Findings · 2026-11-20
Prop-firm challenges with no skill at all: the rulebook decides who wins
We sent a trader with zero edge, picking every direction by coin flip, through two crypto prop-firm rulebooks on six years of real Bitcoin and Ether prices. Under one rulebook, sizing up to the limits came out slightly ahead of the fee. Add a little extra cost per trade and it was gone.
The claim. “You need a winning strategy before a prop-firm challenge is worth paying for.”
What we found. Mostly true. With no skill, one rulebook paid back a little more than its fee, but only at the lowest trading costs; the other lost money at any size.
Follow a signal channel? We score every call it made, after fees. About a minute.
Why test a trader with no skill
A prop-firm challenge is a bet with a capped downside: you pay a fee, and if you pass, you trade the firm's money and keep most of the profit while the firm absorbs the losses. That shape can pay even when the trader has no edge, the same way a lottery ticket can be worth its price if the prize is big enough. The firm sets the fee and the rules so that it is not. We wanted to know who wins when skill is taken out completely.
What we tested
Every trade picked Bitcoin or Ether and long or short by coin flip, so the trader had no edge at all. Each trade risked a fixed share of the account with a 2% stop and a target twice as far, on real 15-minute prices from 2020 to October 2026, paying exchange fees on every trade and closing everything before midnight UTC.
We ran this through two published rulebooks. The two-stage one: make 10%, then 5%, never lose 5% in a day or 10% in total, trade at least four days per stage, no single day may make more than 40% of the profit, and the fee (about 0.6% of the account) is refunded with the first payout. The one-stage one: make 10%, never lose 3% in a day or 6% in total, fee about 1% of the account and not refunded. Funded accounts took 80% of profit each time it reached 5%. We started a new attempt every day, eight coin-flip sequences each.
Three sizing rules were locked before the first run: two-stage at 2% risk per trade (1% in the second stage, to respect the 40% day rule), one-stage at 1%, and two-stage at 0.5% as a control.
The result
Two-stage, sized to the limits: on attempts started from 20 September 2023 onwards, the no-skill trader collected about 0.8% of the account per attempt more than the fee cost. It was positive for attempts started in every year from 2020 to 2025. Four in five attempts still failed; the ones that passed paid for them, helped by the refunded fee.
Two-stage, small size: the same rules at 0.5% risk almost never passed. The trader simply lost the fee, about 0.5% of the account per attempt. The difference between the two is size, not skill: big enough bets give a coin flip a real chance of reaching the target before the loss limit.
One-stage: lost about 0.3% of the account per attempt. The fee is larger, is never refunded, and the 3% daily and 6% total limits stop most runs early.
Why we still call it a fail
Our pass mark requires a result to survive an extra 0.2% cost on every trade, for slippage and wider spreads. At 2% risk each trade already pays about a twentieth of what it risks in fees; the extra cost doubles that, and the two-stage result fell to a loss of about 0.4% of the account per attempt. An effect that disappears with a slightly worse fill is the firm's margin, not an edge.
The rest is outside what a backtest can check: how a firm reads its own consistency rule, any limit on risk per trade, and whether payouts are paid on time.
What this does and does not say
It says the rulebook itself carries a thin margin that can lean slightly toward the trader under one set of rules and not the other, and that trading costs decide which side of zero it lands on. It does not say anyone should buy a challenge. A strategy with a real edge would change these numbers; in our earlier test, eight popular TradingView strategies did not have one.
We are now running the two-stage rules forward on live prices, on paper, with cheaper limit-order entries, to see whether the margin survives real fills.
How it was tested
- Binance USDT-M perpetual futures, BTCUSDT and ETHUSDT 15-minute candles from data.binance.vision, January 2020 to October 2026.
- Entries at the next 15-minute open; stop and target checked on 15-minute highs and lows, with the stop assumed first when both are touched in one bar; flat by 23:45 UTC. Costs: 0.107% taker fees per round trip.
- One attempt per day, eight random sequences each; 120-day limit per challenge stage, funded accounts followed for up to a year. Results judged on attempts started from 20 September 2023 onwards, uncertainty measured across start weeks.
- Rules, sizes and the pass mark were locked on 19 November 2026 before the first run. Pass mark: a t-statistic of 3, the same sign before 2023-09-20, surviving an extra 0.2% cost per round trip and the removal of the five best weeks.
- Rulebooks modelled from the firms' published rules and independent reviews in November 2026; one-stage payout terms are assumed to match the two-stage ones. Fees differ between account sizes and promotions.
Research on past data, not financial advice. Every test we run, failures included, is on the Proving Ground.