My quick and dirty drive-by on Allen is that the velocity drop is a bit of a concern but what jumps out is how both his fastball and curveball have been ineffective.
So I went to look at StatCast's data and found that Allen has had significant issues tunneling his pitches in 2019.
Here are two Cody Allen outings versus the Twins, the first from 2017 and the second from this year. In 2017 his recognition point -- the point where hitters decipher the pitch type -- and the commit point -- where they swing -- have a tight cluster between his two pitches. It's difficult to distinguish the fastball versus the curve. This season however, after tinkering with his mechanics in spring training, it resulted in the two pitches being able to be recognized easier. The knuckle curve now tunnels through a higher spot than the fastball. Many teams now have those types of reports waiting for hitters (i.e. if you see the ball coming from this window, it's a X).
Here it is at a glance:
Allen, despite the velocity in the low-90s now, still has a very high spin rate fastball -- one that would be the highest on the Twins' staff. that can play. Interestingly, perhaps because of his mechanics tinker or maybe because of the Angels direction, Allen did not use that fastball up in the zone as much as he did in his Indians days.
The Twins will attempt to reign all this in. Unlike the Angels, they have a robust analytics team, coaching staff and now have shown they are effective at using both.
I don't know if Allen will be able to get to where he needs to be but if the Twins can tweak him, he could be a very good arm in the bullpen down the stretch. Time will tell.