Download README.md from paperswithbacktest/Options-Daily-Price: direct link, hf CLI and curl.
- Browser
- Download file 8.34 kB
-
https://huggingface.co/datasets/paperswithbacktest/Options-Daily-Price/resolve/main/README.md
- Command line
-
hf download hf://datasets/paperswithbacktest/Options-Daily-Price/README.md
-
curl -L -H "Authorization: Bearer $HF_TOKEN" -o README.md https://huggingface.co/datasets/paperswithbacktest/Options-Daily-Price/resolve/main/README.md
pretty_name: Options Daily Price
language:
- en
license: other
task_categories:
- time-series-forecasting
- tabular-regression
tags:
- finance
- quantitative-trading
- backtesting
- algorithmic-trading
- options
- derivatives
- implied-volatility
- greeks
- ohlcv
- market-data
- time-series
size_categories:
- 100M<n<1B
extra_gated_prompt: >-
This dataset is free to browse and gated for download. Approval is tied to a
Papers With Backtest subscription, which also covers the other datasets of the
organisation: https://paperswithbacktest.com/pricing
dataset_info:
features:
- name: symbol
dtype: string
- name: date
dtype: string
- name: underlying
dtype: string
- name: expiration
dtype: string
- name: strike
dtype: float64
- name: option_type
dtype: string
- name: exercise_style
dtype: string
- name: shares_per_contract
dtype: float64
- name: open
dtype: float64
- name: high
dtype: float64
- name: low
dtype: float64
- name: close
dtype: float64
- name: volume
dtype: int64
- name: vwap
dtype: float64
- name: transactions
dtype: int64
- name: deliverables
dtype: string
- name: underlying_close
dtype: float32
- name: rate
dtype: float32
- name: dividends_pv
dtype: float32
- name: implied_volatility
dtype: float32
- name: delta
dtype: float32
- name: gamma
dtype: float32
- name: vega
dtype: float32
- name: theta
dtype: float32
splits:
- name: train
num_examples: 561107408
Options Daily Price
Daily bars of every US-listed option contract that traded, from June 2014 to today: equity options, ETF options and index options (SPX, NDX, RUT, VIX and their weeklies and minis), expired contracts included. One row per contract and session, with the contract's terms, the session's VWAP and number of trades, and an implied volatility and greeks computed from the close.
561,107,408 rows, 24 columns, 36.4 million contracts on 10,883 underlyings, covering 2014-06-02 to 2026-10-30. Refreshed monthly.
Why It Matters
- Option strategies on real contracts: covered calls, put writing, straddles around earnings, collars, delta-hedged option portfolios: every leg is a contract that existed and traded, at the price it traded.
- No survivorship: contracts are kept after they expire, and underlyings after they delist.
- Option-implied signals for stock selection: implied volatility spreads between calls and puts, smirks, at-the-money term structures and option volume relative to stock volume, per name and per day.
Load It
Installation/Upgrade:
pip install --upgrade pwb-toolbox
Load the Dataset:
from pwb_toolbox import datasets as pwb_ds
df = pwb_ds.load_dataset("Options-Daily-Price", symbols=["SPY261120C00650000"])
print(df.set_index("date").iloc[-1])
The symbol is the contract's OCC code: root, expiration as YYMMDD, C or P, strike times 1000 on eight digits. To work on a whole chain, read the monthly shards and filter on underlying.
Columns
| Column Name | Description |
|---|---|
| symbol | OCC option symbol, e.g. SPY261120C00650000. |
| date | Session, YYYY-MM-DD. |
| underlying | Ticker of the underlying stock, ETF or index, as it was named at the time (FB before 2022). |
| expiration | Expiration date, from the OCC symbol. |
| strike | Strike price, from the OCC symbol. Raw, on the basis of the day (not split-adjusted). |
| option_type | call or put. |
| exercise_style | american (equity and ETF options) or european (most index options). Null for a few contracts missing from the contract reference. |
| shares_per_contract | Deliverable shares; 100 for a standard contract. Anything else is an adjusted contract (after a split, merger or special dividend). |
| open, high, low, close | Prices of the session's eligible trades, per share. close is the last trade, which for a thinly traded contract can be hours before the 4pm close of the underlying. |
| volume | Contracts traded. |
| vwap | Volume-weighted average price of the session's trades, rebuilt from the trades. |
| transactions | Number of trades. |
| deliverables | For adjusted contracts, the deliverable (JSON); null otherwise. |
| underlying_close | Raw close of the underlying on the session (for XSP, XND, MRUT, DJX the index level scaled to the option's multiplier). |
| rate | Treasury rate interpolated at the option's maturity, continuously compounded. |
| dividends_pv | Present value, per share, of the dividends the model expects before expiration: the underlying's trailing 12-month yield applied continuously (SPY, QQQ, IWM or DIA for index options). For an American contract that yield cannot price, the projected discrete dividends due before expiration. |
| implied_volatility | Volatility that reprices close, annualised: Black-Scholes-Merton for European contracts, Barone-Adesi-Whaley for American ones. |
| delta, gamma, vega, theta | Black-Scholes-Merton greeks at that volatility. Vega per volatility point, theta per calendar day. |
How Far to Trust the Implied Volatility
It is computed here from the last trade, not read from quotes. Measured against independent end-of-day values computed from quotes, on about 1.5 million equity and ETF contracts a day, 30 October to 5 November 2026:
| Contracts traded ≥ 10 that day | median absolute gap | median gap |
|---|---|---|
| 8 days to 1 year, strike within 3 % of spot | 1.6 to 2.2 vol points | within ±0.1 |
| 8 days to 1 year, strike 3-20 % away | 1.8 to 3.7 vol points | within ±0.35 |
| 1 to 7 days, or strike more than 20 % away | 3 to 12 vol points | up to +6 |
Fed the same closing quote instead of the trade, the solver lands within 0.25 to 0.65 points of those values in median (8 days to 1 year, strikes within 20 %), and its delta within 0.003: the gap above is the price of using trades, not the model. It is noise, not bias. Average across contracts, or require volume, before reading a single row. Over 2014-2026, the 30-day at-the-money SPX volatility built from this table tracks the VIX with a correlation of 0.985 over 3,114 sessions, 2.8 points below it in median, the usual skew gap.
Gap: 21 to 29 October 2026. On these seven sessions the underlying closes were not yet available when the table was built, so underlying_close, implied_volatility and the greeks are missing on 40 % of the rows of the 21st, 81 % of the 22nd and all rows from the 23rd to the 29th (the bars themselves are complete). They will be filled at the next refresh.
The volatility is null on the expiration day, when the underlying close is missing, when the close is below the no-arbitrage bound (deep in-the-money contracts often trade at parity), and for VIX options, which price the VIX future of their expiry rather than the spot index.
What This Data Does Not Cover
Quotes. There is no bid, ask or mid: prices are trades. A strategy that buys at the ask and sells at the bid must model the spread.
Open interest. Not in this table; it is not available historically.
Who traded. Trades are not signed buy or sell, and carry no account type, so open/close or customer/market-maker volume cannot be rebuilt.
Other markets. US-listed options only: no options on futures, no non-US exchanges, no over-the-counter options.
Contracts that did not trade. A contract with no trade on a session has no row that day.
Access
Browsing the card and the schema is open to anyone. Downloading the files needs an approved request, tied to a subscription: what each plan includes. The same approval covers every dataset of the organisation.
Elsewhere
- Dataset page and coverage charts
- The strategy catalogue
pwb-toolbox, the loader used in the snippet above- Every dataset in this organisation