I'm guessing the only advantage is ease of manufacture. Pretty easy to slide a bridge across a router table,
Mike

saddles leaning back are self compensating as action is raised.

I'll add why not mill the saddle slot after the bridge is glued on AND angle back a bit?

Greg?

I'll add why not mill the saddle slot after the bridge is glued on AND angle back a bit?

Because I don't enjoy butt clenching moments or procedures all that much when there's a perfectly good alternative?

Cause not everyone has the kajillion dollar Collins saddle mill?

The higher the string off the frets, the more it is stretched sharp when fretted and the further south the break point must be moved to compensate.
I think it was Allan Carruth who figured 9 degrees of back angle would be completely self correcting as action was raised or lowered on a typical steel string. Any amount of rear tilt is at least going in the right direction.


Edit: I just re-read your post and get what your saying

I think the tiny change in length as the string sweeps on the end of a 25.5" radius is insignificant compared to the increased sharpness from stretch as the action height goes up.

and are self compensating as action is raised

Dave Farmer. Implicit in your initial post is the assumption that there's something inherently wrong with milling a saddle slot at 90 degrees. I disagree with that premise. I'm one of those "small" builders who persists in doing just that. 've had no problems at all with that set up. I have access to several guitars that I made from 1980-'82, and there have been no structural or intonation problems due to my saddle slots. In fact, for the past 6 years I've gone to 3/16" and 1/4" saddles and still have had no bridge splitting even with the increased amount of material removed. So, I consider in my case "If it ain't broke, don't fix it."

Also, regarding tops developing a belly. Not even my earliest customers have reported any noticeable deformation of the top, that you seem concerned about. I think I have to go with Clay on this one; "if a guitar develops a significant belly over time it was probably too lightly built." or perhaps it's a true flat-top rather than one that has an arched X-brace.

At any rate, if you feel it necessary to angle the saddle back, by all means do so. At least one notable luthier, Paul Woolson, does. But, I'll stick with what works for me.

When I mentioned to him that the design was thought to be self compensating, he asked whether we had ever reduce saddle height on an older instrument to maintain the player's desired string action heights, and how desirable it might be to that player to see the intonation change over time as the instrument aged and further reductions in height were needed. I was stuck for an answer and opted to retreat to the student bench to ponder the point.

In thinking about this, the movement in intonation points on an angled saddle is only a benefit of the design when changing string action over a limited period of time, such as for initial setup. It seems to me that intonation change is an undesirable attribute where the action adjustments occur over time due to the normal aging of traditionally constructed instruments, no?