Super simple ( and stupid ) floating point error propagation calculating type I dont guarantee that this is correct. If you know you can do something better just contribute! :)
