The ‘distance effect' measuring the elasticity of trade flows to distance has been found to be rising since the early 1970s in a host of studies based on the gravity model, leading observers to call it the ‘distance puzzle'. However, this puzzle is regularly challenged by new developments in the specification of the gravity equation or in its estimations. We propose an original survey on the existing methods used to quantify the distance puzzle – basically the computation of an average distance of trade, a meta-analysis on existing gravity papers and the implementation of recent econometric developments, all on a well-specified gravity equation both in cross-section and panel data. We apply all these methods to a unique large database (124 countries from 1970 to 2006). It appears that if all these new developments can change the amplitude of the increase in the trade elasticity to distance, none solve the distance puzzle. We confirm the existence of this puzzle and identify that it only applies to low-income countries which exhibit a significant rising distance effect on their trade of around 18% between 1970 and 2006 while the distance ‘puzzle' for trade within richer countries disappears.