Then it is still operating.
But he does not claim silencing it, either.
If he did not switch it off then what is with your "perfectly reasonable" story of love-night, privacy and protection against Lumumba plus father?
Probablity is 1/100 or 1/1000?
If the network is operating then delivery is immediate, unless the phone is inaccessible. There were no problems with the network that night so it is only the accessibility of the phone that counts.
Bolint, it's my position that it's hard, even when you're not having a smoke, to remember the minutae of what you did and exactly when during a time period which seems very normal to you at the time.
No matter what Raffaelle remembers about whether his phone was switched off/ on silent / on, it's my position that it doesn't matter.
If it was on, and Raffaelle remembered correctly, it has no implications for guilt, and it's perfectly and reasonably explained because blips in coverage happen all the time. Most of the time, the signal in my bedroom is fine, then one day, I receive a text 5 hrs late. However, if that day my movements happen to be under intense scrutiny, that fact becomes viewed as suspicious, even though it happens all the time...
If it was off, and Raffaelle remembered incorrectly, this also carries no implication of guilt, as it's perfectly reasonable that Raffaelle's recollection was just an error (a stoned error?), and it's also reasonable that he switched his phone off so that he could enjoy a private evening with his girlfriend.
As for your question about the probability of this dual occurrence (RS 'happening' not to receive a text until 7 hrs after it was sent, and that this 'happened' to occur on the night Meredith was murdered) - I think this ignores the problem of multiple endpoints. Whereas the odds of a particular coincidence, specified in advance, in other words a single outcome or endpoint in a specified window of time, may be very low, the odds of any coincidence or a set of coincidences which could include any number of outcomes or endpoints in a vague or not at all defined window of time, none of which are specified in advance, could be pretty high.
In other words, to know how much of a coincidence this is, we'd have to have data which told us exactly how unusual this is generally.