The properties of the first 5–12 classic initial breakdown pulses (IBPs) of three cloud‐to‐ground (CG) lightning flashes were determined using a modified transmission line model. As part of the modeling, the current with respect to time of each IBP was determined from the measured electric field changes at multiple sites using three theoretical methods called Hilbert transform, Hertzian dipole, and matrix inversion. In the transmission line modeling the length of each IBP was estimated from high‐speed video data of the IBPs. The modeling provided the following properties of the larger classic IBPs in each flash: peak current, velocity, total charge, charge moment, radiated power, and total energy dissipated for successive IB pulses in three developing lightning flashes. For the main initial leader channel in the three CG flashes (and for one long branch), IBP peak current was largest for the first or second classic IBP and declined mostly monotonically with successive IBPs. For the same channels, IBP current velocity was smallest for the first classic IBP and increased mostly monotonically with successive IBPs. The smallest velocities were (2.0, 2.5, 2.5, 3.0) × 107 m/s, respectively, while the largest velocities were (9.2, 12.2, 11.5, 12.0) × 107 m/s, respectively. These data support earlier hypotheses that it is the classic IB pulses during initial leader of normal negative CG flashes that change the nonconductive air into an ionized path that is sufficiently long and conductive to start the stepped leader.