Beta User Tutorial

Exact Sports Facts is a Cognal Sports product, bringing immediate database facts to sports fans.
– you ask for a Baseball Fact in English and it gets the fact from an MLB database:
just speak if you’re on a mobile phone or type into the textbox.
– it’s driven by an ‘ad hoc’ database query generator: Cognal’s SayQL
– since SayQL is ‘ad hoc’, you’re not limited to drilling down “By Player”, “By Team”, “By Year”.
just ask for some Fact, like “American League pitchers with the lowest ERA”.

SayQL converts your Natural Language request into a database query (SQL command)
Based on the terms and phrases you asks, SayQL gets the exact fact from the database.

About the Baseball Database we’re using:
It’s the world’s most popular open source baseball database, authored by Sean Lahman (source: and Dr. Theodore Turocy.
It’s a Historical database, updated yearly (not real-time).
So the baseball facts in this database are all about yearly MLB stats.

Here is the data model of the historical MLB database:

Some popular Requests:
“American League Players with the most home runs, by season”
“Texas Rangers players with more than 40 total home runs”
“2013 allstars with the most home runs”
“National League managers with the most wins”

Important tip: notice that nearly all example requests have some “limiting filter” (otherwise the request would return ALL the rows (e.g., all 18,000+ players that ever played, etc.):
Players with “the most”, “more than”, of a particular batting or pitching stat, by season, OR limited by a year (e.g., 2010), a club (e.g. Texas Rangers), etc.

About the Beta Program:

All of Cognal’s NLQuery-based products are designed to provide the user with the most successful UX (user experience) possible. For a NL database search product to be considerer “successful”, it must provide the requested facts from target databases at an extremely high “success ratio”. So the Exact Sports Facts beta program is a way for a group of volunteer sports fans to use the product and help Cognal get its success ratio to a target level of 80%.
(in comparision, regular search tools can’t get successful queries from databases: their “success ratio” is more like an “unsuccess ratio” – around 20%).