ψ Psi Quant Research

Signal Report · US

ADI BUY

Composite decision report grounded in 1-XH strategy signals and multi-agent debate, generated by DeepSeek at 2026-08-07 09:26.

🎯 Final Decision

Action
BUY
Weight
35%
Confidence
55%
Variant
xiaohan_gogo_v2

Reasons

Bull Case (Analyst)

- Technical signal is cleanly bullish: the 1-XH family prints a buy with score 0.70, and the gogo_detail shows stacked bottoms across 1h–4h timeframes plus a blue-upper-above-yellow-upper alignment — a momentum-confirming structure at $377.10, above both fast EMA ($374.5) and slow EMA ($376.8).
- The cross-sectional factor view corroborates: ADI's composite z of +0.25 ranks 75/251 (top 29.9%) in a BULL regime catalog with zero direction-flips. The dominant driver is xh_gogo_score at +2.39 — the same family as the technical signal — meaning the momentum/trend factor is the primary lift, not a one-off.
- Secondary factor support: beta_60d (+0.22) and amihud_illiq_60d (+0.11) both contribute positively, suggesting the stock has market sensitivity and limited liquidity drag in this regime — favorable for continuation in a risk-on tape.
- News flow is fundamentally constructive: Q4 results "significantly exceeded consensus estimates" (item 3), the stock surged +4.82% on Aug 4, and a technical breakout setup is flagged (item 6). The CEO's 10b5-1 sale of 10,000 shares at $363 (item 7) is pre-scheduled and small relative to holdings (~137.5k remaining) — not a red flag.
- Price action is confirming: after the Aug 4 surge, ADI advanced another +5.09% to $380.29 on Tuesday (item 5), showing follow-through buying rather than a one-day pop.

Bear Case (Analyst)

- The cross-sectional factor view is mixed-to-negative on quality: the top draggers are close_to_high_60d (-0.89), sharpe_60d (-0.77), and volume_momentum_60d (-0.71). ADI is not near its 60-day high, has weak risk-adjusted returns, and volume momentum is fading — all suggesting the recent rally lacks breadth and durability.
- ADI is NOT in the production basket. The cross-section ranked it 75/251 (top 29.9%), but the model explicitly excluded it from the top decile. The composite z of +0.25 is modest — this is a marginal long, not a high-conviction one, and the technical buy may be a lower-quality signal in this regime.
- Insider selling is a caution flag: SVP Martin Cotter sold 17,594 shares at $418.00 on June 2 (item 1) — a $7.35M disposal near higher prices than current ($377). Combined with the CEO's 10b5-1 sale at $363 (item 7), insiders are monetizing into strength, not adding.
- News sentiment is noisy and lagging: the most recent item (item 4) shows ADI dipping -1.51% while the S&P gained +1.48% — the stock underperformed the broad market in the latest session, and item 5 confirms it "underperforms Tuesday when compared to competitors" despite the daily gain. Relative weakness is a bearish tell.
- The stock is cyclical (item 6) and the factor drags on sharpe_60d (-0.77) suggest the risk-adjusted return profile is deteriorating even as price rises — a classic late-stage momentum setup where the technical buy may be chasing a move that the cross-section is already fading.

Devil's Advocate

COUNTER TO BULL:
- The "stacked bottoms" and EMA alignment are purely descriptive of price history; they do not establish that the next marginal buyer appears above $377. The buy score of 0.70 is barely above the typical 0.5 neutral threshold — a weak conviction signal dressed up as clean.
- The cross-sectional z of +0.25 is statistically indistinguishable from zero in most factor models; ranking 75/251 is not "top decile" and the model's own exclusion from the production basket is the strongest evidence that this signal is not tradeable at size.
- Beta_60d (+0.22) is a double-edged sword: in a risk-on tape it helps, but it also means ADI will give back more than the market on any macro hiccup. Amihud illiquidity (+0.11) is not "support" — it means you pay a spread to get out, which only matters when you're wrong.
- The CEO's 10b5-1 sale is "pre-scheduled" — but the schedule was set by the CEO, and the size (10k shares) is irrelevant; the signal is that the insider chose to sell at $363 when the stock is now $377, implying they saw no near-term upside worth waiting for.
- The +5.09% follow-through on Tuesday is one day. The bear case's item 5 explicitly says ADI underperformed competitors that same day — so the "follow-through" is relative underperformance dressed as absolute gain.

COUNTER TO BEAR:
- The quality drags (close_to_high, sharpe, volume_momentum) are all 60-day lookbacks — they are lagging indicators that will mechanically improve if the current rally persists for another week. Using them to fade a fresh technical buy is like shorting a stock because it was weak last month.
- "Not in the production basket" is a model construction choice, not a market verdict. Top 29.9% with a positive z is still a long candidate in most systematic frameworks; excluding it may reflect risk limits or sector neutrality, not a bearish forecast.
- The June 2 insider sale at $418 is 3 months old and at a price 10% above current — that seller already took profit and is not a marginal seller at $377. The CEO's 10b5-1 is explicitly mechanical. Two small insider sales do not constitute a distribution pattern.
- The -1.51% day vs S&P +1.48% is a single session; item 5 shows the stock still gained +5.09% that week. Cherry-picking one down day to call "relative weakness" while ignoring the prior day's +4.82% surge is inconsistent time-window selection.
- Calling it "late-stage momentum" because sharpe_60d is negative is circular: a stock that just broke out from a base will have poor 60-day Sharpe by definition. The technical signal is designed to catch the transition, not confirm the past — the bear case is using the wrong diagnostic for the setup.

Technical Signal (1-XH)

action
buy
score
0.70
variant
xiaohan_gogo_v2
timestamp
2026-08-06 15:30:00
close
377.1
ema_low_fast
374.5
ema_high_slow
376.8
gogo_score
0.7
gogo_detail
1h_bottom,2h_bottom,3h_bottom,4h_bottom,1h_blue_upper_above_yellow_upper

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

Trades
2
Win rate
50%
Avg / trade
+15.0%
Total (compounded)
+30%
best / worst
+30.8% / -0.8%
avg holding
32.6 days
open position
yes — marked-to-market -0.8%
Entry Exit Entry Exit Return Exit reason
2026-07-21 open 380.36 377.14 -0.8% open_position_mark_to_market
2026-03-31 2026-05-19 316.36 413.93 +30.8% 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)