Why a backtest looks good

You already saw the ending. That is the trick.

The finished chart

The later bars are on the screen when you choose the entry. The path looks clean because you can see where it went. That is hindsight. It is not a number this page can quote, and it is not a second journal.

What this sitting hides

File → New Training. Set Session mode to Blind (not Standard).

• The symbol name stays hidden until you end.
• The calendar stays hidden until you end.
• Prices stay the real numbers. Only the name and the calendar are hidden.
• The window is a random 90–180 days drawn from history already on this PC. The same screen also says 3–6 months.
• Prop Challenge is off.

What you need first

Each series needs at least 150 days on the period you pick. The wizard says Not enough local history for a blind window when it does not.

The Free EURUSD sample is seven calendar days. It cannot host this sitting. Get more history in Data Center until that period is long enough, and until the period is one your plan includes. If the wizard says the plan does not include that period, pick another period. Do not download a random file from a forum.

The sitting

1. File → New Training. Session mode: Blind. Choose the period. Start blind session.
2. Before Play, write one rule from prices you can see. Invalidation is the high and the low of the first bar on the chart. If a later bar trades beyond either price, this experiment is over. This page adds no pip distance and no percent of an account.
3. Play, then Pause, then Step. Leave Indicators closed. Market Watch is hidden. Strategy Tester, optimize, walk-forward, and Monte Carlo stay closed. Those entries say Not available during a blind session.
4. File → End Training. The dialog title is This session was. It shows the symbol and the UTC window. Choose Show results.
5. Fill Went well, Mistakes, and Next plan about the rule you wrote before Play.

Educational training content only. BarionQuant is not a broker and does not provide investment advice. Past practice results do not predict future outcomes.