If you use the Garmin co-ords for the summit, then yes. I usually am not at the true summit when at the Garmin-summit.
On the other hand, OSM summits are pretty darn good. I will generally be on the true summit, or at least very close to, when I'm on the mapped summit point.
Did you see if the same error occurrs with the new map? I'll attach an averaged waypoint at the end, if that helps.
The 50-500m margin of error seems a bit large to me.
I've walked along the same track separate days and they are all roughly on top of each other.
This is from three days, five tracks, and over a week separation.

I have a variance of around 10m, which fits in with the link you sent me:
http://www.gpsinformation.org/dale/dgps.htm10m variation in my track would equal a margin of error of 5m, of course, it's a small sample size so the margin of error could be twice that. Even doubled, it's still an order of magnitude less than 100m.
Of course, in canyons and forests quality decreases, but clear mountain summits and ridgelines should not have that problem.
I don't think that guy included GLONASS in his calcs, which should theoretically improve accuracy even more.
I have also tried waypoint averaging. This example was for ten minutes while I had a quick lunch break.
I was in the true saddle, the lowest point between the two knolls on the ridgeline.

300m error this time.
This is the waypoint
Men wanted for hazardous journey. Low wages, bitter cold, long hours of complete darkness. Safe return doubtful.