When I bought my Incra miter gauge on the recommendation of a forum member - one of my better recent acquisitions by the way - I read that the way to check the accuracy was to cut the parts of a hexagon and check the fit of the joints - what would be an insignificant error when checking a single part becomes more noticeable when it's an accumulation of errors. As I cut a fair number of octagons, I thought that I would use this idea to check the accuracy of my miter saw which was set and checked using an adjustable bevel and protractor. I had some lengths of 3/4" red oak and used that, with my new stop system, to cut out the parts of an octagon. When clamped, one of the joints had a very minor gap at the outside of the joint - but I can juggle the parts and get the same gap at an inside joint - so I'm assuming that this gap is the accumulated error produced by the combination of the saw setting, any loosenes in the saw bearings/mandrel and operator technique. The adjustments on the saw consist of loosening clamp bolts, moving the parts and retightening the bolts - which normally requires adjusting the parts slightly past the required setting to compensate for movement as the bolts are torqued - so it's kind of trial and error. As the error in each cut, theoretically anyway, is 1/8 of the gap seen, I think that this is as close as I'm going to get and, for what I'm doing anyway, that this is close enough.
In the past, for larger parts, I've glued octagons into half assemblies, checked the straightness across the outer edge (toe-in or toe-out) and trimmed it straight by either clamping the assembly on a panel with the short corners against the edge of the panel and running the assembly through the TS to get a straight edge or doing the same thing by clamping a straight edge to the part and trimming with a router and pilot bit. Any minor differences at the outer points of the octagon after trimming can be blended afterwards - this works for assemblies where the initial miter angle is close and the error is not extreme, basically where you have your fixture about a close as you can get it.
To finish off the story, my wife walked out into the shop after I'd finished the clamp-up, saw the assembly sitting there and decided that it would make a nice picture frame for her - one more item on the "Honey-Do" list.
In the past, for larger parts, I've glued octagons into half assemblies, checked the straightness across the outer edge (toe-in or toe-out) and trimmed it straight by either clamping the assembly on a panel with the short corners against the edge of the panel and running the assembly through the TS to get a straight edge or doing the same thing by clamping a straight edge to the part and trimming with a router and pilot bit. Any minor differences at the outer points of the octagon after trimming can be blended afterwards - this works for assemblies where the initial miter angle is close and the error is not extreme, basically where you have your fixture about a close as you can get it.
To finish off the story, my wife walked out into the shop after I'd finished the clamp-up, saw the assembly sitting there and decided that it would make a nice picture frame for her - one more item on the "Honey-Do" list.