ψ Psi Quant Research

Signal Report · US

UPS HOLD

Composite decision report grounded in 1-XH strategy signals and a completed multi-agent debate (agents-3node-v1), generated at 2026-09-17 09:55.

Final Agent Decision

Action
HOLD
Weight
0%
Confidence
35%
Variant
xiaohan_gogo_v2

Reasons

Bull Case (Analyst)

- Technical signal is a clean multi-timeframe bottom: `gogo_detail` shows 1h/2h/3h/4h/daily all at bottom with `gogo_score` 0.70 and action=buy at close 100.2 — a stacked washout setup, not a single-timeframe blip.
- Cross-sectional composite z = +0.38 (rank 47/252, top 18.7%) inside a **BULL** catalog with 0 regime direction-flips, so higher z is directly rewarded in this regime.
- The dominant lifter is `xh_gogo_score` (+2.76), i.e. the same signal driving the technical call — the factor model and the technical trigger are corroborating, not contradicting.
- Basket membership confirms: UPS is already in the production basket (rebalance 2026-08-19, 4.0% weight), so the model's live positioning agrees with the fresh buy signal.
- Secondary lifters `downside_vol_60d` (+0.31) and `rs_1m_vs_spy` (+0.14) suggest contained realized downside and modest recent relative strength; news flow includes a $2B international/logistics investment and sell-side target hikes (Citi to $132, another to $130 overweight), plus a +1.42% outperformance day.

Bear Case (Analyst)

- The rank is only 47/252 — good, not elite — and the draggers are quality/extension factors: `close_to_high_60d` (-1.07) is the single largest negative, meaning price is far below its 60-day high, and `sharpe_60d` (-0.54) flags poor risk-adjusted trend.
- `rsi_14` (-0.47) is a negative contributor, so this is not an oversold-bounce reading the factor model rewards; the momentum/quality sleeve is actively working against the name.
- Price sits below the slow EMA (`ema_high_slow` 101.9 vs close 100.2), so the "bottom" call is unconfirmed by trend — a buy signal into a declining structure with no reclaim.
- News tone is mixed-to-weak on the long horizon: shares down 32.4% over five years, brokerage consensus is only "Hold," and there's board/executive churn (Boratto resignation, Gutmann retirement) plus a leadership reshuffle — governance/continuity noise, not a catalyst.
- No evidence in the bundle of an earnings, guidance, or volume catalyst to force a re-rating; absent that, the factor drag from `close_to_high_60d` and `sharpe_60d` can keep UPS a low-conviction, range-bound long.

Devil's Advocate

COUNTER TO BULL:
- "Multi-timeframe bottom" is a lagging oscillator stack, not a catalyst. All five timeframes can bottom simultaneously and still bleed — that's what a downtrend looks like. Stacked washout ≠ reversal; it's the signature of persistent weakness.
- `gogo_score` 0.70 is not extreme. If 0.70 is the buy threshold, you're buying the marginal signal, not the fat tail. Where's the distribution? A 0.70 in a BULL catalog may be barely above the cross-sectional mean.
- Circular corroboration: `xh_gogo_score` (+2.76) is the *same* signal as the technical call. You're counting one piece of evidence twice and calling it "corroboration." That's not confirmation, it's double-counting.
- z = +0.38 is thin. Rank 47/252 is top 18.7% — that's a B-grade name, not a conviction long. In a BULL catalog with zero flips, *everything* looks rewarded; the regime is doing the work, not the stock.
- Basket membership is stale evidence. Rebalance 2026-08-19 predates the fresh signal — you're inferring agreement from a position the model may already want to exit. Membership is not endorsement.
- `downside_vol_60d` (+0.31) and `rs_1m_vs_spy` (+0.14) are rounding-error lifters. Calling them "secondary lifters" inflates noise. And "modest relative strength" over 1m is exactly what you'd expect from a dead-cat bounce.
- News is narrative, not signal. Target hikes and a $2B investment are sell-side and PR — they don't move the factor model, and they're the kind of thing that gets priced in before you read it.

COUNTER TO BEAR:
- `close_to_high_60d` (-1.07) being the largest dragger is *consistent* with a bottom setup. You can't simultaneously call it a washout and complain price is far from the high — that's the definition of a washout. The bear is using the bull's premise as a negative.
- `sharpe_60d` (-0.54) is backward-looking. If the thesis is mean-reversion off a bottom, trailing risk-adjusted trend *should* be bad. Penalizing it is penalizing the setup.
- `rsi_14` (-0.47) as a negative contributor cuts against the bear's own "not oversold" claim. If RSI isn't oversold, the name isn't extended either — so where's the downside edge?
- Below slow EMA is a single trend filter. In a BULL catalog with zero direction-flips, trend-following filters are the *least* informative sleeve. You're importing a bear-market rule into a bull regime.
- Five-year -32.4%, "Hold" consensus, and exec churn are long-horizon and already public. None of it is new information the factor model hasn't priced. Citing it as a bear catalyst is recency bias dressed as fundamental analysis.
- "No catalyst" is not a bear case — it's an absence of a bull case. The bear needs a *reason to fall*, not a reason to be bored. Range-bound is not short.

BOTTOM LINE (both sides): The bull is double-counting one signal and leaning on regime tailwind; the bear is using the bull's own washout premise as a negative and citing stale, public information as edge. Neither side has identified a *new* piece of information that resolves the setup.

Technical Signal (1-XH)

action
buy
score
0.70
variant
xiaohan_gogo_v2
timestamp
2026-09-17T15:30:00-04:00
close
100.2
ema_low_fast
100.2
ema_high_slow
101.9
gogo_score
0.7
gogo_detail
1h_bottom,2h_bottom,3h_bottom,4h_bottom,daily_bottom

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

Trades
4
Win rate
50%
Avg / trade
+2.5%
Total (compounded)
+10%
best / worst
+10.7% / -0.4%
avg holding
16.0 days
open position
yes — marked-to-market -0.4%
Entry Exit Entry Exit Return Exit reason
2026-09-17 open 100.55 100.18 -0.4% open_position_mark_to_market
2026-09-01 2026-09-01 102.85 102.46 -0.4% adaptive_breakdown_1h
2026-08-19 2026-09-01 103.41 103.47 +0.1% adaptive_breakdown_30m
2026-05-19 2026-07-09 100.16 110.83 +10.7% 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)