# Dragon Followthrough Profit Loop Review

## Scope
- objective: inspect whether false-veto glued weak rebounds deserve a fast trend reentry path
- data_source: `fetch_daily_data+compute`
- latest_bar: `2026-04-10`
- row_count: `2737`

## Subtype Readout
- `mid_zone_very_weak_b1` | blocked_full `2` | shadow_confirmed `1` | confirm_like_3bar `50.00%` | next3_sell_cross_rate `0.00%` | avg_next20_max `7.04%`
- `high_zone_weak_b1` | blocked_full `13` | shadow_confirmed `2` | confirm_like_3bar `15.38%` | next3_sell_cross_rate `38.46%` | avg_next20_max `2.97%`
- `ql_rebound_weak_followthrough` | blocked_full `3` | shadow_confirmed `0` | confirm_like_3bar `0.00%` | next3_sell_cross_rate `100.00%` | avg_next20_max `0.04%`

## Probe Readout
- `probe_mid_zone` -> `mid_zone_very_weak_b1` | trades `1` | same_close_avg `-0.24%` | next_open_avg `-0.50%` | post_exit_max_10b `8.35%` | exit_reasons `knife_take_profit_2_glued:1`
- `probe_high_zone` -> `high_zone_weak_b1` | trades `2` | same_close_avg `-3.52%` | next_open_avg `-5.03%` | post_exit_max_10b `-1.52%` | exit_reasons `knife_take_profit_2_glued:1 | good_to_take_profit_2:kdj_sell:1`
- `probe_ql_rebound` -> `ql_rebound_weak_followthrough` | trades `0` | same_close_avg `NA` | next_open_avg `NA` | post_exit_max_10b `NA` | exit_reasons `none`

## Latest Block Case
- signal_date `2026-04-08` | subtype `ql_rebound_weak_followthrough` | regime `range`
- observed bars after block `2/3` | ql_reconfirm_count `0` | sell_cross_count `0`
- max_close_return so far `3.56%` | confirm_like_within_3 `False`

## Judgment
- `mid_zone_very_weak_b1` is the only subtype that still shows a non-zero delayed-reentry path; it stays the only practical followthrough research lane.
- `high_zone_weak_b1` should not be promoted. Enabling pending there only created losing followthrough trades in this replay.
- `ql_rebound_weak_followthrough` remains a hard-block family. The probe branch still produced zero confirmed reentries there.
- execution timing remains a real question only for the mid-zone path: same_close `-0.24%` vs next_open `-0.50%`.
- entry-specific exit treatment is worth a narrow check only if the mid-zone path keeps showing higher post-entry upside than realized return.
- high-zone probe result is already weak enough to stop here: trades `2`, same_close_avg `-3.52%`.
- ql rebound probe stays off the table for promotion: confirmed trades `0`.

## Trade Detail Highlights
- `probe_mid_zone` | origin `2020-11-30` -> buy `2020-12-01` -> sell `2020-12-09` | same_close `-0.24%` | next_open `-0.50%` | sell `knife_take_profit_2_glued` | post_exit_max_10b `8.35%`
- `probe_high_zone` | origin `2016-04-29` -> buy `2016-05-03` -> sell `2016-05-06` | same_close `-4.31%` | next_open `-4.88%` | sell `knife_take_profit_2_glued` | post_exit_max_10b `-2.73%`
- `probe_high_zone` | origin `2025-10-21` -> buy `2025-10-24` -> sell `2025-11-14` | same_close `-2.74%` | next_open `-5.17%` | sell `good_to_take_profit_2:kdj_sell` | post_exit_max_10b `-0.30%`