RECENT POSTS
Tuesday, January 8, 2019
A Look at Climate Change Disclosures

Wednesday, January 2, 2019
Quants: Point-in-Time Data for Backtesting

Friday, December 28, 2018
Now Showing: Controls & Procedures

Thursday, December 27, 2018
A Reminder on Non-GAAP Reporting Rules

Monday, December 17, 2018
Researching PG&E’s Wildfire Risk

Wednesday, December 12, 2018
Tracking Brexit Disclosures

Thursday, December 6, 2018
Campbell Soup: Looking Behind the Label

Sunday, December 2, 2018
SEC Comment Letters: The Amazon Example

Wednesday, November 28, 2018
Measuring Big Pharma’s Chemical Dependency

Monday, November 26, 2018
Analysts, Can You Relate? A True Story

Monday, November 19, 2018
Digging Up Historical Trend Data: Quest Example

Sunday, November 11, 2018
Cost of Revenue, SG&A: Q3 Update

Monday, November 5, 2018
Lease Accounting: FedEx vs. UPS

Saturday, November 3, 2018
New Email Alerting Powers

Wednesday, October 31, 2018
PTC and Two Tales of Revenue

Tuesday, October 30, 2018
10-K/Q Section Text Change Detection

Sunday, October 28, 2018
Finding Purchase Price Allocation

Sunday, October 21, 2018
Charting Netflix Growth in Three Ways

Wednesday, October 17, 2018
Interesting Data on Interest Income

Thursday, October 11, 2018
The Decline of Sears in Three Charts

Archive  |  Search:
Quants: Point-in-Time Data for Backtesting
Wednesday, January 2, 2019

Here at Calcbench we love all analysts, so today we want to give a reminder to all the quantitative analysts out there — we have plenty of time-stamped financial data that can help you backtest your algorithms, and are happy to share.

Our standard dataset consists of this, from the S&P 500. It’s what you see when you open our Multi-Company Page.

We have all that data above, plus other key data and financial metrics, all time-stamped. Those time-stamped data fields include data reported, period start and end, fiscal year and fiscal period, and a few other items.

If you need that data as the raw material to field-test your algorithms, just drop us a line at us@calcbench.com. We’re happy to discuss what you need and how we can get the right data to you.

For the non-quants wondering what all this is, let us explain. Quants create algorithms based on financial data. They test those algorithms using historical data, and need to know the time each piece of data became available to avoid “look-ahead bias.”

That bias happens when your model is running through a year’s worth of data, but you accidentally include some event that happened the 22nd of the month, while your model is still working as if today is the 18th of the month. The algorithm is simulating events, but mistakenly using data it would not have known at some given point in time.

Quants can download this example of our time-stamped data, looking at Microsoft’s data. We have much more, for many more companies, and are happy to work with you to deliver the data you need. Just ask us at us@calcbench.com or visit our website, www.Calcbench.com, and tap on the “Chat with us!” box in the lower-right corner of your browser.


FREE Calcbench Premium
Two Week Trial

Research Financial & Accounting Data Like Never Before. More features and try our Excel add-in. Sign up now to try the Premium Suite.