How to Find Stocks to Day Trade
To find stocks to day trade, start from the whole market and remove: first what is too thin to absorb your order, then what costs too much to trade, then what does not move enough to pay for the attempt. What survives is a shortlist to watch, not a list to buy.
A screen is a subtraction, not a search. You start with everything and take away what cannot be traded — and if what is left is still fifty names, no filter has actually run.
Before you start
A free screener. Every major charting platform includes one and the free tiers are sufficient for all four filters below.
Your round-trip cost — spread plus commission, for the size you actually trade.
A calendar of scheduled events — earnings dates, rate decisions. The chart does not contain these, which is the boundary the technical and fundamental page draws.
The steps
1. Start from the whole market
Not a watchlist somebody posted. The point of a screen is that it did not know what you were hoping to find.
2. Remove what is too thin
Set a minimum average volume, and check the size available at the best bid and offer rather than only the price. On a thin book your own order moves it, which is the liquidity constraint felt directly.
3. Remove what costs too much to trade
Divide the spread by a typical bar’s range. Under a few percent is workable; a tenth is not, and the penny stocks page is where that stops being a rounding error.
4. Remove what does not move
Set a minimum average true range (ATR) as a percentage of price. An instrument that cannot travel your stop distance in a day has nothing in it for you, however good the chart looks.
5. Check the calendar on what survives
Anything reporting inside your holding period is a decision to accept a gap, not an oversight. Make it deliberately or remove the name.
6. Write the survivors down and stop
Three to five names. The screen has finished; nothing on that list is a trade until it does something at a level.
How to tell it worked
Two checks, and both take under 2 minutes.
Count what survived. 3 to 5 names is a screen. Fifty is a screen that did not run — at least one threshold is set so loosely it removed nothing, and the usual culprit is the cost filter because it is the one people leave out.
Then check that yesterday’s list would have been different. If the same names survive every day regardless of conditions, the filters are describing the instruments rather than selecting among them, and you have built a watchlist with extra steps.
And one thing that is not a check: whether the names went up. A screen that only keeps what rose yesterday is momentum selection wearing a filter’s clothes, and it will pass this test every time while telling you nothing.
The order matters
Run the filters cheapest-first and the whole thing takes minutes; run them in the wrong order and it takes an evening.
Liquidity and cost are properties of the instrument. They barely change day to day, so they can be applied once to build a tradeable universe of a few hundred names that you keep.
Movement and the calendar are properties of the day. They are the only two that need re-running, and they run against the few hundred rather than against the whole market.
That split is the actual time saving, and it is why a screen should feel like a ten-minute routine rather than a project.
How many names you can actually watch
The screen’s job is not to find the best instrument. It is to produce a list short enough that you can be paying attention when one of them does something.
That number is smaller than it feels. A level only matters at the moment price reaches it, and a name you are not watching at that moment is a name that was not on the list in any useful sense.
Three is a working number for one screen. Five is the ceiling for most people, and it stops being a shortlist somewhere around eight — at which point you are not selecting, you are scanning, and the entry procedure has nothing to attach to.
The alternative is alerts. Set them at the levels you marked and let the platform watch, which converts an attention problem into a configuration one and is the single highest-value thing most people can change about their screening routine.
The original data
Across our study of 24,971 trading videos, 105 cover finding stocks to trade. The median one gets 11,258 views, 66% never pass 50,000, and the median length is 11.6 minutes.
The corpus carries description text for 43 of those 105, and across those 43, two mention invalidation, failure, or what a bad read looks like.
About one in twenty, on a procedure whose whole output is a list of candidates that mostly should not be traded.
When it fails
The screen returned everything
A filter that removes nothing is a filter you have not set. Tighten the one that is loosest — usually cost — until the list is short enough to actually watch.
You screened for what already moved
Yesterday’s biggest movers are a list of instruments where the move has happened. That is a legitimate strategy and it is not a screen; be clear which one you are running.
The list became a commitment
A shortlist creates pressure to use it. Most days most of the names do nothing, and the correct outcome of a screen is frequently zero trades — the scarcity numbers on the entry page are why.
You judged the name instead of the level
Passing four filters means an instrument is tradeable, not that it is a trade. Everything that makes it one happens on the chart afterwards.
Related
How to read the market is what happens after the shortlist exists.
How to find an entry is the level and trigger those names have to produce.
And penny stocks is what the cost filter is protecting you from.
I run this the night before and it takes about ten minutes. The filter that removed the most bad trades was the cost one, and it was the last one I added — I spent a long time trading things that moved plenty and gave the whole move back in spread.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.