Read the study before you quote the number

The 97 percent figure comes from a study of Brazilian futures day traders who kept going for 300 days. Among them, 97 percent ended with a net loss.

The paper is called "Day Trading for a Living?" by Fernando Chague, Rodrigo De-Losso and Bruno Giovannetti. It first circulated as a working paper in 2019.

The authors had data on every individual trading the mini-Ibovespa contract, a small futures contract on Brazil's main stock index. It is the contract Brazilian day traders favour most. Full coverage is the study's strength: nobody picked a flattering sample.

The headline sentence gets copied everywhere. The details of who was counted rarely travel with it, and those details change what the number means for you.

The study started with 19,646 people who began day trading that contract between 2013 and 2015. The 97 percent does not describe all of them.

It describes the 1,551 who kept day trading for 300 sessions or more. The paper's question was whether people can day trade for a living. So it looked hardest at the ones who stuck with it.

That choice cuts both ways. It excludes casual traders who quit fast, some of them after a lucky start. It also means the persistent group had every chance to improve with practice. If anyone was going to show a learning curve, it was them.

The sample is individuals only, in one country, on one contract. It says nothing direct about US stocks, options buyers or swing traders holding for days.

Run the numbers the paper gives and the group shrinks fast.

Of 19,646 starters, 1,551 reached 300 sessions. That's 1,551 divided by 19,646, or about 7.9 percent. So the famous figure covers fewer than one starter in twelve.

In that group, 97 percent lost money and 3 percent made money. Three percent of 1,551 is about 47 people. That's out of nearly twenty thousand who started.

The paper also compared profits with real wages. About 1 percent of the persistent group earned more per day than Brazil's minimum wage. Fewer than one in a hundred beat a bank teller's starting pay.

Lose money meant a negative net result over the whole period each person traded. Exchange fees and brokerage fees were subtracted before anyone was called a winner or a loser.

Some costs were left out. Taxes weren't included, and neither were trading courses or platform subscriptions. Adding them could only make the results worse, never better.

The authors also asked whether people got better with practice. They found no evidence that day traders learned over the time they traded. The share of people in profit actually fell the longer people kept going.

Their own conclusion was blunt. In their view, it's close to impossible for individuals to day trade that contract for a living. The competition includes high-frequency firms with faster systems and lower costs.

It can't tell you why each trader lost. The paper measures outcomes, not decisions, so the causes people attach to it are guesses. Overtrading and fees are both plausible causes. Neither was measured.

It can't tell you the figure for other markets. US stocks, options and longer holding periods have different costs and different traders. A different study would give a different number.

It can't tell you about your own account either. A base rate is a warning about the starting odds. Your record is the only evidence about you.

So quote it carefully. "97 percent of all day traders lose money" overstates it. "97 percent of persistent Brazilian index futures day traders lost money" is what the paper found.

Ask your own record the study's two questions. Are you net positive after fees, and are you improving?

For the first, use net P&L after commissions, never gross. Then look at expectancy, the average dollars you make or lose per trade. If it's negative after a hundred trades, your record already matches the study's majority. Check max drawdown too, since a positive month can hide one deep hole.

For the second, compare your first hundred trades with your latest hundred. Put the same stats side by side. If nothing improved, more screen time alone probably won't change it.

If the numbers say stop or shrink, shrink. Cut size, drop the symbols that keep losing, and keep measuring for another hundred trades before you decide again.