Radio observations of Gamma Ray Bursts afterglows are fundamental in providing insights into their physics and environment, and in constraining the true energetics of these sources. Nonetheless, radio observations of GRB afterglows are presently sparse in the time/frequency domain. Starting from a complete sample of 58 bright Swift long bursts (BAT6), we constructed a homogeneous sub-sample of 38 radio detections/upper limits which preserves all the properties of the parent sample. One half of the bursts have detections between 1 and 5 days after the explosion with typical fluxes F>100 muJy at 8.4 GHz. Through a Population SYnthesis Code coupled with the standard afterglow Hydrodynamical Emission model (PSYCHE) we reproduce the radio flux distribution of the radio sub-sample. Based on these results we study the detectability in the time/frequency domain of the entire long GRB population by present and future radio facilities. We find that the GRBs that typically trigger Swift can be detected at 8.4 GHz by JVLA within few days with modest exposures even at high redshifts. The final SKA can potentially observe the whole GRB population provided that there will be a dedicated GRB gamma-ray detector more sensitive than Swift. For a sizable fraction (50%) of these bursts, SKA will allow us to perform radio-calorimetry, after the trans-relativistic transition (occurring ~100 d), providing an estimate of the true (collimation corrected) energetics of GRBs.