How to Lie: An Analysis of Deception Algorithms

I never lie, though I have no way to convince you of that. Either way, I am an engineer and I have been thinking about the mechanics of lying. Specifically, I am interested in the logical thought processes required to avoid being caught.

Someone who lies often must constantly keep track of all their lies. Otherwise, they end up saying the wrong thing at the wrong time. The question is this: what is the best way for them to search through the list of lies and ensure that their next sentence won't be disastrous?

The most basic method would be to search through the lies, one at a time in some random order, until a problematic one is found or until they all check out. That is pretty slow, though.

A faster way is to group the lies by topic. Then, only the lies in topics relevant to the conversation must be checked. If this was a computer, then it would probably be more efficient to group the lies within each group and group those as well and so on. However, people have limitations on their ability to control subconscious thought processes; I would guess that only one or two levels of grouping is efficient for people. While this is a great improvement on the basic method, it still requires that every group topic be compared with that of the intended sentence to decide whether it needs to be searched.

The Tree of Lie

Instead of grouping by topic, they could group by person. Only a few lies are relevant to each person. Therefore, it makes sense to only check the lies that are relevant to the person being spoken to. One benefit this has over grouping by topics is that no subjective, time consuming comparisons need to be made for every group to determine whether it is relevant, as for topics. A problem with this method comes when more than one person is listening. It may take significantly longer to check every one of their packages.

So, combine the two, using person grouping when talking to one or two people and topic grouping when talking to many. Also, for some people, it may be helpful to group all their lies by topic as well.

Occasionally, even with a good deception algorithm, there won't be enough time to finish searching through everything. Therefore, the lies should be sorted from most to least important within each group and the groups sorted likewise. Then, they can always be checked in that order and, if there isn't enough time to finish, at least the important lies have been checked.

All of this is very interesting, but do liars actually employ these methods? I have no clue. These are just the absurd musings of an engineer's mind.