I'm been drawn in to a decade-old domestic fire with 5 fatalities
(against my better judgement; the other side's attorney put Oscar Pistorius away).

Their FI (also appointed many years after the fact) claims Suspicious fire (police FI at the time said Accidental-, bedding onto an electric bar heater),
based on two main attacks :

1.
The 5 deceased (mother and 4 children, hubby was rescued by neighbours whom cut through the burglar proofing) were found scattered in the livingroom lying supine;
FI (a retired fireman) claims in all his many years of fire fighting,
he would have expected to find mummy trying to protect her offspring.

2.
Mom & dad were allegedly woken by their children screaming;
rushing through to the kid's bedroom they found the bed partially alight;
Mum trying to flee the house with the kids (the house was burglar-proofed all round)
whilst Pop tried to put the fire out with water, until driven back by smoke & heat.
FI claims that a fire (bed/bedding) lit my the bar heater would take 15 minutes to go to flashover,
whilst the witness account (hubby, sole survivor) describes it as mere minutes; i.e. must be an accelerated fire.


Now I, a mere chemist/physicist for one am unwilling to testify as to human behavior in fires,
other than to say that to wake from dead sleep into a now-darkened smoke-filled house
where you inadvertently perhaps get a lungful of acrid hot (perhaps toxic) smoke,
doesn't bode well for a measured, logical response...
Any data/case studies on this would be much appreciated.

As for the fire's development, I have seen some NIST/NFPA-comparative tests
of identical livingroom setups where the unaccelerated fire lags the accelerated fire by perhaps a minute;
anyone have such videos using beds & bedding? (or recent better-definition videos)
A picture paints a thousand words...