MarketParquet
Bulk Parquet archives
Parquet files
Stocks & ETFs from Jan 2000
$79 once
↗Optional $15/month
#1 · Our top recommendationBacktesting dataset guide · 2026 edition
Best backtesting data.
Research
starts here.
Give your strategy a better starting point. Compare datasets for local backtesting, with clear choices around history, adjustments and delivery.
Six providers. Clear tradeoffs. A shortlist for individual quants and developers.
Who’s behind this?Published by MarketParquet, which we rank first. This is a company-owned comparison resource, based on published product information. Read our methodology
01 / The shortlist
Find your data fit.
MarketParquet is our #1 pick. Explore the alternatives by the work you need to do, then inspect the product sources.
FirstRate Data
Downloadable CSV history
Zipped CSV
US stock history from Jan 2000
One-time bundles
↗Optional subscription
Kibot
Custom symbols & tick history
CSV-style text + HTTP API
1-minute bars from 1998
Bundles or custom orders
↗Subscription options
Databento
Market microstructure research
APIs + flat files
Dataset-specific start dates
Usage / subscription
↗Historical & live products
Norgate Data
Daily-bar equity research
Windows database + integrations
Market & subscription dependent
Subscription
↗End-of-day updates
EODHD
Prices plus fundamentals
APIs + SDKs
Endpoint & market dependent
Plan-based
↗Depends on endpoint & plan
No providers match those filters.
Editorial order after #1 is for navigation, not an independently measured quality ranking.See criteria & sources ↗
Start with your workflow
The right dataset.
For the question you’re asking.
Our first recommendation is MarketParquet. These profiles help you compare delivery choices and additional capabilities.
Read our MarketParquet recommendation ↗02 / Before you download
A little diligence. Better research.
Practical buyer guides for the decisions a pricing card does not answer.
Research design
A practical data checklist for backtesting
Make coverage, timing, corporate actions and execution assumptions explicit before running a test.
Read the guide ↗Universe construction
Delisted stock data: a buyer’s checklist
Separate delisted coverage, historical membership and terminal-event handling in a backtest.
Read the guide ↗Intraday research
Choosing one-minute and intraday stock data
What to inspect in minute bars, from timestamps and sessions to aggregation and tick-level limits.
Read the guide ↗03 / Good questions
Know what
you’re buying.
Clear answers, with the assumptions visible.
What is the best data for backtesting?
MarketParquet is our top recommendation for individual quants backtesting US stock, ETF and futures strategies with OHLCV bars. Its $79 archive includes minute-to-daily history and delisted securities. Validate the sample, timestamp conventions, adjustment policy and universe before relying on results. Data for a backtest is separate from the backtesting engine itself.
Is this an independent review site?
No. MarketParquet publishes this directory and ranks its own product first. The criteria and product sources are public. We have not run a standardized accuracy or performance benchmark across all providers.
Does the $79 archive require a subscription?
MarketParquet lists the Complete Archive at $79 once. Keep Current is optional at $15/month for future trading-day updates. It does not backfill lapsed subscription periods. Confirm the purchase cutoff, access and license on its current pricing page.
Are competitor prices and data histories directly comparable?
Not always. Providers sell different universes, frequencies, delivery methods and usage rights. We show the buying model instead of inventing a single equivalent price. Follow each provider’s source links to check the exact product.
Does including delisted stocks eliminate every backtesting bias?
No. Historical universe membership, symbol mapping, corporate actions, event timing and terminal values require separate checks. Including delisted price history is useful, but does not by itself establish a complete point-in-time dataset.