To detect the extremely feeble redshifted HI signals a very substantial investment of LOFAR observing time will be required. The EoR KSP project plan outlines the current strategy. We aim to target between 3 and 5 windows, each about 50-100 square degrees in size and covering the top-part of the HBA tile beam, in the Galactic halo. In each window we wish to observe for about 75-100 nights, with about 4h useful integration time per night. To obtain information on all redshifts where HI signals are to be expected (from z=6.5 at 190 MHz, to z=11.5 at 115 MHz) a minimum of 2 frequency settings will be needed. This corresponds to a total of about 400 subbands of each 195 kHz wide. The 4-bit observing mode will be crucial to collect the large number of beams. The low surface brightness of the signals, about a few milliKelvin, only requires imaging with baselines up to a few km in length. However, in order to isolate and eliminate confusing discrete sources (AGN, starbursts and clusters) much longer baselines will be required as well. Depending on the outcome of extensive end-to-end simulations we expect the longest baselines for EoR experiment to be about 20 km. However, to limit the effects of ionospheric phase corruptions we may well need baselines up to possibly 50 km to the core; we are developing a technique to allow 3-D ionospheric tomography to assist with the non-isoplanatic chromatic effects of the ionospheric phase corruptions.