I took a look at 538, and it's league and CL predictions are laughable. Not just for us, but others. For example, they have Spurs at a lower percentage to get top four than Chelsea. There is absolutely nothing so far to suggest that is realistic.
The point is, the APLT is so much better
To be fair to 538, their model is predictive, while APLT is descriptive. APLT tells us where we all are at the moment, but if you use it for predictive purposes, then you have to assume that all teams will continue to perform as they have been so far. Predictive models take into account the fact that some, like Spurs, are unlikely to do so.
APLT has Spurs on top, but if 538 had Spurs as favourites for the title, that would be even more laughable.