CNTD20 CNTD

Difference of up/down day proportions

Formula

Mean($close>Ref($close, 1), 20)-Mean($close<Ref($close, 1), 20)

The expression above is verbatim from qlib Alpha158/360 feature engineering (qlib/contrib/data/loader.py, Apache-2.0). The platform qlib-compatible expression engine evaluates it on the deterministic research panel.

Cross-section statistics & information coefficient

IC mean (60M)

-0.0191

ICIR

-0.343

Cross-section mean

0.0143

Cross-section std

0.3498

Min

-1.0000

Max

0.9000

IC = Spearman-style rank correlation between the factor cross-section and next-month returns, computed over the last 60 months on the deterministic research panel.

Data export

alpha158.json

Related factors · CNTD

CNTD5

Mean($close>Ref($close, 1), 5)-Mean($close<Ref($close, 1), 5)

CNTD10

Mean($close>Ref($close, 1), 10)-Mean($close<Ref($close, 1), 10)

CNTD30

Mean($close>Ref($close, 1), 30)-Mean($close<Ref($close, 1), 30)

CNTD60

Mean($close>Ref($close, 1), 60)-Mean($close<Ref($close, 1), 60)

Alpha158 — Alpha158 (Microsoft qlib) — 158 features… Factor Library Factor Hub Model Lab