Islandboy - I do remember the topic of undersized sprocket hub thickness from several years ago. At that time I had tried a 17 tooth front just to see if I like the ratio change. For my riding I didn't like it and I went back to the original. But the 17 tooth sprocket I tried was the same hub thickness as the 16 tooth OEM and it mounted tight to the shaft with no freeplay. The OEM design is not intended to have freeplay, and a proper replacement should have the same thickness and have no freeplay.
So I (respectfully) disagree with Mcromo - by design intent - there should be no play of the sprocket on the shaft. The shaft itself has a bit play but the sprocket is intended to be held tight to the shaft.
But I do understand that many people have used aftermarket sprockets that have a thinner hub than OEM and have not had objectionable issues. I do imagine that the freeplay could cause the chain to wear differently. That might not be a big issue either. And I have heard others say that different bike manufacturers have sprockets mounted with freeplay. I don't have any first hand knowledge about that.
Just my two cents...
Historically, I always operated on drive sprockets being end clamped with no float.
I really should correct myself now, some are end clamped with no float, some are not.
Certainly, my old Norton Commando and SOHC CB750s had end clamped drive sprockets.
My recollection is that my OEM 919 drive sprocket had a hair of float.
The 520 Renthal I put on about 10 years ago had more end float, to the point I did some head scratching and fishing around which ended up with me being able to sleep with the end float left alone.
Earlier today I did a super fast rip and found a website about some Honda cruiser types whose OEM drive sprockets have end float.
I didn't note down particulars, but saw enough to satisfy myself that there are some end float OEM drive sprockets in use.
From an engineering perspective, I'll make the following observations.
An end float sprocket/shaft spline class of fit can be enhanced to provide more, and more even, contacting load areas.(sort of like % of thread engagement for male/female threading in the face loaded state)
This is because any end face error will not be attempting to incorrectly load the sprocket's spline faces against those of the shaft.
This could allow a higher power rating and/or service life for the combination of the sprocket/shaft.
A end clamped sprocket will place the entire shaft in strain axially, in other words induce a stress.
This could result in a lower power rating for the shaft.
This could result in being forced to use a lesser class of fit for the sprocket/shaft splining, and a reduced power rating and/or service life.