Desk · risk
Accuracy is the word that gets us in trouble
A draft weekly said 九转准确率 63%. I asked 63% of what. Silence. Then a spreadsheet of unmatched 9s colored green because the next day was up. I killed the slide.
The replay rule we actually run: a finished low-9 waits for the next high-9. The later close versus the earlier close is the historical comparison. If the opposite 9 has not printed, the row is open. Open is not a win. Open is not a loss. Open is waiting.
Calling that 准确率 teaches the IC the wrong object. They will hear forecast. They will ask for a bigger number next week. They will get a bigger number, because unmatched rows are easy to torture.
What I allow on the page
- How many 9s finished this week, by venue, daily.
- How many pairs closed this week.
- Of closed pairs, how many later closes were higher, how many lower. History. Sample size in the same cell, always.
- A sentence: this does not forecast the next bar.
If you need a percentage to feel professional, you are already shopping for a story. Use the pair count.
A count, not an order. Questions: support@nineradar.com.