For those into stats, here are the runners I have start details of running at Moonee Valley over 2040. I haven't seen too many instances where barrier is a huge issue. Some mught say that the middle barriers are best here. But that would have to be weighted up by an assessment of the horses in those barriers. On those details, on face value, the worst barrier ratio wise is barrier 1. Of the last 14 years, I have only 4 horses to have run from barrier 14 in the Cox Plate. They rarely get a capacity field.
Yes! The above is probably like teaching most here to suck eggs. I realise this is not a research and statistics site, however there are some on these racing social media sites (and in senior governance and management roles in racing) who simply do not grasp those two basic concepts.
The classic primary school level example is that if you are playing roulette and the last 20 spins have been black, what is the chance of the next spin being red? I won't answer that. I know Thommo would get it wrong which is why he keeps failing his remedial maths classes. The other critical point that you make above mardi aside from the population statistic cf individual event one, is that SR is useless information as far as informing the prediction of future results is concerned. It is the SR compared to a decent estimate of predicted strike rate that might be useful. So, as you say, a comparison of the actual SR for each barrier with a reasonable assessment of the likely SR based on the chance of each runner without considering barrier draw.