The first three parts of this series measured the bar for each list, how fast each turns over, and
who sits on them. This part asks a question a scanner setting cannot avoid: how long should a
gainers list be before it stops hiding things.
1. How many stocks actually move
Counting only stocks that closed at $1 or more and traded at least $1,000,000, here is what a
session contains.
| Move from the open | Median count | 10th percentile | 90th percentile | Busiest session |
| 5% or more | 211 | 111 | 449 | 886 |
| 10% or more | 61 | 35 | 140 | 211 |
| 20% or more | 15 | 8 | 27 | 55 |
| 30% or more | 6 | 3 | 10 | 20 |
Sixty-one tradable stocks clear 10% on an ordinary session, and on a busy one it is 211. The number
of genuinely large moves is small: six names clear 30%.
2. What a list of a given length holds
| Move | Top 30 | Top 50 | Top 100 |
| 5% or more | 14.2% | 23.7% | 47.4% |
| 10% or more | 49.2% | 82.0% | 100.0% |
| 20% or more | 100.0% | 100.0% | 100.0% |
| 30% or more | 100.0% | 100.0% | 100.0% |
A thirty-name gainers list is complete for the 20% club and covers about half of the 10% club. It
holds one in seven of the 5% names.
To hold half the 10% movers a list needs 30 rows at the median session, and to hold nine in ten it
needs 55. On the busiest tenth of sessions those become 70 and 126.
3. The dollar-volume list is not a substitute
If a trader watches the most active names by turnover and expects the day's movers to pass through,
the overlap is close to nothing.
| Share of 10% movers appearing in the top 100 by dollar volume | Value |
| Median session | 1.3% |
| 90th percentile session | 3.7% |
| Best session in the sample | 6.4% |
One name in a hundred, on a typical session. That is consistent with the rest of this series: the
two lists are built on quantities that barely interact, and a hundred rows of one of them is not a
window onto the other.
4. Choosing a length
The right length depends on which club you want complete rather than on a round number.
If the interest is 20% and larger moves, thirty rows is enough and always has been in this sample.
If the interest is 10% and larger, thirty rows shows about half and fifty shows most of it. If the
interest is 5%, no list of a reasonable length is close to complete, and a filter is a better tool
than a longer list.
The count also moves with the session. A list sized for the median day holds far less on the busiest
tenth, where 140 names clear 10%, so a fixed length quietly becomes a tighter filter exactly on the
days with the most to look at.
5. What this does not measure
We count membership, not opportunity. Nothing here says a name outside the list was worth trading,
and we did not measure returns.
Daily bars decide the move at the close, so this is what a list would have needed to contain in
hindsight. A live scanner ranks on partial information and its membership will differ, which we
measured separately in
https://traderwe.com/research/does-a-faster-stock-screener-catch-more-movers-we-tested-501-sessions
The qualified filter of $1 and $1,000,000 is deliberately loose. A stricter one reduces every count
in the first table, which shortens the list you need.
6. Sample accounting
The sample is 181 sessions between 2025-12-02 and 2026-08-21, with exchange test symbols excluded
for the reason given in part one. Moves are measured from the session open to its close.
Coverage is the share of that session's qualifying movers that a ranked list of the given length
would contain, computed session by session and reported as the median across sessions. Where a
session had fewer movers than the list length, coverage is complete by construction and enters as
100%.
