Historic prices not matching Yahoo historic reports

Here's a review of 'Ahmed's' code, pinpointing the causes for the observed mismatch between QuantConnect's historical data and reported (e.g., Yahoo) prices:

1. Intraday History Request Timing

• When you request self.History(self.Symbol(symbol), 1, Resolution.Daily) at an intraday time like 15:59 or 09:31, QuantConnect will give you the most recent completed daily bar. This means:
  • At 15:59 on 2025-05-01, you'll get the 2025-04-30 daily bar (because the 05-01 bar isn't finished yet).
  • At 09:31 on 2025-05-02, you'll get the 2025-05-01 daily bar (since 05-02 hasn't closed).
• As a result, you're logging/using prices from the prior day instead of the intended target day.

2. Partial Bars Not Official Closes

• Even if the daily bar for the "current" date were accessible before the market closes, its data (especially the 'close' price) would not reflect the official 16:00 closing auction, but just the last trade processed as of your request time.

3. Symbol Mapping / IPO Date Issues

• Stocks like "FRGT" might not have data on the dates you request, especially soon after IPOs; QuantConnect may silently provide prices from a different mapped symbol/ticker or return incomplete/missing data for very recent listings.

4. Stale Dictionary Entries

• If history is empty for a symbol (e.g., stock not trading yet), the corresponding entry is just skipped, retaining any prior value for that ticker and potentially leading to wrong/stale numbers in your logs.

5. DataFrame Indexing

• The use of history.loc[symbol]['close'][0] assumes a specific two-level DataFrame index. In practice, for a single symbol, you may need to use .iloc or .at to robustly access the price. Since structure varies by the number of symbols and bars, this can cause silent errors or missed data.

6. External Data Differences

• Yahoo and QuantConnect may differ due to:
  • Data vendors (different trade feeds and regional conventions)
  • Time zone (QuantConnect is US/Eastern for US equities)
  • Price adjustments (QC gives raw, not split/div adjusted, prices unless specified)
  • IPO data latency or differing definitions of the first trading day

Suggestions to Fix and Align Prices:

• For 'close' prices: Schedule at TimeRules.AfterMarketClose(symbol, 0) (i.e., as soon as the official closing price is available) and fetch the latest completed daily bar.
• For 'open' prices: Schedule just after market open, and fetch the .iloc[-1] row from the prior day's daily bar, not the current, since the current day is incomplete.
• For new tickers: Always check if the history frame is empty and explicitly handle missing data (e.g., log or set None).
• Hold Symbol objects: Store the Symbol objects from AddEquity and use those in history requests to avoid any symbol mapping issues.
• DataFrame access: Use iloc indexing for reliability—e.g., history['close'].iloc[-1].
• Cross-check external closing time conventions (including timezone and auction logic) before comparing prices.

In summary:
Your code samples the “previous day” bars because of when and how the history requests are made. QuantConnect’s prices are accurate, but your scheduling and data-access approach need to account for how completed bars are made available versus the live market session.

Let me know if you'd like corrected code or more specific examples for accessing exactly the prices you’re after!

Sources: Data Preparation - QuantConnect.com, History Responses - QuantConnect.com, US Equities - QuantConnect.com, History Algorithm, Consolidate Hour Bars Into Daily Bars Regression Algorithm