Denser demo data for deep/daily SMA crosses (SMA 2/4 over AAPL.US 2006 is already populated) #19
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: Brummel/Aura#19
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Found by the cycle-0011 fieldtest (docs/specs/fieldtest-0011-ingest.md).
AAPL.US — the symbol named in the ingestion carrier and the gated integration test — has only ~21 M1 bars for all of August 2006 and ~100 for the whole year. An end-to-end SMA-cross backtest over it runs but yields all-zero metrics (total_pips 0, max_drawdown 0, exposure_sign_flips 0): the slow SMA never warms over so few bars, so the signal never fires. This is correct C2 warm-up behaviour, NOT a bug, and the ingestion boundary itself is faultless.
But the Walking-skeleton milestone's promise (C22: 'a newcomer sees a populated trace') is only technically met — 'a backtest runs over real bars' is true, yet the shipped demo data can't demonstrate a non-degenerate run, and nothing on the public surface signals that 21 bars/month is the expected density (vs a window-bounds bug or normalization error).
Options: document expected per-symbol bar density; OR name a denser default symbol/window for the milestone demo so the showcase run is non-degenerate. No aura-ingest code change implied — this is about which data the milestone showcases. Worth weighing before the Walking-skeleton milestone close.
Walking-skeleton demo data: AAPL.US too sparse (~21 M1 bars/month) for a non-degenerate signal-quality run (C22)to Denser demo data for deep/daily SMA crosses (SMA 2/4 over AAPL.US 2006 is already populated)Closing — load-bearing premise falsified.
The original claim (AAPL.US 2006 too sparse → all-zero degenerate metrics) is empirically false for the shipped SMA(2)/SMA(4) config: the Walking-skeleton milestone fieldtest (
docs/specs/fieldtest-milestone-walking-skeleton.md, finding[spec_gap]) ran over real AAPL.US 2006-08 bars and produced a populated trace (total_pips=-1.5,max_drawdown=65.5,exposure_sign_flips=6). The gating milestone has since closed.The only residual — that deep/daily crosses with longer windows won't warm over ~21 bars/month — is a low-value, idea-labelled demo-data preference with no current demo that needs it. Closing as moot; re-file if a deep-cross showcase is ever wanted.