Taiwan Point-in-Time Backtest Data — snapshot_date on every row
Backtesting on restated data silently uses information not knowable on signal day — look-ahead bias, and it inflates results. This bundle stamps snapshot_date on every row, ready to feed straight into a backtest engine.
Why look-ahead bias kills backtests
A company announces EPS 3.2 on Aug 10, restated to 3.8 on Nov 12. A conventional source overwrites in place, so a backtest querying Sep 1 gets 3.8 — but the market saw 3.2. Your strategy silently used future data.
How the bias is blocked in the data layer
Monthly revenue is aligned to its actual disclosure date rather than the reporting month, and financial statements use the statutory filing deadline as a visibility gate. Every row carries snapshot_date.
That makes the guarantee checkable rather than promised: your engine can assert that any value used on signal day was already public on signal day. Look-ahead bias is blocked in the data, not left to researcher discipline.
The same principle, one layer down: an ex-dividend price gap is contamination that belongs in the data layer too. The backward-adjusted price chart shows the adjusted candles against the raw close with the per-day adjustment factor, and quantifies how much a raw-price return understates the holding period. Free on the demo ticker.