My preferred method for calculating the global average temperature anomaly for HADCRUT4.6 is by triangulating the measurements over the surface of a sphere. This gives almost identical results to those of Cowtan & Way but is a far simpler and faster method. Here is my update for July 2018.

Monthly comparison of HADCRUT4.6 classical (green) and 3D triangulation. Cowtan & Way use kriging to extrapolate H4 into polar regions in 2D
Here is the detail for the Northern Hemisphere
Finally here is the Southern Hemisphere (their winter) which shows cooler than normal temperatures over the tip of South America.
Pingback: July’s Hadcrut4.6(3D) = 0.65C – Climate Collections
Your chart seems to make it clear that the large spike in temperature around 2016-2017 was a temporary phenomenon caused by redistribution of heat from the oceans to the atmosphere. All three of your estimates indicate that temperatures have now returned to pre-El Nino levels. The magnitude of this effect looks to be about 0.5 degrees C. I remember Michael Mann claiming in 2016 that only about 0.1 degree C could be attributed to the effects of El Nino. I wonder to what he now attributes the 0.5 degrees C drop in temperature.
I agree we have returned to pre-El Nino monthly temperatures levels. It will be interesting to see at the end of the year how 2018 compares to 2014.