Neither FlightStats or the CAA provide a true on-time performance figure. As has been mentioned FlightStats do not have 100% data coverage of all flights. The CAA's data is also flawed as they use a fixed average taxi time (8 mins at MAN) to calculate an assumed departure time.
We all know there is a mass of taxi time variance between say T3 and 23R and the far end of T2 to 05L. The CAA ia aware of this, however their defence is this that this is the only way they can collate data from all UK airfields to a common standard. This is why their published data for any large airport (LHR/LGW/MAN/STN etc) will always be inaccurate.
For the record, true OTP for MAN in Nov was 86.7% and for the calender year of 2013 - 81.4%.
Someone mentioned about MAN-ZRH. Yes this is one of the worst performing city pairs, however it is nothing to do with MAN. Operational runway restrictions at ZRH create significant destination ATFM delay for which there is no avoidance.