Rules

This site does not apply the market breadth filter, so it differs from the backtested rule.
Rule set (same for every stock)
+1 price > MA20; +1 price > MA50; +1 MA20 > MA50; +1 MA20 rising (vs 5 days ago)
+1 RSI14 between 45 and 70; -1 RSI14 > 75
+1 projected volume >= 1.5x 20-day average on an up day; -1 same on a down day
The score (max 6) is context only. It does not decide BUY.
BUY: price above the 20-day resistance (breakout) AND price > MA50 AND MA20 > MA50 AND MA20 rising,
     AND a price bar exists for today, no one-day move >10% in the last 50 sessions, fewer than 4 no-range days in the last 20.
AVOID: score <= 1, OR price < MA50 with MA20 < MA50 (downtrend)
WAIT: everything else
Support / resistance = lowest low / highest high of the previous 20 sessions (today excluded).
Est. entry = last price + 1 tick (0.10): a buy order fills at the ask, not at the last trade.
Stop = est. entry - 3 x ATR14. Risk % = distance from entry to stop. There is no price target.
Exit plan for a BUY: sell if the price trades at or below the stop; otherwise sell after 40 sessions.
Backtest of this exact BUY rule: 72 stocks, 2020 to Nov 2026, 0.5% commission each side, no sale before T+2, 454 trades.
  Average +2.5% per trade, but only 39% of trades won; the typical (median) trade lost 5%, a few large winners carried the average.
  Rule was chosen on 2020 to mid-2024. On the later, unseen period (233 trades) it averaged +0.9% per trade: not statistically reliable.
  2024 alone: about 30 trades, average about -9%. Paying 1% above the assumed entry removes the unseen-period profit entirely.
The stop is not guaranteed: shares bought cannot be sold before settlement, and a sell fills only if there are bids (none at lower circuit).
Intraday volume is projected to a full session using the share of 10:00-14:30 elapsed (an estimate, scaled up 5x at most).
Data: stocknow.com.bd daily history, fetched when a stock is looked up and reused for up to 15 minutes.