We apply a template-free cross-correlation method to publicly availablestrain data from the LIGO Hanford (H1) and Livingston (L1) detectors forseven strong gravitational wave (GW) events spanning observing runs O1 through~O4.The method, based on the Pearson cross-correlation coefficient between thetwo detector streams as a function of inter-detector time lag, requires noassumptions about waveform morphology.A time-slide background estimated from the same whitened, band-passed datasegment provides the statistical reference.Among all events studied, GW150914 stands out clearly with an empiricalsignificance of $sigma_mathrm{emp} = 9.1$ and $p = 0.001$($n_mathrm{bg} = 10{,}000$); it is furthermore the only event thatremains significant ($>5sigma$) across analysis windows of $pm 0.5$,$pm 1.0$, and $pm 2.0,mathrm{s}$.GW170814 is also detected ($sigma_mathrm{emp} = 4.2$, $p = 0.0005$)but its significance is window-dependent.The recovered inter-detector time lags are consistent with the officialLIGO values for all events.Events with long in-band chirp durations (GW170817: $approx!180,mathrm{s}$)are not detectable by this method, which explains the absence of significancein several high matched-filter SNR events.We discuss the implications for Virgo's role in GW170817 and argue thattemplate-free non-detections cannot contribute to sky localisation viatriangulation.