With the RR RB211 series the real thrust was measured by a load cell on the test bed and with all the usual corrections applied a resistor was selected and installed in the IEPR system on the engine so that indicated EPR equalled a measured amount of thrust. There was an tolerance of plus/minus 1% N1.
The result was picky crews occasionally wrote up that N1's disagreed by x amount with EPR's aligned generally inferring an EPR error which was a pain to troubleshoot. Digging into the test bed figures invariably revealed that the engine with low (or high) N1 passed off the test cell like that. An engine with a fully refurbished fan with nice sharp leading edges on the blades might achieve its thrust on the low side of the allowed N1 band whereas one put through test with only minor cleanup to the fan might run on the high side of the band.