Re: I'm asking...
zog said:
Oh! Please do!
Zog.
P.S. I'm serious, I'd like to see how it's done!
ok, here's the essentials of it. [gimme a couple days and I'll submit a formal paper after having checked a few things]
trust, can be calculated as the likelyhood or probability that given a known set of stimulus/circumstance, an entity will react in a predictable manner. one calculates this by statistical analysis over time.
therefore, as an example.
when I submit an invoice to a particular client, because they have consistantly in the past payed me within a given amount of time. I can predict with a high level of certainty that they shall pay the next invoice as before.
ok, opaque example.
[originally i had this conversation over a period of three days with someone who was writing his PhD thesis in e-commerce in the early 90s.]
we analyse behaviour and probability against a known reference point. credit companies do something similar.
if we were to then implement a way to assess trust, from one entity to another entity. there would have to be a history that is documented and the results, positive/negative of each interaction analysed in terms of predictability/probability.
best thought in terms of a 4x4x4x4 matrix where x=previous-result y=consistancy z=probability t=time.
the trust level moves forward in time, and either up or down as a calculated probability level. [scale 0-99] Y is expressed as green/yellow/orange/red
its a little more complex to apply to humans. [how does one calculate or quantify "gut feelings" ??]
disclaimer: I've from time to time made trust assessments that turned out to be utterly and completely dead wrong. [everyone does that]
I'll post URLs to relevant papers if asked.