(...)
Mach accepts "but if I were a prosecutor I would assume that the cross-questioning of a person who does not intend to tell the truth would not bring any further information about what happened" does Mach accept that in this circumstance it is wrong of Nencini to rely on Guede for a motivation. Nencini p334 translation "the fact that Rudy Hermann Guede insisted on repeating this circumstance in every one of his interrogations, together with the certain proof that the sum of 300 euros had in fact been set aside by the victim for the payment of her share of rent, makes the story of the Ivorian objectively credible" both because the testimony had not been given in their trial giving no opportunity to examine Guede, and because Guede should be regarded as unreliable in all he says.
I don't have made up my mind on the point yet, as for what concerns Guede's reliability on the 300 euros cash, frankly. I would need to re-read Nencini and pay attention to the specific point. But I can say preliminary that your two questions have two answers:
1) "because the testimony had not been given in their trial giving no opportunity to examine Guede", the answer to this question is: No. I disagree with you, this is not a valid reason, in my opinion, as for the law. A piece of information or a finding can be used in a trial whenever there are reasons to believe it truthful, also when it comes from somewhere else, like from some other venue or from some other trial. Judicial findings from other trials specifically can be used in cases different from the original ones. No, there is no requirement that the original source of the evidence is "examined" in a cross questioning as if he were a witness.
2) the second reason instead, that is "because Guede should be regarded as unreliable in all he says" is much more serious. Though I am not sure this can be considered universally valid; there might be some circumstances in which there is reason to believe that some information from an unreliable person is relevant, if such information happens to be crossed and corroborated. For example, this happens if it's about a detail which the witness was not supposed to know about.
On this particular point, I am not sure, but I acknowledge that Guede's telling the story of Meredith argueing over missing cash is a bit peculiar on the part of Rudy Guede, since there is no reason why he should know about a specific 300 sum and there is no reason why he should put this unnecessary element into his story. The fact that he puts it costantly into each one of his stories despite their changes and mutual cotnradictions, it's peculiar and suggests that maybe there a reason for this.
However I don't know the reason. It might have to do with Knox's 300 euros allegedly missing from her drawer. This money happens to be a much peculiar element of the story also in Amanda Knox's acount of facts, because she says she forgot to check about them. Which is intrinsically not credible and contradictory with other parts of her story.
The 300 € in Guede's story might also tehoretically be explained by other reasons, for example maybe it was Guede himself who stole the money and this is why he knows about them.
So the element exists, it is true that it has some peculiarity because it is present in all Guede's stories, it is unnecessary and he is not supposed to be aware about it, but it is also ambiguous, in my opinion, if taken by itself in isolation, at least until not crossed with Knox's testimony. Taken by itself it could be pointing in opposite directions. However if you cross it with the oddity of Knox contradicting herself about the money in her drawer, well, it casts a peculiar light; I'm not sure but I think once you assume Knox is guilty (and you also observe her unacounted for cash flow and her frequentation of drug dealers), the missing 300€ can be used as part of the picture to understand something of the motives in play in the murder, to build an explanatory conjecture.
