This gets tricky for me.
My computer generated running styles are mostly geared towards pace projection and not the preferred running style of the horse when he's sharp.
If a horse showed speed when it was in form but has been getting outrun lately, it might still technically be an E, but I'd expect it to contribute little to nothing to the fractions and pace pressure today.
The same is true of E routers dropping into a sprint or even long sprinters (7F) with some speed dropping back to short sprints (5F or 5 1/2F). I make adjustments for distance.
The trickiest adjustments are turf to dirt and vice versa because not only are you dealing with different average paces between the surfaces, you are also dealing with surface preferences. In those cases I make a small adjustment but in the end wind up just looking at the PPs and taking my best guess about surface preference and what the horse might do.
I also want to eventually add a class adjustment to it.
A horse that has been showing speed against much weaker (even if the fractions seem fast enough to lead today) will often just get outrun or quickly disposed of against better horses at higher class levels that are quicker to accelerate and take control.