@just a fan: strakes on fighters are high aoa devices. In a certain sense they prevent separtation - but they porduce lift, the so called vortex lift.
In f1 they are after vortex downforce. The vortex has a very low pressure (related to crntripetale acceleration to force air along a curved helical path). If you get a vortex to run unedr a downfacing surface you get downforce directly, in the same way as you get lift on strakes or delta wing.
Then as mikey says, if you increase the velocity of the vortex along its axis you strech and narrow the vortex tube; there is a theorem by helmoltz that says that says that this will increase the vorticity (conservation of angular momentum, like a ice skater that accelerates her rotation when closes the arms). That's why I think the ebd is so effective, because it accelerates an existing vortex. Also the leadinf edge vortices are accelerated by the accelartion around the leading edge.
So if you manage to accelerate the vortex, you get a narrower vortex and a even lower pressure peak. The influence of low pressure peaks is local.
So from a certain point of view, the more vortices under the floor the batter; the problem is that vortices are very dissipative and introduce a loss of energy in the flow. So too many vortices is bad; it's a compromise I think.
There are some vortices that are let's say "unavoidable" feature of the flow: once they are generated, why not exploit them? But generating a lot of vortices aiming for downforce is not very efficient.
We can see one nice example of "unavoidable" on the rebull last front wing: at the transition between the mandated central wing section and the outer part of the wing there is a vortex, because the two adjacent section are very different in size and generate different loading; thus rbr have chosen to make a small rounde curl on the mainplane to house that vortex, take a gain in downforce from its low pressure before it is released downstream.