Day Trading Failure Rate:
What 30 Studies Actually Found
Between 70% and 97% of day traders lose money, depending on the market, time period, and how “day trader” is defined. After reviewing 30 studies spanning 8 countries and 25 years — plus our own dataset of 15.1 million scored stock-days — the consistent finding is: the vast majority of discretionary day traders lose money, while systematic signal-following at longer horizons (60+ days) produces meaningfully better outcomes.
Out of 100 day traders
70–97 lose money
What Every Study Found
14 studies with quantitative failure rate data, ordered by failure rate. Each dot represents one study — hover for details.
It Doesn't Matter Where
Across 8 countries and 25 years of research, the pattern holds everywhere.
Brazil
1–97%
2 studies · 2020, 2025
Taiwan
1–95%
3 studies · 2004, 2014, 2020
India
92.8%
1 study · 2024
France
89%
1 study · 2014
United Kingdom
82%
1 study · 2016
Spain
82%
1 study · 2016
United States
1–80%
3 studies · 1999, 2003, 2000
Finland
80%
1 study · 2005
It Doesn't Matter When
From 1999 to 2025, the failure rate hasn't improved. More traders enter every year — and most still lose.
Dot size represents sample size. The failure rate has not improved over 25 years of research.
Why the Numbers Range from 64% to 97%
Every article cites a different number. Here's why — and which ones to trust.
| Factor | Lower End (64–72%) | Higher End (93–97%) |
|---|---|---|
| Market | US equities during a bull market | Brazilian futures, Indian F&O derivatives |
| Definition of "day trader" | Anyone at a day trading firm (includes casual traders) | Persistent traders who traded 300+ days |
| Observation period | Single quarter or year | Multi-year tracking (2+ years) |
| What counts as "loss" | Any negative P&L (even -$1) | Cannot earn more than minimum wage |
| Cost inclusion | Gross returns (before commissions) | Net returns after all trading costs |
| Market conditions | Bull market (1998–1999 dot-com era) | All conditions (bull, bear, crash) |
Market
64–72%
US equities during a bull market
93–97%
Brazilian futures, Indian F&O derivatives
Definition of "day trader"
64–72%
Anyone at a day trading firm (includes casual traders)
93–97%
Persistent traders who traded 300+ days
Observation period
64–72%
Single quarter or year
93–97%
Multi-year tracking (2+ years)
What counts as "loss"
64–72%
Any negative P&L (even -$1)
93–97%
Cannot earn more than minimum wage
Cost inclusion
64–72%
Gross returns (before commissions)
93–97%
Net returns after all trading costs
Market conditions
64–72%
Bull market (1998–1999 dot-com era)
93–97%
All conditions (bull, bear, crash)
Both numbers are real. They're answering fundamentally different questions. The 64% figure asks “did your account end positive?” during a historic bull run. The 97% figure asks “can you earn a living doing this?” across all market conditions. The second question is the one that matters.
Why Do 95% of Traders Lose Money?
The “95%” figure is commonly cited but not precisely sourced to a single study. The actual range across peer-reviewed research is 70–97%. The reasons are consistent across all studies:
- Transaction costs consume gross profits. Barber et al. (2009) found that Taiwan's individual investors lost the equivalent of 2.2% of the country's GDP annually — mostly to commissions and bid-ask spreads. Even traders who pick winners lose money after fees.
- Overconfidence drives excessive trading. Odean (1999) showed that stocks investors purchased underperformed those they sold by 3.3% over one year. Barber & Odean (2001) found men traded 45% more than women and earned 2.65 percentage points less per year.
- The disposition effect — holding losers, selling winners. Linnainmaa (2005) found that Finnish day traders' seemingly high returns were illusory because they refused to realize losses, creating a misleading picture of profitability.
- No persistent skill for the vast majority. A 15-year study of 360,000 day traders on the Taiwan Stock Exchange found less than 1% demonstrated reliable profitability after fees (Barber, Lee, Liu, & Odean, 2014). The top 500 traders (0.14%) earned 37.9 basis points per day after costs — but they are extreme outliers.
How Much Do Most Day Traders Lose?
The average loss varies dramatically by market:
- France (CFDs/Forex): Average loss of €10,887 per client over 4 years (AMF, 2014)
- India (F&O): Average loss of ~₹2 lakh (~$2,400) per trader per year. Aggregate losses exceeded $21.5B over three years (SEBI, 2024)
- US (day trading firms): The average day trader needed $464/day ($111,360/year) just to break even after commissions (U.S. Senate, 2000)
- Brazil (COVID era): Aggregate gross losses of R$9.9 billion (~$1.8B) across 968,512 individuals who day traded during 2020–2023 (Chague & Giovannetti, 2025)
What Is the 1% Rule in Day Trading?
The “1% rule” has two distinct meanings in trading:
- Risk management rule: Never risk more than 1% of your total account on a single trade. This limits the damage of any single loss and prevents account blowups.
- The academic finding: Less than 1% of day traders demonstrate persistent, reliable profitability. Barber et al. (2014) found that out of 360,000 day traders in Taiwan over 15 years, fewer than 1% showed genuine skill after accounting for transaction costs.
The irony is instructive: the risk management “1% rule” exists because the academic “1% finding” is true — most traders will lose, so managing how much you lose per trade is critical.
Original Research: 15 Million Scored Stock-Days
We applied our Ripeness Score algorithm to 12 years of market data across ~5,000 stocks — then validated with 5,900 live production signals.
According to Banana Farmer's analysis of 15.1 million scored stock-days over 12 years, Ripe-rated signals showed a 57.7% win rate at 60 days — and live production data confirmed a 58.5% win rate at 3 months across 5,547 measured signals.
By contrast, academic studies find 70–97% of discretionary day traders lose money.
Win Rate by Signal Type and Holding Period
Backtested data: ~29,000 samples per cell, 2014–2026. Outliers trimmed. Win rate = % of signals where forward return > 0.
Short-term (5-day) signals are essentially coin flips across all badges. The signal works at 60 days.
Live Production Signals: Win Rate by Holding Period
5,900 real signals captured by the live system (September 2023 – March 2026). These are not backtested — these are actual calls with measured outcomes.
Win rate climbs steadily from 48% at 1 day to 58.5% at 3 months. The longer you hold a high-scored signal, the better it performs.
What This Data Does — and Doesn't — Show
- Short-term signals are NOT an edge. 1–5 day win rates are 44–50% across all badge types — essentially a coin flip. This directly supports the academic finding that day trading doesn't work.
- Patience is the differentiator. At 60 days, Ripe signals hit 57.7% win rate. Live data confirms 58.5% at 3 months. The scoring system identifies momentum that plays out over weeks and months, not minutes and hours.
- Ripening outperforms Ripe at 60 days (63.9% vs 57.7%). Stocks still building momentum outperform those at peak. The early signal is the better signal.
- Performance varies by year. Win rates ranged from 42% (2020) to 77% (2019). The system works best in trending markets and struggles in choppy or bear markets. This is expected for any momentum-based approach.
Methodology
Academic Study Selection
We identified studies through SSRN, NBER, Google Scholar, and regulatory body publications (SEC, ESMA, SEBI, AMF, FCA, CNMV, NASAA). Inclusion criteria: must report a quantitative measure of day trader or active trader profitability with a defined sample. We excluded literature reviews that cited no original data (one was identified and removed). Final set: 30 studies spanning 1999–2025 across 8 countries.
Verification Process
Every study underwent two rounds of fact-checking. Round 1: web-based verification of titles, authors, publication venues, and key findings. Round 2: PDF-level verification — agents fetched and read the actual papers to confirm specific numbers, sample sizes, and methodology details. Seven corrections were found and applied. All corrections are documented in our public repository.
Original Data: Backtested Scores
We applied our Ripeness Score algorithm to ~4,989 stocks daily from March 2014 to March 2026 (~15.1 million scored stock-days). The scoring engine uses technical indicators (RSI, EMA20/50, MACD, CoilScore, ATR) and momentum signals (returns, volume). Social sentiment data was not available historically and was excluded from backtested scores (weight redistributed to technical and momentum factors).
Forward returns were measured at 1-day, 5-day, 20-day, and 60-day windows. Max Favorable Excursion (MFE), Max Adverse Excursion (MAE), and days-to-MFE were computed for each window. SPY returns over the same windows serve as the benchmark.
Original Data: Live Production Signals
From September 2023 to March 2026, our production system captured 5,914 real-time signals with scores ≥75 (“Ripe”). Outcomes were measured at 6 milestone windows: 1 day, 3 days, 1 week, 2 weeks, 1 month, and 3 months. 5,901 of these signals have measured outcomes. These are live calls, not backtested.
Limitations and Caveats
- Survivorship bias: Backtested data only includes the ~4,989 assets currently tracked. Stocks that were delisted, went bankrupt, or were removed during 2014–2026 are excluded.
- No social signal in backtest: Historical scores use technical and momentum signals only. Live scores include social sentiment from Reddit and X.
- Outlier sensitivity: Raw return averages are contaminated by penny stock moves (>1,000%). We report trimmed statistics (returns capped at -90% to +500%) and medians alongside means.
- Regime dependence: Year-to-year win rates range from 42% to 77%. The system performs best in trending markets and underperforms in choppy or bear markets.
- PostgREST sampling: Each analysis cell is based on ~29,000–30,000 samples drawn proportionally across year ranges. This is large but not exhaustive of the full ~15.1M rows.
Data Sources
Market data: Tiingo (stocks), CoinGecko (crypto). Database: Supabase (PostgreSQL). Scoring: proprietary Ripeness Score algorithm. Academic papers sourced from SSRN, NBER, journal publishers, and regulatory body websites.
All 31 Studies
Sortable database of every study referenced in this analysis. Click any row for details and paper link.
| Study | Year | Market | Sample | Rate ↓ | Type |
|---|---|---|---|---|---|
Chague 2020 BR 98K 99.87% Peer-Reviewed | |||||
Chague 2020 BR 20K 97% Working Paper | |||||
Barber 2014 TW 360K 95% Peer-Reviewed | |||||
FPFX Technologies 2024 US — 93% Industry Data | |||||
Securities and Exchange Board of India 2024 IN 11.3M 92.8% Regulatory | |||||
Chague 2025 BR 969K 90% Peer-Reviewed | |||||
Autorité des marchés financiers 2014 FR 15K 89% Regulatory | |||||
Barber 2009 TW 3.9M 82% Peer-Reviewed | |||||
Financial Conduct Authority 2016 GB — 82% Regulatory | |||||
Comisión Nacional del Mercado de Valores 2016 ES — 82% Regulatory | |||||
European Securities and Markets Authority 2018 EU 2.0M 81.5% Regulatory | |||||
Barber 2004 TW 3.9M 80% Working Paper | |||||
Jordan 2003 US 334 80% Peer-Reviewed | |||||
Linnainmaa 2005 FI 26K 80% Working Paper | |||||
U.S. Senate Permanent Subcommittee on Investigations 2000 US — 75% Regulatory | |||||
NASAA Day Trading Project Group 1999 US 68 70% Regulatory | |||||
Dalvi 2023 IN — 70% Working Paper | |||||
Beckmeyer 2023 US — 65% Working Paper | |||||
Barber 2020 TW 450K 56% Peer-Reviewed | |||||
Zarattini 2024 US 7K 36% Working Paper | |||||
Barber 2000 US 66K 6.5% Peer-Reviewed | |||||
Coval 2021 US 66K 6% Peer-Reviewed | |||||
Odean 1999 US 10K 3.3% Peer-Reviewed | |||||
Barber 2001 US 35K 2.65% Peer-Reviewed | |||||
U.S. Securities and Exchange Commission 2000 US — — Regulatory | |||||
Gallegos-Erazo 2024 — — — Peer-Reviewed | |||||
Mahani 2007 — — — Peer-Reviewed | |||||
Grinblatt 2009 FI 10K — Peer-Reviewed | |||||
Gao 2015 TW — — Peer-Reviewed | |||||
Kaniel 2008 US — — Peer-Reviewed | |||||
Kelley 2013 US — — Peer-Reviewed | |||||
Click any row to expand details and access the original paper. Verification status indicates whether claims were confirmed against the primary source document.
Frequently Asked Questions
Between 70% and 97% of day traders lose money, depending on the market, time period, and how "day trader" is defined. The most cited figure — 97% — comes from a study of 19,646 Brazilian futures traders who persisted for 300+ days. Studies of US equity day traders find 64–80% lose money.
Disclosures
- This is research content, not investment advice. Past performance does not guarantee future results.
- Backtested results include survivorship bias — only currently tracked assets are scored. Delisted stocks are excluded.
- Historical scores use technical and momentum signals only (no social sentiment data prior to 2023).
- Short-term (1–5 day) signals show no meaningful edge. The system is designed for medium-to-long-term horizons.
- Banana Farmer is a stock and crypto signal ranking platform. Aaron Browne-Moore is the founder.
Last updated: April 2026 | Data sources: SSRN, NBER, SEC, ESMA, SEBI, AMF, FCA, CNMV, Tiingo, Supabase