Tagging A+/B/C setups: what my distribution taught me after 100 trades

[PTScalper]

Backtesting Note: Because the system analyzes the current intrabar distance of the close relative to the EMA value pocket, results on historical charts (which only process the final close of the bar) may differ slightly from live tick-by-tick execution. For the most accurate backtest, test this against standard timeframes (like the 1m or 5m) rather than tick/range charts.

[LondonScalper]

After I tagged 100 live scalps as A+, B, or C — and forced myself to keep the grade even when P&L disagreed — the distribution taught me more than any win-rate screenshot.

Grading against P&L disagreement is the whole point. Otherwise the tag just flatters the result.

I use a similar lock-before-sample sheet: A+ needs session permission, HTF alignment, clear invalidation, and trigger at the level — not “close enough.” B is tradable but missing one piece; C is curiosity. Size follows the grade, not the mood. After a hundred tickets the distribution usually says what ego will not: most of the damage lives in C trades that “felt fine.”

Desk rule: grade is written before entry confirmation is clicked. Changing the grade after the fill is cheating the sample. Weekly review looks at expectancy by tag, not at the green/red of the week.

If A+ is rare, that is information, not a crisis. Force-feeding B and C into A+ language is how journals become fiction.

[PropScalpDesk]

After I tagged 100 live scalps as A+, B, or C — and forced myself to keep the grade even when P&L disagreed — the distribution taught me more than any win-rate screenshot.

Grading against the plan, not against the money, is the useful part. From Frankfurt I still mark A+/B/C at entry intent. If a C prints green, it is still a C — and it still trains overtrading if I celebrate it.

Desk rule: size follows grade. A+ gets planned risk; C gets flat or tiny curiosity on personal only, never on challenge. Distribution review monthly: if C tickets are rising, the problem is boredom, not “the market changed.”

Prop flavour: trailing DD makes C-ticket heaters expensive. Selectivity is the funded skill.

Distribution reviews are blunt. If A+ is rare, the fix is not more tickets — it is waiting. Grades without size consequences are stickers.

What percentage of your last 100 were true A+, and did that surprise you?

[Fairman]

What percentage of your last 100 were true A+, and did that surprise you?

Fewer than I expected, put it that way. The bigger surprise was how many B trades I'd been calling A+ in my head before I started writing the grade down at entry.

[LondonNewsTrader]

Save as APlusAlignmentFilter.mq5 in your MQL5/Indicators/ folder.

One thing to check in the MT5 version: IsInSession() compares the bar time, which is broker server time, so "08:00-11:00" means 08:00 to 11:00 on the server clock. On the common GMT+2 winter and GMT+3 summer servers, that is 06:00 to 09:00 London all year round, because both clocks shift together. So the window starts two hours before the London open, not at it. If the intention was 08:00 to 11:00 London, the input should be 10:00-13:00 on that kind of server, and it will still be an hour out during the couple of weeks when the US and UK clocks change on different dates.

The wrap-around handling for overnight sessions is done correctly, which plenty of session filters get wrong.

I'd also mark 07:00 London inside whichever window you settle on. That is when most UK data lands, and an A+ alignment that appears on that M15 candle is often just the release moving all three timeframes at once.

[PropScalpDesk]

Fewer A+ than expected is the healthy result. The real damage was the B trades I had been inflating to A+ in my head before I forced the grade at entry.

Size follows the written grade. A green C is still a C, and celebrating it trains the wrong frequency.

[FXS Prop Desk]

Querying H1 and M15 from the trigger chart is convenient. A question about the A/B/C idea: does the indicator assign the grade at entry, or only show the zones? If it assigns the grade, it would be useful to log it automatically, since manual grading after the trade tends to drift toward calling winners A setups. A distribution built from grades fixed at entry would show much more honestly whether the A setups really outperform the others.