As that article mentions, data quality was abysmal. I once saw a remark about the error bars being larger than the supposed effect - though o cannot find it and

As that article mentions, data quality was abysmal. I once saw a remark about the error bars being larger than the supposed effect - though o cannot find it and think that might be overly harsh. The 1922 test of starlight deflection by the Sun was much, much more accurate. And it did take about 100 years to detect gravitational waves. Though the earliest test was probably the perihelion of Mercury, which I think was covered in the original GR paper.