring.
A raw ranking still isn’t enough, because the same score means different things in different contexts. A name freshly emerging from a long base is a different proposition than one that has been extended for weeks, even at an identical score, the distinction drawn by the setup-state classifier. And a strong-looking list in a deteriorating market is worth less than the same list in a broad advance, which is why the scan is read against the prevailing market regime.
Why pre-market, and why ranked
The timing is deliberate. Running the scan before the open means the ranked list is built on settled, complete prior-day data plus the opening context, a stable picture, not a moving target chased intraday. It is a calm, once-a-day read rather than a flickering real-time feed demanding constant reaction.
And the ranking, not just the filtering, is the point. A filter hands you fifty names that “qualify” and leaves you to guess which matter. A ranked scan hands you the field in order, with the score and the context attached, so you can see where the strength is and how steeply it drops off. It is research output you can read top-to-bottom, not a pile of matches, and it is a list, not an instruction. What you do with the ranking is your decision; the scan’s job is only to do the measuring, the same way, every day.
Filter vs. ranked scan, in one line
A screener tells you which names pass a few conditions. A ranked pre-market scan measures every name many ways, weighs the measurements so no single one runs away with it, places each name in its setup-state and market-regime context, and hands you the whole universe sorted by strength, before the bell. The first is a yes/no question. The second is a daily, ordered piece of research. Shishin publishes the second.