COST — Costco Wholesale Corporation

Report date: 2026-05-26 | Sector: Consumer Defensive | Last price: $1,028.24 | Horizon: 30d | Generated: 2026-05-26T07:15:47.367968Z

Forecast summary

Ensemble point -1.81%

80% CI [-7.19%, +3.85%]

95% CI [-9.83%, +6.96%]

Method dispersion 0.0175

Beats RW baseline YES

MC drift (annual) -29.60%

MC sigma (annual) 20.22%

MC paths 1000

Bull target (90th pct) +5.11% -> $1,080.75

Base target (50th) -3.66% -> $990.66

Bear target (10th pct) -11.94% -> $905.45

Macro regime risk_on_growth

10y yield 4.56%

3m yield 3.59%

Yield-curve slope +0.97%

VIX level 16.59

VIX z-score (252d) -0.48

Sector ETF XLP

Sector relative (90d) +4.72%

Forecast plot (interactive)

Realised volatility

Yang-Zhang annualized (60d) 20.22%

Close-to-close annualized (60d) 19.89%

Per-method comparison

Method	Weight	Point	80% lo	80% hi	95% lo	95% hi
linear	25.0%	-3.46%	-7.15%	+0.20%	-8.72%	+2.21%
monte_carlo	25.0%	-3.66%	-11.94%	+5.11%	-16.18%	+9.78%
ar1	25.0%	-0.13%	-1.65%	+1.40%	-2.44%	+2.22%
random_walk	25.0%	+0.00%	-8.01%	+8.70%	-11.98%	+13.61%

Factor contributions (interactive waterfall)

Factor contribution table

Factor	Loading	Ticker value	Contribution
value_score	+0.0495	-2.0980	-0.1038
quality_score	+0.0192	+1.4651	+0.0282
momentum_score	+0.0286	-0.2813	-0.0080
lowvol_score	+0.0173	+0.5054	+0.0087
revisions_score	+0.0236	+1.7750	+0.0419
news_activity_score	+0.0402	-0.0288	-0.0012
(intercept)	-	-	-0.0005

Factor decay over horizon

Factor IC over time

Per-factor IC backtest summary

Factor	Mean IC	ICIR	% positive	Cumulative attribution	n periods
value_score	+0.2563	+2.690	100.0%	+0.5383	252
quality_score	+0.1105	+1.096	86.5%	+0.2401	252
momentum_score	+0.1428	+1.354	92.5%	+0.2460	252
lowvol_score	+0.0878	+0.887	82.1%	+0.1870	252
revisions_score	+0.1375	+1.319	88.9%	+0.2738	252
news_activity_score	+0.3540	+3.832	100.0%	+1.0893	252

Snapshot — fundamentals + technical

Market cap $456.18B

P/E (trailing) 53.5

P/B 27.6

Forward P/E 45.6

PEG 5.22

Dividend yield 57.00%

Beta 0.91

52w high / low $1,096.50 / $844.06

Distance from 52w high -6.23%

Cross-sectional rank 50 / 92 median

Panel source (trained model: 2026-05-26 (0d old), universe_size=92, lookback_days=252)

RSI(14) 50.4

SMA50 / SMA200 $1,007.01 / $954.89

ATR(14) $22.07

Avg volume (20d) 1.94M

Profit margin 2.99%

ROE 29.65%

Revenue growth (YoY) +21.50%

Earnings growth (YoY) +45.50%

Debt/Equity 60.3

Current ratio 1.06

Short ratio 3.69

Forecast accuracy — walk-forward backtest (60d lookback)

Horizon	n predictions	Directional accuracy	MAE (return)	RMSE	80% CI hit rate	Pearson(pred, real)
1d	691	52.2%	0.96%	1.30%	82.8%	-0.068
30d	662	62.8%	6.80%	8.58%	68.9%	-0.062
60d	632	72.9%	9.93%	12.60%	65.0%	+0.069

Volume momentum (Granville / CGW / Quong-Soudack)

Composite z +0.448

Active signals 7 of 7

MFI(14) 61.78

CMF(20) +0.1210

OBV z (252d) -0.217

VPT z (252d) +1.325

VW-momentum z +0.375

Volume z (60d) +0.371

Relative volume (vs 20d ADV) 1.05x

CMF z (252d) +0.426

Sector rotation & AI-spillover (v6)

Sector ETF XLP

Rel-return 5d / 20d / 60d -2.18% / -0.05% / +8.38%

Sector mom-z 1m / 3m / 6m -0.41 / -1.10 / +0.59

Rotation phase cyclical

AI-factor beta (60d) -0.618

AI spillover score -0.0568

Risk-off corr regime (60d) 0.33 (>0.6 = risk-off; <0.35 = stock-picking)

Sector breadth (% > SMA50) -

Sector dispersion 20d -

Peer comparison (sector-relative valuation & momentum)

Peer signal

MIXED

score -0.56 (9 peers)

P/E percentile

100% (100% cheaper than this)

20d momentum percentile

44% (44% lag this)

Market-cap percentile

89%

Peer set (mcap band 0.2x-5x, same sector)

KO PG PM WMT PEP BUD BTI UL MO

9 peers in Consumer Defensive (mcap band 0.2x-5x); signal=mixed; score=-0.56

Decision input: cheap_leader → nudge bias one bucket toward LONG; expensive_laggard → nudge toward SHORT; cheap_laggard = value-trap (neutralize); expensive_leader = crowding risk (neutralize). Folded into Claude critique and one-line PM summary.

Multi-timeframe technical analysis (1wk / 1d / 1h)

Per-timeframe verdict (TradingView-style aggregate of 17 indicators: RSI, MACD, BBands, ADX, Stoch, %R, OBV, MFI, CMF, ATR, SMA50/200, EMA9/21, Donchian, price-vs-SMA200). Counts = how many indicators fired BUY / SELL / NEUTRAL; net strength is the weight-sum delta. Crosses (golden/death, MACD signal cross, Donchian breakout) and the SMA200 trend filter carry 1.5x weight.

1wk verdict

NEUTRAL

4 buy / 1 sell / 12 neutral · net +7.0

1d verdict

NEUTRAL

4 buy / 0 sell / 13 neutral · net +5.5

1h verdict

BUY

6 buy / 2 sell / 9 neutral · net +8.0

1wk NEUTRAL 4/1/12

RSI(14) 58.3

MACD hist +3.6207

Bollinger %b 0.78

ADX(14) 18.2 RISING

ATR % 4.31%

MFI(14) / CMF(20) 58.3 / +0.356

Donchian pos / break 0.71

vs SMA200 ABOVE

n_bars521

1d NEUTRAL 4/0/13

RSI(14) 50.3

MACD hist +2.1425

Bollinger %b 0.51

ADX(14) 15.3 RISING

ATR % 2.15%

MFI(14) / CMF(20) 61.8 / +0.121

Donchian pos / break 0.39

vs SMA200 ABOVE

n_bars1253

1h BUY 6/2/9

RSI(14) 30.8

MACD hist -3.2508

Bollinger %b 0.23

ADX(14) 45.2 RISING

ATR % 0.74%

MFI(14) / CMF(20) 26.9 / -0.234

Donchian pos / break 0.04

vs SMA200 ABOVE

n_bars1729

Cross-timeframe confluence

Confluence LONG (0-1) 0.00

Confluence SHORT (0-1) 0.00

Patterns are intentionally selective (Elder triple-screen, daily/hourly divergence, BB-squeeze+breakout, SMA-alignment, MACD/OBV). Most tickers fire 0-2 patterns on any given day. Below: each pattern with its current state (✓ firing / ✗ not firing).

Long patterns

✗ elder_triple

✗ daily_oversold_hourly_bull

✗ bb_squeeze_break_up

✗ sma_align_long

✗ macd_cross_obv

Short patterns

✗ elder_triple

✗ daily_overbought_hourly_div

✗ bb_squeeze_break_dn

✗ sma_align_short

✗ macd_cross_obv

Intraday nowcast — 1h-bar short-horizon (v6)

Next 24h nowcast

+0.19%

over next 24h on 1h bars

80% band

[-3.39%, +3.90%]

n_bars = 1729 · sigma_1h = 0.0058

Direction vs 30d ensemble

AGREES

1h direction: LONG · 30d direction: SHORT

Source: intraday_forecaster_v10_trained_lgbm.
(trained model: 2026-05-26 (0d old), mean_directional_accuracy=0.523, n_features=17) Decision input only -- the 30d ensemble + factor regression remain the primary call. Conflicts surfaced into the Claude critique as horizon_conflict flag when |1h move| > 1%.

Meta-label gate — AFML Ch.3 secondary classifier (v6)

Verdict

meta-label: NEUTRAL

p_take = 0.45 · threshold 0.60

Bias adjustment

meta-label NEUTRAL (p_take=0.45 between 0.40 and 0.60) -- no adjustment.

Method: secondary LightGBM trained on (features + primary direction) -> P(primary is right). Source: meta_labeler_v10_trained_lgbm.
(trained model: 2026-05-26 (0d old), roc_auc=0.641, accuracy=0.608, n_train_rows=8,284) When ABSTAIN with p_take < 1-threshold AND bias non-neutral, the bias is demoted one bucket and a meta_abstain flag is added to the critique. See AFML Ch.3.6 (Lopez de Prado 2018).

Implied vol vs realized vol — options-trade gate (v8 W2)

Regime

NEUTRAL

no decisive vol mispricing — delta-1 setup over options

IV / RV ratio

1.22

z-score: n/a (insufficient history)

30d IV vs 30d RV

IV 25.9% (Polygon)

RV 21.3% (yang-zhang)

Provenance: source polygon_options:contracts_call; expiry 2026-06-26 (31d) · strike $1,030.00 · 1 strike(s) used.
Caveat: 15-min delayed end-of-day quotes; free-tier Polygon (no Greeks, no real-time). Regime is qualitative input only — NOT a numeric factor on the regression.

FMP fundamentals & analyst consensus (v7)

Quality / leverage

ROE 26.6%

ROIC -

FCF yield 2.0%

Debt / EBITDA -

Current ratio 1.06

Margins / valuation

Gross margin 12.9%

Operating margin 3.8%

Net margin 3.0%

P/E (TTM) 53.4

EV / EBITDA 31.8

Analyst consensus

Target mean $1,083.27

Target high / low $1,275.00 / $769.00

Upside vs last close +5.4%

Revisions score +0.112

Trade-bias signal — foundation for trade-guidance layer

Bias neutral

Composite z-score -0.854

Conviction 0.3414201643317319

Recommended playbook cash

Suggested position size 0% (no Kelly)

no Kelly: bias=neutral / meta_p_take=0.45 <= 0.5; singleton-proxy: news_activity_z=-0.03 (nonzero=True), wf30d_dir_acc=0.6283987915407855 (>0.55=True)

Macro override risk_on_growth_downweights_bearish

Strategies on bias side

•Cash / no position

•Iron condor if IV rank > 50

•Wait for next factor refresh

Why this bias

Sensitivity (OFAT tornado ±1 sd) and stress scenarios

Factor	Loading w_i	Ticker z-value	+1 sd impact	-1 sd impact
value_score	+0.0495	-2.098	+0.0495	-0.0495
news_activity_score	+0.0402	-0.029	+0.0402	-0.0402
momentum_score	+0.0286	-0.281	+0.0286	-0.0286
revisions_score	+0.0236	+1.775	+0.0236	-0.0236
quality_score	+0.0192	+1.465	+0.0192	-0.0192
lowvol_score	+0.0173	+0.505	+0.0173	-0.0173

Stress scenarios (forecast shift from base point)

Scenario	Description	Stressed point	Delta vs base
rates_+100bps	Yield curve +100bps (lowvol -1sd, value -0.5sd)	-7.66%	-4.20%
recession_risk_off	Recession (momentum -1.5sd, lowvol +1sd, news -1sd)	-10.05%	-6.58%
quality_flight	Quality flight (quality +1sd, revisions +1sd, momentum -0.5sd)	-0.60%	+2.86%

Multi-source live sentiment (free real-time)

Composite z-score -0.571

Active signals 3 of 3

Sources stocktwits, apewisdom, iv_skew

StockTwits bull 5

StockTwits bear 0

StockTwits net +1.00

Message volume 30

ApeWisdom mentions 3

24h change +0

Reddit rank 113

25-delta IV skew +0.9589

IV-skew read bearish

Analyst critique

Agreement with model agree_with_caveats

Confidence 0.55

PM one-liner COST Consumer Defensive: quant -1.81%/30d, disp +0.0175, beats RW, macro neutral, 1 catalyst(s).

Sensitivity concern If value_score (the highest-loading factor) is mis-measured by +/-1 stdev, the bias likely flips.

What the model may have missed

Peer set (9 Consumer Defensive names) ranks this ticker in the 100th P/E percentile and 44th momentum percentile (signal: mixed) -- no decisive sector-relative edge.
1 catalyst(s) (?) scheduled -- model integrates calendar dates but not expected-move size from options skew
IV-RV neutral (25.9% IV vs 21.3% RV, ratio 1.22x) -- no decisive vol-mispricing edge; delta-1 setup over options.

Critique flags

{'type': 'factor_conflict', 'severity': 'medium', 'evidence': 'value_score drives 57% of total contribution -- single-factor bet', 'suggested_check': 'stress-test what COST looks like if value_score z-score reverts to 0'}
{'type': 'placebo_audit_failed', 'severity': 'high', 'evidence': 'GT-score 0.53 > 0.5: real-acc 52.2% vs max placebo-acc 55.4% (AMG -0.031)', 'suggested_check': 'setup may be spurious -- inspect feature-importance for noise-like features; consider increased abstention threshold'}

Factor contribution methodology — where the numbers come from

Point return formula (linear factor model): r_hat = alpha + sum_i (w_i × f_i) where w_i is the cross-sectional loading for factor i (estimated from the cross-section of US large-cap names) and f_i is the z-scored factor value for this ticker (Grinold-Kahn winsorization at ±3).

Per-factor contribution shown in the waterfall above is exactly w_i × f_i (decimal-return units), so the bars sum (with intercept) to the linear forecast point. The ensemble point combines linear / Monte Carlo / AR(1) / random-walk per per-method-comparison table.

Bias mapping (composite z-score → LONG/NEUTRAL/SHORT): weighted sum of value (1.0), quality (0.7), momentum (1.0), low-vol (0.5), revisions (1.0), news-activity (0.6); thresholds ±1.0 with macro override (risk-off + LONG → reduce to NEUTRAL; risk-on-growth + SHORT → reduce to NEUTRAL). LONG → Playbook A (long-call / call-debit-spread / covered-call); SHORT → Playbook B (long-put / put-debit-spread / bear-call-spread); NEUTRAL → cash.

Sensitivity (one-factor-at-a-time tornado above): impact = w_i × 1sd per factor (MSCI / Two-Sigma Venn convention). Stress scenarios are canned multi-factor shocks (rates +100bps; recession risk-off; quality flight).

Factor decay uses exponential half-life per factor type: catalysts 7d, news 14d, revisions 30d, momentum 45d, value 90d, quality 90d (Tetlock 2007; Cohen-Malloy-Pomorski 2012). The heatmap above shows contribution × exp(-ln(2) × t / half_life) on a 5-day grid.

Factor IC backtest: Spearman rank correlation between each factor's cross-sectional ranking on day t and the realized h-day-forward return, aggregated across the panel (Grinold-Kahn 2000 ch.4 fundamental law of active management). Source for this report: synthetic_panel_252d_100tickers_seed42 (parallel runner).

Placebo audit (v6, GT-Score): GT-score = 0.53 SPURIOUS — setup likely lacks real signal . Computed from 3 seed(s) across placebo kinds: shuffled, gaussian_iid, garch. Method: real directional accuracy = 52.2%; max placebo accuracy = 55.4%; AMG = -0.031. Source: placebo_audit_v6_ar1_runner. Provenance: arXiv 2604.15531 "Spurious Predictability in Financial ML" (Paper #3). Risk-flag spurious_predictability_audit_failed added to narrative.

Catalysts in window (30d)

Date	Type	Source	Confidence	Description
2026-05-28	earnings	yahoo	high	Earnings announcement (COST)

Analyst narrative

Bull case

COST ensemble forecast is -1.81% with 80% band [-7.19%, +3.85%] over 30 days. Positive drivers: revisions_score, quality_score. 1 catalyst(s) in the window provide event-driven upside potential.

Bear case

Inter-method dispersion (+0.0175) and beats random-walk argue caution. Negative drivers: value_score. Macro regime risk_on_growth historically caps single-name conviction.

Synthesis

Weight the ensemble's -1.81% center against +0.0175 method dispersion. Risk On Growth regime + 1 catalyst(s) define the setup. Treat the 80% band as the working trade-sizing envelope.

Risk flags

spurious_predictability_audit_failed (GT-score=0.53 > 0.5; AMG=-0.031)