Why market cap matters for momentum trading

What market cap tells a momentum trader

Market cap is the market value of a company's outstanding shares: shares outstanding × share price. A company with 50 million shares at $8.00, for example, is worth $400 million. Investor.gov has a short definition of market capitalization if you want the regulator's version.

Size doesn't decide whether a move continues, but it shapes the conditions you trade in: how wide the spread is, how deep the book is, how violently price reacts to news, and how likely a corporate action is to interfere. Size labels vary by source. One common rule of thumb calls companies under about $300 million micro-caps, $300 million to $2 billion small caps, $2 billion to $10 billion mid-caps and anything larger large caps. Treat these as conventions, not rules.

Small caps and large caps trade differently

These are tendencies, not laws. A $3 billion company can have a wide premarket spread on a quiet morning, and a $150 million company can trade cleanly on a busy one. The glossary entries on volatility and liquidity cover the underlying ideas.

Dilution

When a company sells new shares, each existing share represents a smaller slice of the business. Small companies whose stock has just spiked sometimes use the higher price to raise cash, and an offering announced mid-run can stall or reverse the move. For a momentum trader the risk is practical: the catalyst that lifted the price can be followed within hours by new supply that absorbs the demand.

Reverse splits

A reverse split combines shares, for example one new share for every ten old ones. The company's value doesn't change: 100 million shares at $0.50 ($50 million) become 10 million shares at $5.00 (still $50 million). But any calculation that mixes the old share count with the new price breaks badly. Multiply the old 100 million shares by the new $5.00 and you get $500 million, ten times too high and enough to clear a $500 million floor. That's why the date and source of a share count matter.

What a market-cap filter does

A market-cap filter removes candidates below a minimum size before other rules look at them. It's a coarse universe filter: it cuts out the smallest and typically hardest-to-trade companies, but it doesn't measure the spread, depth or float directly, so it works alongside those checks rather than replacing them.

Any market-cap filter has to answer four design questions:

• What happens exactly at the threshold? Does a $500,000,000 company pass a $500 million minimum?
• What happens when the figure is unknown? If a missing number counts as a pass, the filter fails precisely when the data is weakest.
• Where does the number come from, and how old is it? Offerings, buybacks and splits change share counts between reports.
• Which price is used? On a stock up 40% premarket, the current price inflates the market cap by 40%.

Momentum Lab's mandatory presets

In Momentum Lab, the market-cap filter can't be switched off. Each strategy uses one of three presets:

A stock exactly at the minimum passes. A stock without a known market cap never passes, and the skip reads "Skipped — Market cap unavailable" rather than "below minimum", so missing data is never confused with a small company. A trigger that arrives while the figure is still loading is held, not refused. The preset names describe how selective each floor is, not textbook size classes: by the rule of thumb above, a $2B floor still admits mid-caps.

Analytics can be broken down by market-cap preset, and backtests can compare presets, with one caveat: backtests apply today's market cap to every past day, so a company that has since grown or shrunk is judged at its current size.

Where the figure comes from

With Interactive Brokers data, Momentum Lab uses the first figure available, in this order:

1. IBKR. IBKR's own market-cap figure.
2. SEC EDGAR, derived. If IBKR has none: shares outstanding from the company's latest 10-K or 10-Q, multiplied by the lower of the last price and the previous close, and labelled "derived". PatternX derives a figure only when that report is no more than 200 days old, the company doesn't file on Form 20-F or 40-F, no reverse-split 8-K has been filed since the report, the issuer isn't a fund, and the symbol is the company's main listing.
3. Polygon. If neither works: Polygon's figure, for common stock only.

Alpaca's data carries no market cap, so with Alpaca data the figure has to come from Polygon, where PatternX provides a Polygon feed. ETFs on PatternX's list get no figure, so they're always skipped as market cap unavailable.

These choices lean conservative. Using the lower of two prices means the move being traded can't lift a stock over the floor by itself. For example, a company with 50 million shares that closed at $8.00 ($400 million) and trades at $11.00 premarket would be $550 million at the current price; the derived figure uses $8.00 and leaves it below the $500 million Balanced floor. The reverse-split check prevents the tenfold error described above. And 20-F and 40-F filers are foreign issuers whose reported share count may not match the shares traded in the US, for example when the US listing is a depositary receipt.

With IBKR data, the movers scan itself ranks the top gainers and losers among stocks above your market-cap floor, so the list isn't crowded with companies the filter would reject anyway.

You can check a share count yourself: the cover page of each 10-K and 10-Q states the shares outstanding, and filings can be searched on SEC EDGAR.

Limits of any market-cap filter

• A market cap is a snapshot. Offerings, buybacks and splits change the share count before the next report.
• Market cap isn't float. A company whose shares are mostly held by insiders can trade like a much smaller one.
• A higher floor reduces, but doesn't remove, gap, halt and spread risk, and it means fewer opportunities.
• Applying today's market cap to past days keeps backtests from being point-in-time accurate.

For the other liquidity gates, see relative volume and premarket volume and premarket liquidity and spreads. The Momentum Lab guide lists every skip reason, and the Momentum Lab methodology is the place to start for the engine as a whole.