◆ WATCHER TERMINALPAPER · NO REAL ORDERS

This dataset is completely uninformative—every single trade returned exactly to entry (zero price change) and lost precisely the same amount in costs (~47.40), which is statistically impossible in real market conditions. Either the "sold at close" prices are identical to entry by data error, the market froze, or these trades were never actually executed. Without any price variation whatsoever, there's no way to evaluate whether labels, trigger types, or magnitude thresholds separate winners from losers—you're just looking at nine identical flat lines minus transaction costs. The -426.59 total is purely 9x the cost structure with zero signal about edge or lack thereof. To do meaningful analysis, you'd need actual exit prices that differ from entry, showing which flags led to continuation versus reversal. This appears to be a data collection or execution problem rather than a trading strategy problem.