FAQ

A shot list of the most Frequently Asked Questions.

Reach out with any questions you have that relate to Date, System Approaches or Application of Results.
Note – We can sign an NDA for sensitive data uses.

FAQ Data Related Questions:

  • How is the data normalized?  And why does it matter?

Precision Alpha is available as processed location-level data.  Data is presented as a very simple, flat, tab-delimited file with each row corresponding to a closing price date.  No need for JSON, and data is never stored in a relational database.  

  • What files do you provide to subscribers?

Subscribers receive the following file:

  • Daily File, e.g., NASD_07_01_2019.txt 
    • Processing begins 4 hours after market-close each day 
    • Processing and delivery targeted for 5 hours after market close
    • Contains all historical data (six months) used for machine learning
    • File size can vary across exchanges, but current for NASD the size of a daily file is between 30MB-35MB
  • Which exchanges do you currently offer?

Precision Alpha currently offers access to over 80 global exchanges. All tickers are covered for each of the exchanges.  

  • What are some of the key data fields? Do you provide back?
    1. Next day probabilities, market emotion/energy, market power, market resistance, market noise
    2. All data used for processing (6 months) and the key data fields are returned for a whole exchange (assume roughly 3000 equities).
  1. Do you provide Backtesting, and for how long time back in history?

Yes, we do. 3 years of historical data for an entire exchange.

  1. How do subscribers receive the data?

Data is delivered as a text file to S3, or directly to a customer endpoint (SFTP, S3, etc.). 

  1. What is the granularity of the data? How often is the data delivered?

Every security on the exchange, subject to minimum data quality requirements.  Delivered daily, approximately 5 hours after market close.

  1. The upcoming data source category question is designed to communicate with future clients how the data was procured. Observation Data: Systematic data collection (eg: Sensor Data Collection) Manual Submission Data: Data collected through user or panel inputs (eg: Survey Data) Open Source Data: Data originating in the public domain (eg: scientific & governmental)What is the granularity of the data? How often is the data delivered?

Closing prices procured from market data supplier, Xignite.

  1. Source of data?

Licensed from Xignite.

  1. When did you start selling this data?

2018-05-01

  1. What Is the size of each file? How much data Is added with each update?

The size of the data depends on the number of symbols on the exchange.  As an example, each daily file for NASDAQ is approximately 30MB to 35MB, and contains all historical data used for the processing.

  1. In which time zone is each time series measured?

Data observations are provided In UTC time and converted to local time. 

  1. How Is the data collected and sampled? 

Precision Alpha data uses six months of closing prices to return exact machine-learned results. No sampling.

  1. Does your company store or test data using a cloud infrastructure?

Yes, store data in Amazon AWS S3, and process data using AWS Lambda.

  1. How long does your data go back? Do you store historical data?

More than a year of processed data for exchanges.  We store processed exchange data in S3. No customer-specific information is persisted. 

  1. Sharpe ratio?

The Sharpe ratio is defined using statistical expressions, namely, the average return and the standard deviation.  It is clear, however, that financial markets are not in statistical equilibrium, and this measure is completely inappropriate in most financial markets.  The non-equilibrium generalization of the Sharpe ratio can be shown to be always greater than the equilibrium case. Therefore we call the non-equilibrium expression a “Sharper ratio”, and is defined as the ratio of the expected PL minus the risk free return, divided by the expected PL minus the average return.

  1. What asset classes is your data set relevant for?
  2. Equity (let’s chat)
  3. Macro
  4. Credit
  5. Do you map your equities in your data to industry standards like CUSIP, SEDOL or ISIN?  

Yes, market data provider maps SEDOL (Stock Exchange Daily Official List) 

  1. Is your data time-stamped?  

Yes.

  1. Does your data contain PII (Personally identifiable information) or MNPI (Material Nonpublic Information)?

No.

  1. How does Market Power compare with Relative Strength Index? 

Precision Alpha relies on the same measurement data as the Relative Strength Index.  However, Precision Alpha uses the measurements to calculate exact, closed-form algebraically-related expressions for market energy, market power, market resistance and market noise.  The Relative Strength Index is closest dimensionally to market energy.