thank you both , theres a recommended value that slipped past me when documenting ..
edgemaxlen =agent_radius*8 ...

thanks sivan i saw some other weird path related things i will look into this ..
edit...
yes there is a bug but not precisly what you mentioned sivan thanks i will correct it
i changed the speed of my test so i could observe more path action and i think
perhaps its the trace thingy i thought worked
because this never happend before i tried switching to external tracing wich doesnt trace into recast geometry ...


oh and hight is not an actual concern when it comes to the path on the navmesh i remember someone bringing that up with mikko but thats how it functions an actual concern would be direction .. the problem with developement on trial for me is that i only am able to observe for a short period at a time so a real test would be observing it running for an hour or more ..

by the way this is not all the features of the actual release version i made a copy and removed that code because as i agreed with 3run the actual first release i wil offer up to tust if the team deems it fitting but have to say that they do have their own system their working on and perhaps also i dont know how easy navmeshing would be for begginers but anyway thank you for checking