> decrease of probability that a particular individual will act, with decrease in group size
I believe the study did show that. I heard an interview with one of the authors on the radio (so my memory may be faulty).
So I don't think it really debunked the bystander effect, rather he made a point along the lines of, 'it doesn't really matter, the more people present means it's the more likely that one of them intervenes'.
I believe the study showed that the probability of any individual getting involved increases if there is no-one else present. i.e. they are the only one who can help.
Surveillance cameras might make people feel someone else will take responsibility.
I never really played it when I was a kid, but knew all about it from The Beano and Oor Wullie.