Roux is clever, but his timeline was an utter crock. I don't blame him for doing it - it's his job, but his timeline is still nonsensical and doesn't even manage to support Pistorius's own story.
This is why I am more interested in a statistical approach as opposed to Masipa's flawed logic.
Roux's timeline requires bending key parts of the evidence - for example changing what time it was on Stipp's clock by 10 mins.
Similarly reconciling Mr Fossils timeline also requires changing Johnson's evidence.
Nel was correct not to produce a timeline. Rather he called for reason.
i.e. the timeline cannot be reconciled exactly and only OP knows what really happened. The state cannot be expected to produce an exact timeline. And anyway it is not critical.
Rather Nel pointed out that overall, the witnesses did a great job of describing what happened - but necessarily there must be mistakes and we can't know exactly what the mistakes are.
Rather than trying to determine the exact marks in the sand - I believe we should examine the entire net.
What are the common, corroborated elements in the story?
Which things are more likely?
Mr Fossils spreadsheet (far more sophisticated than Masipa's approach) allows one to examine the various pieces and consider how they fit in the overall picture.
So for example, once you realise that the 4 nearest witnesses - Dr & Mrs Stipp & Mr & Mrs N - report no shots after 3:15:51, this is necessarily a devastating combination of data points. Furthermore none of these 4 witnesses hears Reeva after 3:15:51.
From this highly reliable centre point, once can work outwards in terms of what is more likely than not.
Nel realises Johnson cannot be exactly reconciled with this. Yet Johnson & his wife hear broadly the same things
Just because Johnson hears 5-6 shots - we don't say - oh well in that case all of Johnson's evidence is wrong - we realise that he has mistakes.
And if Johnson can hear the wrong number of shots, he can also be mistaken about sequence many months later. We can never know exactly.
However given that 2 other witnesses heard a woman screaming, and a woman was shot dead at that exact time - it is really likely that Johnson:
Really heard a man
Really heard one person not two people over many minutes
Really heard cricket bats not guns?
Or is the more likely inference that Johnson has a less vivid, supporting detail incorrect?
Then once can conduct the same analysis multiple times.
Given Mr & Mrs N hear no shots after 3:15:51, is it really likely Mr & Mrs Stipp miss the cricket bats at 3:17?
Given 4 people heard a woman and 2 people, it is likely what they actually heard was OP (a point not in evidence)
Given a woman was shot dead, is it likely she never screamed when 4 people said they heard her?
Given the immediate neighbours heard OP crying for help after 3:15:51 for some minutes - is it likely that they failed to notice him smashing down his toilet door with a cricket bat @ 3:17?
Given Mrs N was awoken by gunshots, it is likely she would then not notice cricket bats sounding like gunshots when awake only 2 minutes later?
Given Mrs Stipp heard a commotion and it said 3.02am on her clock - is it likely that what she actually heard was moments before 3:15:51?
Could Ms Burger, a talented musician, hear the “bangs” in the tempo matching the ballistics, yet actually this was the cricket bat?
it is the overall power of all of these inferences combined that is critical - not each of them individually.