ψ Psi Quant Research

Signal Report · US

DUK HOLD

Composite decision report grounded in 1-XH strategy signals and a completed multi-agent debate (agents-3node-v1), generated at 2026-09-30 10:07.

Final Agent Decision

Action
HOLD
Weight
0%
Confidence
35%
Variant
xiaohan_gogo_v2

Reasons

Bull Case (Analyst)

- Technical signal is a clean `buy` (score 0.60, xiaohan_gogo_v2) with the full 1h/2h/3h/4h bottom stack confirmed (`gogo_detail: 1h_bottom,2h_bottom,3h_bottom,4h_bottom`) — a multi-timeframe exhaustion/reversal setup, not a single-bar blip.
- Price at 114 sits just above the fast EMA (113.8) and below the slow EMA (116), i.e. early in a reclaim rather than extended — the entry is near the low end of the recent range, consistent with the `drawdown_3m` lifter (+0.23) and `downside_vol_60d` (+1.09) reading.
- The dominant positive factor is `xh_gogo_score` at +1.70 z — the same signal family driving the technical action, so the cross-sectional rank (98/252) is being pulled up by the exact factor that generated the trade.
- News flow is mildly supportive: Pluang carries a ~8.26% upside target ($127), the board appointment headline came with earnings "holding up," and Morningstar lists DUK among its best utilities to buy.

Bear Case (Analyst)

- Composite z is only +0.06, rank 98/252 (top 38.9%) — the cross-section does **not** put DUK in the top decile, and it is explicitly **not in the production basket**. The technical buy is not corroborated by basket membership.
- The draggers are quality/trend factors, not noise: `sharpe_60d` (-1.72) and `rsi_14` (-1.46) are the two worst readings, with `close_to_high_60d` (-0.97) confirming the stock is trading well off its highs. Weak risk-adjusted momentum plus weak RSI is a poor backdrop for a durable long.
- The bull case leans almost entirely on one factor (`xh_gogo_score` +1.70); strip it out and the remaining lifters are small (+1.09, +0.23) against larger negative contributions — a narrow, single-factor signal.
- Valuation/relative coverage is mixed-to-negative: Simply Wall St flags DUK as ~12% overvalued on dividend value, and two head-to-head comparisons (GE Vernova, and the Zacks #3 Hold framing) put DUK on the weaker side of the pair.
- No evidence in the bundle of a fundamental catalyst (no guidance raise, no rate-case outcome); the news set is price-target commentary and list inclusions, which is weak support for a 0.60-score technical entry.

Devil's Advocate

COUNTER TO BULL:
- `gogo_detail` bottom stack is a *condition*, not a trigger. Four timeframes "bottoming" simultaneously is equally consistent with a slow bleed that keeps re-printing bottoms — no evidence of a turn, only of oversold persistence.
- "Just above fast EMA, below slow EMA" is not "early in a reclaim." It's the definition of a downtrend with a bounce. Reclaim requires crossing the slow EMA; that hasn't happened. The bull is calling a setup that hasn't triggered.
- `xh_gogo_score` +1.70 is doing all the work, and the bull admits the technical signal *is* the same factor family. That's circular: the factor is being cited as confirmation of itself. Cross-sectional rank 98/252 is not corroboration, it's the same number restated.
- `downside_vol_60d` (+1.09) as a *lifter* is suspicious. High downside vol is a risk characteristic, not an edge. If the model rewards it, the bull is inheriting a factor that may be regime-fragile.
- News support is soft: a price target, a board appointment, and a list inclusion are not catalysts. "Earnings holding up" is the absence of bad news, not the presence of good news. None of this explains why *now*.
- Score 0.60 is not high. The bull never states the threshold for action or the hit rate at this score. "Clean buy" is asserted, not evidenced.

COUNTER TO BEAR:
- "Not in production basket" is a process fact, not a forecast. Baskets lag signal generation by construction; exclusion today says nothing about forward return. The bear is treating a filter as a verdict.
- `sharpe_60d` (-1.72) and `rsi_14` (-1.46) being the worst readings is *exactly* what a mean-reversion setup looks like at the low. The bear is penalizing the stock for being early — which is the thesis. Weak trend + weak RSI is the entry condition, not a disqualifier.
- "Strip out xh_gogo_score and it's narrow" applies to almost every factor-driven signal. The bear needs to show the factor is *unreliable*, not that it's *singular*. No evidence offered on its historical hit rate.
- Valuation flags (Simply Wall St ~12% overvalued, Zacks #3 Hold, GEV comparison) are slow-moving and largely orthogonal to a multi-timeframe technical entry. Citing them against a short-horizon trade is a horizon mismatch.
- "No fundamental catalyst" cuts both ways — the bear also has no fundamental *deterioration*. Absence of catalyst is neutral, not bearish, unless the strategy requires one.
- Composite z +0.06 with rank 98/252 is *slightly positive*, not negative. The bear frames "not top decile" as if it were "bottom decile." It isn't.

Technical Signal (1-XH)

action
buy
score
0.60
variant
xiaohan_gogo_v2
timestamp
2026-09-30T15:30:00-04:00
close
114
ema_low_fast
113.8
ema_high_slow
116
gogo_score
0.6
gogo_detail
1h_bottom,2h_bottom,3h_bottom,4h_bottom

📈 Backtest Track Record xiaohan_gogo_v2 · signals simulated over available history (~3-month 30-min window, accumulating daily)

Trades
2
Win rate
0%
Avg / trade
-3.5%
Total (compounded)
-7%
best / worst
-0.7% / -6.3%
avg holding
50.0 days
open position
yes — marked-to-market -6.3%
Entry Exit Entry Exit Return Exit reason
2026-08-25 open 121.64 114.00 -6.3% open_position_mark_to_market
2026-05-13 2026-07-16 125.54 124.70 -0.7% daily_sell_after_daily_blue_cross

复盘 = each past xiaohan_gogo_v2 buy signal simulated: enter next bar, manage adds / adaptive sells / −25% hard stop. Demonstration only, not investment advice.

News & social (8 items)