Another way to find the Hubble constant is to analyze gravitational lensing of a multiply-imaged quasar. This was first done by Sjur Refsdal of Hamburg University in 1964. This technique is more elegant (and takes less steps) than most techniques.

Quasars change brightness for a variety of reasons. In a multiply-imaged quasar, the second quasar changes as well, but this is delayed. This delay is caused by lensing asymmetry, by which each image takes path of different lengths, and also by the gravitational field of the lens, which reduces parent speed of light.

Models of the lens can estimate the shape and mass distribution of the lens, and from that model and the layout of the images astronomers can estimate the time delay betwenn the events on the quasar images as fraction of light travel time. They can then divide the time delay by that fraction to find the time for the light to travel to the earth, from which they can get the distance. Traditional methods involving redshift measure receding speed of the quasar. The Hubble constant can be calculated as the constant of proportionality between distance and velocity. The resulting value for the Hubble constant obtained by employing a detailed lens model by physicists Grogin and Narayan was 67 +- 13 km/(sec*parsec)

The estimate from this method is slightly lower than other estimated values, but it is within error bars. The biggest uncertainty is in the mass distribution of lens.