I am all for taking a scientific approach when it comes to analysis of player performance, however, I am extremely sceptical about what we can tell about a GKs ability from a metric like 'expected saves'.
Firstly, I have to caveat that by saying that IF two GKs are very similar in more traditional metrics (i.e. % shots saved, crosses claimed, successful sweeps, successful passes etc...) THEN I think looking at the expected saves Data could be useful. I am not dismissing it as useless, you would rather have more Data than less Data, as long as you're interpreting it correctly...but like any Data, it can also be very misleading/dangerous if you don't interpret it correctly.
For example, 'expected saves' doesn't take into account any of the 'preventive' measures goalkeepers can take to ensure they don't face shots with a high xG. For example, we all know that De Gea likes to remain rooted to his goal-line. Now, this means he's basically buying himself an extra split-second to make a reflex save and possibly results in a small amount of saves being made that a more 'aggressive' goalkeeper couldn't make, all other things being equal. So in that sense, he might gain 'xS' points. However, it also means he is likely to face shots with a higher xG and a lower xS...because he is doing nothing to prevent strong goal-scoring opportunities (like closing the angle, claiming crosses, sweeping his area etc....)
We have seen that with modern goalkeeping that prevention is far better than 'cure', i.e. closing the angle, sweeping and dominating your penalty area are all more 'plus EV' in terms of preventing goals than just standing on your goal-line and trying to pull-off improbable reflex stops. So how does all of that translate? You mention a goalkeeper like Allisson....how many goals does he prevent with his aggressive, progressive style? How many goals does he prevent by rushing out to smother a shot taken by a forward before they had chance to get set or create a better angle? How is all of this factored into 'expected saves'?
Personally, I think we're going too deep into the 'meta' sometimes with this analysis. A basic eye-test and the very obvious statistics posted by a few people above (save percentage lower than 50% at times when the average is 70%+) all tell me that DDG is not a very good goalkeeper now.