There's alot of misunderstanding of WebBot and how it works and despite people's attempts to explain it, there are those that just refuse to 'get
it' and, yes, those that take the high-probability position so they can make themselves feel smarter. Here are some key points that I made on the
various Oct 7th threads. Again I'll point out that I have made my living doing sophisticated data modeling for a very long time.
First, WebBot is predicated on the fact that humans are individually and collectively precognitive (Cliff uses the term 'psychic'). We can sense
the 'future'. This idea has been borne-out in a large number of published, scientific studies. See
this paper for an overview of some of those studies and/or consult Princeton's
Global Consciousness Project. These studies are carefully designed and rigorously controlled studies whose
data is statistically validated. And for anyone about to parrot-back the pathetically cliched 'statistics can say anything you want it to' retort,
save your effort. You have no clue what you're talking about.
Second, WebBot works by doing pattern recognition modeling around a pre-determined set of words/phrases. It looks for
changes in
liguistics around those target words/phrases (not specific values). We can't speculate on how those words/phrases were selected (since their
methodology is proprietary) but
usually in this kind of data modeling we look for data elements that are generally central to changing data
spaces. There are statistical methodolgies for identifying and selecting these data points (in WebBot's case, words/phrases). The changes in the
linguisitics around the target words are theoretically being brought on by our subconscious precognition. WebBot is using linguistic shifts as a
proxy.
Third, when patterns are 'identified' they are scaled as to their importance statistically against the model space. Any 'value' (such as
date/time) is determined probablistically. Remember the famous 'bell curve'? The 'bell' can be squat and long or tall and tight. But in any
case it is centered on a singe data point. An equal number of data points lie on either side of that 'centroid'. It's a bit more complex than a
simple normal distribution curve but the concept is similar. The pattern (or in this case the 'event') falls within the distribution with the
highest probability and highest scale around the centroid. Oct 7th and Oct 15th are centroids. It's a time window.
The 'event' could be a single event (like 9/11 IF you can call 9/11 a single event. A single day maybe, but hardly a single event) or a cascade
event (like the DotCom Crash). We talk about the DotCom Crash as if it was a single event but we all know that it was not. What we saw on the 7th
was a cascade event of historic proportions. To deny that requires a truly remarkable level of denial.
Also, keep in mind that these predictions were made for the most part over a year ago. As data built-up over time the importance and intensity of
certain time periods became more detaled.
Finally, making sense of the inter-relation of the patterns requires human interpretation. This is clearly a source of error but unavoidable in
modeling exercises of this type. Even in my work there comes a point where the application of a model requires a certain degree of interpretation.
WebBot ins't a crystal ball. Nor a psychic. It's not going to be perfect but it has shown some remarkable abilities to serve as an early warning
system. Over time, it will become even more accurate as patterns can be assessed after the fact to see where things can be improved and to identify
where things were missed. But in general, you'd do well to not discard WebBot's predictions. Don't attampt to make them something they are not
but pay heed. There is most definitely 'something there'.