Everett pretty well summed it up. But if you look carefully you can usually find a spike or some other artifact. I was once following a creek where I thought there might have been a grade, but could find no sign of one. There was a steep narrow place where I thought no railroad would go, but about a mile further on I found a piece of cross tie with a spike in it.
I have seen photos of 14% grades on logging spurs, and heard of one that was 18%. On such steep grades a train would often consist of only two loads. As the timber receeded and lines became longer such steep grades became uneconomical.
Another thing to consider is the cable tramway or "incline" in logging terminology. These were as steep as 76% though 30 to 50% were more common. Generally there were adhesion grades both above and below the incline, and the only difference was the steepness of the grade.
Early truck roads often used old RR grades, but even when built from scratch they used the same design methods. One way to tell the difference is in bridges. Truck bridges and trestles were usually flimsier than those built for railroads.
As far as curves go, a small Climax or Heisler could handle a 50' radius curve. A Shay, with its side drive line, would require slightly more. On the Wallowa-Whitman N.F. we used a minimum of 55' radius for articulated log trucks and 75' for semi's and truck & trailer lashups.
What this all boils down to is that the only sure way to tell, on the ground, if it is a road or RR grade is to find an artifact. If the grade has not been disturbed though, you can sometimes see depressions where the ties were. And, even a spike is not always proof positive. When I worked in the logging industry in Western Washington we used RR spikes for anchoring guy lines, tail holds, deadmen, and anything else where they could be used.
