Signal Tearsheet — rank_on_20

Strategy: lagging_indecies  •  Quantiles: 5  •  Horizons: [1, 5, 20]  •  Delay used: 0 bar(s)
IndexStrategy (QuantStats)Attribution (this signal)
Avg coverage: 0.672
Avg q-turnover: 0.239
Mask: long_candidates

What is this?

How to read this page
This is an ex-ante signal diagnostics page (Alphalens-style). It answers: does the signal predict forward returns (IC / quantile return spread), and is it stable enough to trade (coverage/turnover)? For realized impact under portfolio constraints and costs, use the Attribution page.
coverage: Fraction of instruments with a valid signal value each day.
quantile_turnover: Fraction of names that change quantile bucket vs previous day.
rank_ic: Spearman rank correlation between signal and forward returns.
ic_tstat: t-stat of the IC series; bigger magnitude suggests more consistent IC.
ls_fwd_ret: Forward return of top quantile minus bottom quantile (ex-ante, unconstrained).

Navigation

Quick jump
Signal Quality Information Coefficient Quantile Returns Stability & Coverage Config

Signal Quality

High-level diagnostics
Avg Coverage
0.672
Avg Quantile Turnover
0.239
Horizons
[1, 5, 20]
Delay Bars
0
coverage: Fraction of instruments with a valid signal value each day.
quantile_turnover: Fraction of names that change quantile bucket vs previous day.
coverage turnover
xsec std ic decay

Information Coefficient

Rank IC (Spearman) vs forward returns
countmeanstdtstatminmax
1324.00000.02310.72740.5728-1.00001.0000
5327.00000.02410.73020.5969-1.00001.0000
20317.00000.14350.72843.5086-1.00001.0000
ic series 1 ic hist 1
ic series 5 ic hist 5
ic series 20 ic hist 20

Quantile Returns

Forward returns by quantile + L–S spread
qret bar 1 ls cum 1
qret bar 5 ls cum 5
qret bar 20 ls cum 20
Notes: quantiles computed from the signal aligned to execution timing using delay=0.

Stability & Coverage

Data availability & cross-section behavior
coverage 2 xsec std 2 turnover 2

Config

Exact parameters used
signals                    [rank_mom, rank_on_20, rank_day_20]
horizons                                            [1, 5, 20]
quantiles                                                    5
store_values                                             False
store_quantile                                            True
store_rank                                               False
within_mask                                    long_candidates
signal_delay_bars                                            0
max_instruments                                           None
store_only_traded_names                                  False
Generated by quantdsl_backtest • Signal tear sheet focuses on ex-ante predictiveness (IC / forward returns), not portfolio constraints.