Cohort snapshot · 2026-09-18

COST

290 historical chart-shape analogs of COST as of 2026-09-18, plus what they did over the next 1, 5, and 10 trading days. No prediction — the empirical distribution.

Consensus

0.19

weak

5d median

+0.69%

hit-rate 55%

5d IQR

[-2.78%, +3.36%]

middle 50%

Conformal 80%

[-5.67%, +7.33%]

p10–p90

10-day forward distribution

Median

-0.30%

IQR

[-4.40%, +5.94%]

Hit-rate

48%

Top 5-day features

rs_in_sector_rank(negative)

importance 1.40

qqq_minus_spy_60d(positive)

importance 1.12

credit_spread_state=tight(positive)

importance 0.90

Historical playbook

300 analogs split into outcome modes by what they actually did over the next 5 days. Not a prediction — a distribution.

Outcome modes

300 analogs · 3 clusters · 5-day return

Uptrend

middle 50%: -0.6% to +7.9%

26 setups

+5.0%

69% up

Smooth sideways

middle 50%: -1.7% to +3.4%

169 setups

+1.1%

60% up

Smooth downtrend

middle 50%: -4.9% to +1.5%

68 setups

-2.2%

34% up

Was this useful?

How to read this

Consensus 0.19 is weak — historical analogs disagree on direction. The IQR is wide. Recommendation: wait for resolution (an earnings print, macro catalyst, or a price move that strengthens the consensus) before sizing into COST.

Live comp set

This snapshot is built daily. For the live retrieval against the freshest bar, distribution chart, and per-subject drivers:

Open COST in Chart Library →