A naive approach in blueprints would be to generate a set of sample locations at fixed times along the spline, calculate the distance to each racer to each point to find their approximate time on the spline, and then sort by that. That would have a host of issues to work through such as figuring who is closer if both racers are closest to the same sample point on the spline. Theoretically GetInnacurateNearestPosition() should be a better approach than what I described, but conceivably there could still be edge cases that report incorrectly, especially if you had overlapping splines and whatnot.