Thinking about the developing table (https://aklwei.wordpress.com/2009/12/11/an-approach-at-looking-at-the-evidence/) and starting to make sense of the data. Additional columns can be added that will start with an assumption, then show all of the things that would result from that assumption. For example; start with the assumption Knox is innocent. Then running down the table at the entries (both current and future) many of the diads/triads only then work out to ‘innocent’ if knox & sollecito’s testimony and statements are unreliable. Then you also start having to conclude other things as well, such as- the vicitm’s DNA on the knife was either a problem of the DNA testing or contamination. You also have to state that the footprints revealed by luminol are either not the duo’s, or they’re made at another time with some other fluid. And the footprint on the pillow then doesn’t match sollectio, guede, or the victim, which means someone else was in the room and potentially arranged the scene. And almost all of the witness for the prosecution are wrong. And so on..
Then set up a second assumption- Knox is truthful (a different thing then ‘reliable’) Overlooking for a moment the conflicting statements, run the same analysis. This approach starts putting some entries in the ‘involved’ column.
Potentially one could look at different scenarios where only one of the two were involved.
As these scenarios are developed, then they can be compared to find out which produces a set of conclustions that are more ”plausible. ” Then theoretically if the ‘innocent’ scenarios contain too many implausible conclusions, it may be possible to show beyond a reasonable doubt involvement.. and if you add in then the testimony at guede’s trial of three people required for the crime.. then involvement and guilt become inseparable. I haven’t arrived at that conclusion yet, or any of them really- still looking at the evidence and testimony.