Intraday Trading: Cost Times Frequency

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-09-06 · Updated 2026-10-03 · Reviewed 2026-10-03

How it works

Everything is closed before the session ends. No positions are carried overnight, so nothing is exposed to what happens while the market is shut.

That is the definition, and everything else about the style follows from two consequences: no overnight risk, and many more trades. A third consequence is easy to miss: no share of whatever the market does between the close and the next open, which in SPY since 1993 has been most of the price gain.

The arithmetic

Every round trip pays a fixed cost: the spread plus any commission. What matters is its size against the moves you trade, and that size changes through the day.

Quiet bars make the same cost much bigger. A cost that is a small slice of a busy opening bar can be several times that share of a slow afternoon bar. The cost does not shrink when the market goes quiet.

Multiply the cost by the number of trades and you have the hurdle. It is computable before you place a single trade, and it does not depend on any view about markets, only on your costs and your intended frequency.

The same edge held longer pays the cost far less often. A swing trader capturing a move over five days pays one round trip; an intraday trader capturing the same move in fragments pays several. That is the structural argument for longer holding periods, and it has nothing to do with being right more often.

A worked example

Take a hypothetical cost of $0.02 a share for a round trip in SPY, on 100 shares, so $2.00 a trade. Over the 60 sessions from 10 July to 2 October 2026, the median five-minute SPY bar spanned $0.51 from high to low. The $0.02 is 3.92% of that.

Now add frequency. Four round trips a session for 252 sessions is 1,008 round trips a year, or $2,016.00 on 100 shares. Measured in bars, 1,008 trips at 3.92% each is about 39.5 median five-minute bars of travel handed over every year before anything is earned.

Time of day moves the number. The same $0.02 is 2.27% of the median bar from 9:30 to 10:00, where the median was $0.88, and 5.56% of the median bar from 3:00 to 3:30 PM, where it was $0.36. Against a whole session the cost is small: the median SPY day in those 60 sessions spanned $5.30, so $0.02 is 0.38% of it. That gap, between the cost per bar and the cost per day, is why trading frequency matters more than the headline commission.

What you get in exchange

No overnight gap risk, and this is genuinely valuable. A gap can move price past a stop without trading at it, so the loss on an overnight position is not bounded by the stop distance. Closing before the bell removes that.

The price of that protection is the overnight return itself. In SPY since 1993, the hours the intraday trader gives up are the hours in which most of the fund’s price gain happened, as the trading vs investing comparison measures session by session.

Volume and range decide which hours are usable, and they are not the same thing. In the July to October 2026 sample the busiest five-minute bars came in the last half hour, a median of about 831,000 shares each, yet the bars from 3:00 to 3:30 PM were the narrowest of the day.

In practice: the constraints nobody puts in the brochure

Your results are limited to the hours you can be present. That is a constraint about your life rather than about the market, and it is the one that decides whether the style is available to you at all.

In the United States, the pattern day trader rule restricted margin accounts under $25,000 to a limited number of same-day round trips per rolling five days. FINRA Regulatory Notice 26-10, published 20 April 2026 and read on 3 October 2026, replaces those day trading margin requirements with intraday margin standards from 4 June 2026, and lets firms phase the change in until 20 October 2027. Some brokers may still cap frequency the old way until then, so check yours.

In the US, short holding periods are taxed as short-term gains rather than at long-term rates. Taxes on trading covers the general shape, and the details depend on your circumstances.

Tighter stops are hit more often. The instrument’s volatility has not changed; the room given to it has.

And the accounting is asymmetric in how it feels. A good day is one number you remember; a thousand round trips is a cost you never see itemized against the trades that paid it.

What intraday trading is not

It is not lower risk because positions are shorter. It is lower overnight risk with more frequent exposure to execution costs.

It is not the same as scalping. Scalping is a subset with the shortest holds and the highest frequency, where the cost problem is most severe.

It is not free of the overnight market. The opening gap still sets where your day starts.

And it is not a full-time income by default. The frequency, the costs and the hours are constraints that exist before any question of edge.

The original data

Start with what the style gives up. The trading vs investing comparison splits every SPY day since 1993 into the session and the night, and most of the price gain arrived at night. Two checks here extend it to what an intraday trader in particular misses, using Yahoo Finance’s SPY daily bars and dividends from 29 January 1993 to 2 October 2026, 8,476 sessions each paired with the prior close.

• The dividends fall overnight too. SPY went ex-dividend 136 times, and each price drop lands between a close and the next open. Adding each dividend back on its ex-dividend morning lifts $1 held only overnight from $13.90 on price to $25.37, while $1 held only from open to close ended at $1.26. A trader who is flat every night collects none of those payments.
• It was not one stretch of years. The session moves added up to a loss in 13 of the 34 calendar years counted, against 5 years for the overnight moves with dividends.

The early opening prices need a caution. In 1993, 33 of 233 opens equal the prior close exactly, and in 1994, 24 of 252 do; from 1996 on, no year has more than 13. That may be old data copying the close into the open, and this page could not check it against a second source. Leaving the early years out changes the size, not the direction: from 2 January 1996 the session part grew $1 to $1.11 and the overnight part to $11.26 on price, and from 3 January 2000 to $1.31 and $4.00, or $6.40 with dividends. The full daily table is published with every value.

Inside the session, range is not spread evenly. The best time to day trade page measures each hour’s average range and share of volume; the measure here is smaller, the median span of a single five-minute bar, high minus low, which is the unit a fast trader pays the cost against. Over the 60 sessions from 10 July to 2 October 2026, 4,680 regular-hours five-minute bars, the median by half hour was:

The narrowest bars came from 1:30 to 3:30 PM, not at midday. The same $0.02 was 2.27% of the median bar from 9:30 and 5.56% of the one from 3:00 PM, and in every half hour from 1:30 to 3:30 PM it was more than twice the opening share. These are 60 sessions of one fund in one season, a small sample, and the half-hour table is published so it can be redone on a longer stretch.

The subject draws plenty of attention. Of the 24,971 trading and investing videos measured for this site, 960 have “day trading” in the title, from 477 channels, at a median of about 16,076 views; 122 say “intraday”, at a median of about 23,381.

When it fails

In a quiet range the cost is most of what is available. Small bars, small targets, and a fixed round trip that does not shrink with them.

The second failure is frequency creep. A method designed for two trades a day becomes six on a slow morning, and the cost multiplies while the edge does not.

A third is trading the narrow hours because you are at the screen and the screen is on. In the sample above that was the afternoon from 1:30 PM, where the same cost was more than twice the share it was at the open.

A fourth is the tight-stop illusion. A smaller stop feels like less risk and produces more stop-outs, and the total risk taken can easily be higher.

And a fifth is comparing your results to a longer-horizon trader’s without adjusting for what each held. The holder of an index fund collected the overnight move; the intraday trader, by design, did not.

How this connects

Order choice changes the cost side. A market order pays to get in immediately; a limit order can avoid paying away the spread but risks missing the move entirely.

Vwap is the reference many intraday traders watch. It is an average weighted by where volume actually traded that day, so it resets each morning in step with the style.

A trading journal is the reliable way to see frequency creep, since each extra trade feels reasonable in the moment, and expectancy after costs is the number it should end on.

Related

Day trading is the same idea treated as a business decision. Scalping is the highest-frequency version, where the cost problem is worst. And swing trading is the alternative that pays the same cost far less often and keeps the overnight move.

This page is educational, not financial advice. Test every idea on your own charts before risking money.