I posted a link to an article the other day about the misuse of statistics. The applicability to this case is very high.
Essentially the police assumption was "this was the kind of thing Cannan did, so it was probably him", and they then looked for stuff that supports this. This is not valid reasoning. The correct issue to examine is "based on what we know, what are the chances this was JC?"
This requires what a statistician would call a "prior", which usually just means what is the range of outcomes you expect. Suppose you're on a golf course and you can see the green, the pin and the hole, but not the players teeing off. A ball is teed off out of your sight, it lands on the green and rolls straight into the hole. A hole in one! So who played the hole in one? Next thing you see is Tiger Woods striding towards the hole. OK, old, but a very skilled player, long career, won many tournaments, expert at getting the ball to go where he wants, and he's holding a driver. It's obviously Tiger Woods, right? QED.
Then behind Tiger appear Jon Rahm, Rory McIlroy and Dustin Johnson. The last three world #1 golfers and they're all holding a driver too. Is it still obvious now that it was Tiger's hole in one? No, obviously it's not. It's still the kind of thing he'd do, but it's also the kind of thing the other three would do, and you have no way to know which of them it was. The chances this was Tiger are at best about 25% and maybe less, because a/ Tiger's a has-been and b/ you don't know who's following just behind Rahm, McIlroy and Johnson. Justin Thomas might turn up.
In the case of SJL the crucial prior never established was, how many people might have done this? That number is not 1. As we've discussed before, it's the total of everyone in the area who had done, or would go on to do, something like this. Cannan's one of them for sure, but one out of how many? Your prior is to establish that number, which might be 50, and on that basis, the chances that it's any of them in particular are 2%.
What you do not do is pick one, decide it fits him and then declare he therefore must have done it. You do not do this because there are 49 other people you have failed to consider, and for all you know (because you didn't look), the same might be said of all of them.
The Cannan narrative proceeds exactly to an unsustainable conclusion via this wrong reasoning. Yes, it could be him, but for all we know it might have been 49 others. Why can't it be one of them? Where's the evidence it could only have been him? Tumbleweed - nobody, until DV, tried cogently to show that it could have been someone else,and we still don't have a baseline for how many people that was.