Finding leakers

Event: Summer Sale 2016 discount Approvals
---------------------------------------------
Please note that this information is confidential.
Do not post event details on the Steam Community or on your own site(s) until the sale has officially begun.

 - The Summer Sale 2016 discount runs from 23 June 2016 @ 9:45am Pacific to 4 July 2016 @ 10:00am Pacific

______________________________________________________________

A way of narrowing down the potential leakers drastically is by issuing each person who receives this a version that varies slightly based on wording, punctuation, and spacing. This can also include format, but I'll not cover that here. Note thtat which random variation they receive can be seeded, so the messages will always be consistent *to them*.


Let's look at the title: This can be split into three parts; the "Event" declaration, the event's name, and the specific thing being addressed. The way this is formatted in the shown screenshot is as:

{event}: {eventname} {specific}

The three components could be ordered in the form of:

{event} {eventname} {specific}
{event} {specific} {eventname}
{eventname} {specific} {event}
{eventname} {event} {specific}
{specific} {eventname} {event}
{specific} {event} {eventname}


As a simplification, purely to make a point, let's paint all of these with a broad brush.There may be some variation (such as the need for the word "for"), but each of the cases should generally have the same variables:
 - The separator between {event} and {eventname} ( :, -, >, ->, {space} )
 - The separator between {eventname} and {specific}, as either having the same separator or not would seem just as natural as eachother. ( :, -, >, ->, {space} )
 - How {event} is enclosed ( (), {}, [], nothing )
 - How {eventname} is enclosed ( (), {}, [], nothing )
 - How {specific} is enclosed ( (), {}, [], nothing )
 - The wording of {event} ( "Event", "Sale", "Notice")
 - The wording of {eventname}, can be:
 - - Ordered by {name} {year} or {year} {name}
 - - May have a separator ( ",", "-" )
 - The wording of {specific}, which:
 - - May be "discount approvals", "approved discounts"
 - - May or may not capitalise "discount" or "approvals"

So, by examining the title *alone*, we can establish a number of areas of variation that could be used to identify leakers. This is a simplification, but we can deduce that, from thje title alone in this instance, there are 6*5*5*4*4*4*3*2*2*2*2*2 = 921,600 possible variations. Obviously, they may use more separators, synonyms, or rewordings, and they may have a blacklist of formats that look odd and obviously trying to do this, and so wouldn't use.

Use this same method throughout the entire message, and you could theoretically use the variation to pinpoint the exact source of the leak, even among millions of possible candidates.