ψ Psi Quant Research

Signal Report · US

SO HOLD

Composite decision report grounded in 1-XH strategy signals and a completed multi-agent debate (agents-3node-v1), generated at 2026-10-02 10:08.

Final Agent Decision

Action
HOLD
Weight
0%
Confidence
35%
Variant
xiaohan_gogo_v2

Reasons

Bull Case (Analyst)

- Technical signal is a clean buy: xiaohan_gogo_v2 scores 0.60 with `gogo_detail: 1h_bottom,2h_bottom,3h_bottom,4h_bottom` — a synchronized multi-timeframe bottom, with close (83.73) sitting just above ema_low_fast (83.26) and below ema_high_slow (84.46), i.e., early in a reclaim rather than extended.
- Cross-sectional composite z = +0.16, rank 75/252 (top 29.8%) in a catalog explicitly flagged **BULL** with 0 regime direction-flips — the factor regime is not fighting the long side.
- The dominant lifter is xh_gogo_score (+1.76), which directly corroborates the technical buy signal — the same signal family that fired on the chart is the largest positive contributor to the cross-sectional rank. Volatility lifters (downside_vol_60d +1.16, volatility_60d +0.99) also score positively in this regime.
- Fundamentals/news lean supportive: Q2 EPS of $1.13 beat and rose from $0.92 y/y; 2026 consensus EPS revised up 6.74% to $4.59 and 2027 up 7.12% to $4.91; a new power agreement with a "leading Big Tech" name ties SO to data-center demand with contracted load; ~3.8% dividend yield with a long raise history.

Bear Case (Analyst)

- The cross-section did not put SO in the production basket and it is not in the top decile — rank 75/252 is "above average," not a conviction long; the composite z of +0.16 is marginal.
- The heaviest dragger is sharpe_60d at -2.08 — by far the largest single factor contribution — meaning risk-adjusted trailing performance is poor; close_to_high_60d (-0.83) confirms the stock is not near its 60-day high, and beta_60d (-0.79) shows it isn't providing the beta the regime is rewarding.
- The bull case rests almost entirely on one factor (xh_gogo_score +1.76) plus volatility factors; strip out the gogo score and the remaining factor profile is net negative. That is thin, single-signal support.
- News flow is mixed-to-soft on price action: one item notes SO is "down 15% from all-time highs," and an officer filed a Form 144 proposing a (small, 100-share) sale — neither is material alone, but neither adds fresh upside catalyst.
- No evidence in the bundle of a fundamental deterioration or negative earnings surprise — the bear case here is a positioning/quality-of-signal argument (weak Sharpe, no basket membership, single-factor dependence), not a thesis that the business is breaking.

Devil's Advocate

COUNTER TO BULL:
- "Clean buy" is overstated: close 83.73 sits *below* ema_high_slow (84.46), so the reclaim is unconfirmed. A synchronized multi-TF bottom is a mean-reversion setup, not a trend signal — it fires often and fails often. No volume, no follow-through bar, no invalidation level given.
- xh_gogo_score (+1.76) is doing almost all the work. The bull case admits the same signal family "corroborates" itself — that's circular, not independent confirmation. One factor wearing two hats.
- Composite z = +0.16 is noise-level. Rank 75/252 is top 30%, not top decile. "Catalog flagged BULL with 0 flips" describes the *regime*, not this name — regime tailwind ≠ stock edge.
- Volatility lifters (downside_vol_60d, volatility_60d) scoring positive is ambiguous: in many regimes high vol is rewarded, but it also means wider stops and worse realized risk. Framing it as bullish support is convenient.
- Fundamentals are stale/soft catalysts: an EPS beat already printed, consensus revisions already in the price, a "leading Big Tech" power deal with no disclosed MW, counterparty, or margin. Dividend yield is a floor argument, not an upside one.
- Hidden assumption: that the factor regime persists long enough for a marginal, single-factor, mean-reversion signal to pay. No holding-period or decay analysis offered.

COUNTER TO BEAR:
- Rank 75/252 and z=+0.16 being "marginal" cuts both ways — it's not a short either. Absence from the production basket is a sizing statement, not a directional one.
- sharpe_60d -2.08 is a *trailing* statistic; it's mechanically depressed by the same drawdown that produced the 1h–4h bottoms. Using it as a forward dragger is backward-looking and may be exactly the contrarian setup.
- close_to_high_60d (-0.83) and beta_60d (-0.79) are descriptive, not predictive. "Not near 60d high" is the *precondition* for a bottom-reclaim trade, not evidence against it.
- "Strip out xh_gogo_score and it's net negative" is an arbitrary decomposition. You could strip any single factor from any name and get a different sign — that's not a falsification test.
- Form 144 for 100 shares is noise dressed as a signal. "Down 15% from ATH" is a level, not a catalyst — no comparison to sector, no valuation anchor.
- Bear case concedes no fundamental deterioration and no negative surprise. That means the bear is purely a *quality-of-signal* argument — which is the same weakness the bull has, just inverted. Neither side has an edge claim that survives contact with a P&L.

Technical Signal (1-XH)

action
buy
score
0.60
variant
xiaohan_gogo_v2
timestamp
2026-10-02T15:30:00-04:00
close
83.73
ema_low_fast
83.26
ema_high_slow
84.46
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
-3.1%
Total (compounded)
-6%
best / worst
-0.3% / -5.9%
avg holding
24.1 days
open position
yes — marked-to-market -5.9%
Entry Exit Entry Exit Return Exit reason
2026-08-28 open 88.95 83.73 -5.9% open_position_mark_to_market
2026-05-15 2026-05-28 93.24 92.97 -0.3% adaptive_breakdown_30m

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)