Why Most Day Traders Lose: The Numbers Behind the Attrition

Three months in, your account is down $600. Scroll back through the trade log and the green fills are right there — the morning gap you faded, the breakout you caught at 10:15. You were right more often than you were wrong, and the balance still moved the wrong way.

That gap between “I had winners” and “my money shrank” is not a discipline problem. It is arithmetic, and almost nobody runs it before funding the account. Trading costs are a fixed toll charged on every round trip, win or lose. The edge that is supposed to pay that toll is small, uncertain, and measured in pennies per share. Multiply both sides by three hundred trades a month and the certain number usually beats the uncertain one.

The toll booth you pass twice per trade

Every round trip carries three separate frictions, and only one of them ever appears on a statement.

The spread comes first. You buy at the offer and sell at the bid, so a stock quoted 40.02 by 40.03 costs a penny a share just to get in and back out at the quoted prices. Commissions come second, and the phrase “zero commission” is doing real work: brokers charging nothing per trade are generally paid by wholesalers for routing order flow, which moves the compensation into the fill price instead of onto the invoice. Slippage is third — the distance between the price on your screen when you clicked and the price you actually got. It grows with order size, with volatility, and with how badly you want out.

Individually these are rounding errors. A penny a share on 150 shares is a dollar fifty. Nobody quits over a dollar fifty. But friction is charged per transaction, not per profitable transaction, so the variable that decides everything is how many transactions you make. An investor who buys four positions a year pays the toll eight times. A day trader running fifteen round trips a day pays it six thousand times a year.

One month, fully specified

Call her Farida. She has $10,000 in a margin account and trades a single liquid mid-cap name around $40 a share, in 150-share lots, so roughly $6,000 of exposure is on at any moment. Fifteen round trips a day, twenty trading days a month. Three hundred round trips.

Her execution is decent. She is not chasing, she uses limit orders where the tape allows, and her all-in friction works out to a quarter of a cent per share on each side. On 150 shares that is 37.5 cents in and 37.5 cents out: 75 cents per round trip. Across 300 round trips, $225 for the month.

Her results are decent too. She is right 52% of the time. Winners average 12 cents a share, losers average 11 cents — she is cutting losses slightly shorter than she lets winners run, exactly as instructed.

Run it out:

• 156 winning trades × $18 = $2,808
• 144 losing trades × $16.50 = $2,376
• Gross P&L: +$432
• Costs: −$225
• Net: +$207

Farida made 2.07% on her account in a month while being right barely more than half the time. Annualize that and it sounds spectacular. This is the version of the story that gets posted.

Now look at what she is actually living on. Gross expectancy is $1.44 per round trip — $432 spread across 300 trades. Cost is 75 cents per round trip. She keeps 69 cents. Her whole business runs on a 69-cent margin per transaction, and the 75-cent side is the only part she knows in advance.

What flips the sign

Farida breaks even at a cost of $1.44 per round trip, which on 150 shares works out to 0.48 cents per share each way. Under half a cent. That is the entire buffer between a profitable month and a wasted one. Watch what a single moved assumption does to it.

Two percentage points of win rate — three trades a month landing the other way — puts her at exactly zero. Four points puts her at −$207. Widening her effective spread by a quarter of a cent per share, which is what a slightly thinner ticker or a fast tape does to her, sinks the month without her changing a single decision.

Notice what none of those rows require. In every one Farida still wins about half her trades or better. The folk wisdom that a losing trader must be a bad trader with a bad hit rate is simply wrong. She can be right most of the time and still hand the month to the firm on the other side of her fills, because her edge is measured in fractions of a cent per share and so is her cost.

The obvious fix — trade more, since the strategy works — does nothing to the ratio. Push Farida to 25 round trips a day and gross goes to $720 while costs go to $375. Net $345. Both sides scale together. Frequency amplifies whatever per-trade expectancy already exists; it cannot manufacture expectancy from nothing. If her real edge is 60 cents a trade against a real cost of 75, trading more often gets her to zero faster.

Why the regulator put a floor under this

Anyone executing four or more day trades within five business days in a margin account gets flagged as a pattern day trader and must maintain at least $25,000 of equity in that account. Drop below the threshold and the account is restricted until it is funded back up.

So Farida, at $10,000, is not eligible to trade this way in a margin account at all. That is the point of the rule. It is not a verdict on her ability. It recognizes that a strategy whose margin of safety is half a cent per share, run with leverage, in an account small enough that one bad afternoon is material, produces a risk profile a small account cannot absorb. The mechanics — including how the day-trade count is tallied — sit in FINRA’s rule on pattern day trading.

What the account-level data shows

The most useful evidence on how often this works out comes from Taiwan, for a boring reason. Taiwan Stock Exchange data let researchers see every account and every transaction on the exchange, rather than a survey sample of people who chose to answer questions about their own trading.

Brad Barber, Yi-Tsung Lee, Yu-Jane Liu and Terrance Odean worked that data across several papers, including Do Day Traders Rank Among the Top Performers? and The Cross-Section of Speculator Skill: Evidence from Day Trading. The finding was blunt and it repeated: the large majority of day traders lost money net of costs. A small minority were profitable — and here is the part that gets dropped from summaries — that minority tended to stay profitable in later periods. Persistence across years is the standard test separating skill from a lucky run, and by that test the skill is real. It just belongs to a small group.

The two-population pattern turns up wherever someone gets full account-level visibility instead of self-reports. One honest limit: the strongest datasets are non-US, because that is where regulators had transaction-level access to an entire market. Nothing comparable for US retail brokerage accounts is public. Read the cross-market result as directionally informative, not as a US statistic.

Why the drip stays invisible

Farida’s $225 of monthly friction never shows up as a line item she has to confront. It arrives as 300 slightly worse fills. Each one looked fine at the time, on a trade she was thinking about for other reasons entirely.

Her wins arrive differently. Each is an event — a green number, a moment of being right, a small hit of confirmation, fifteen times a day at unpredictable intervals. That is the reinforcement schedule slot machines are built around. Wins are vivid and episodic. Costs are flat and continuous. Human memory does not weight those two things correctly, and no amount of knowing that fixes it.

A second distortion is not psychological at all. It is sampling. People who lose for four months and quietly close the account rarely write about it. People who had a good March post the screenshot. Everything visible about day trading online is filtered through who chose to speak, which makes the visible distribution of outcomes and the actual distribution of outcomes two different objects.

What this does not tell you

It does not say day trading cannot be done profitably. Every full-population dataset researchers have examined contains a persistent-winner group, and the literature does not settle what separates them — information, execution speed, risk control, capital, or something unmeasured. “Most people lose” and “nobody can win” are different claims, and only the first one is supported here.

Farida is an illustration with stated assumptions, not a benchmark pulled from a study. Her 52% win rate, her 12-cent winners, her quarter-cent-per-side friction — all chosen to be plausible, none measured. Your own figures will differ by broker, by ticker, by order type, by hour of the day. The structure of the arithmetic transfers. The specific numbers do not.

Taxes are missing entirely, and for frequent trading that is a large omission. Short-term gain treatment, wash sale rules, and account type each move the net figure materially, and that is a separate calculation from the trading math. Position sizing and leverage are also unmodeled beyond noting that Farida runs $6,000 of exposure against $10,000 of equity. Change that ratio and the volatility of every row in the table changes with it, in both directions.

FAQ

Is day trading illegal or restricted for retail investors?

It is legal. But once you execute four or more day trades in five business days in a margin account, US rules classify you as a pattern day trader and require at least $25,000 of equity in that account, per FINRA’s pattern day trading rule. Below that line the account gets restricted. The threshold exists because of the risk profile of frequent leveraged trading, not because the activity is banned.

Can you have a high win rate and still lose money?

Yes, and it is the most common route to a losing account. Farida wins 52% of her trades and lands at exactly break-even if that slips to 50%, or −$207 at 48% — while still winning nearly half the time. Win rate only means something in combination with average win size, average loss size, trade frequency, and cost per round trip. Any one of those four can flip the sign by itself.

What counts as a day trade, exactly?

Buying and selling — or short selling and buying to cover — the same security within the same trading day, in a margin account. Holding overnight breaks the pair, so it does not count. The formal definition is on investor.gov’s glossary entry for day trading.

Why do some traders stay profitable for years while most don’t?

The Taiwan account-level studies found profitability persisting for a small group across multiple years, which is the standard test distinguishing skill from luck. What produces that skill is unsettled. It is unlikely to be one factor, and the datasets that identify the group are not granular enough to explain it.

Does this apply to swing trading or long-term investing?

The friction problem scales directly with trade count, so it lands very differently. Farida pays her toll 3,600 times a year. Someone rebalancing a portfolio four times a year pays it eight times. Nearly three orders of magnitude of difference in cumulative friction, before comparing anything about the strategies themselves.

The one calculation worth doing

If you already trade this way, the useful exercise is tedious rather than clever. Export your fills. Total realized spread, commissions, and slippage as a single number, kept separate from gross P&L, the way Farida’s month is broken out above. Divide that total by your round-trip count to get real cost per trade. Divide gross P&L by the same count to get real gross expectancy.

If the second number is not comfortably larger than the first, nothing about your entries or your indicators is the binding constraint. FINRA’s investor education material and the SEC’s investor.gov resources cover account requirements and risk disclosures in more detail.

This article is general information, not financial advice. See our disclaimer.

More in Trading · all topics · calculators · how this was checked