Absolute tip–tilt recovery using a tilt signal measured on a Laser Guide Star is a central problem in the framework of the development of Adaptive Optics Systems reaching full sky coverage down to visible wavelengths. In the past few years, various techniques aimed at solving this problem have been proposed. However only a couple of these has been recently tested in practice. We report about an experiment aimed at evaluating the performance of one of these techniques called the ‘Elongation Perspective’ technique. Our experiment has been performed using the ALFA system in Calar-Alto (Spain) and involves the simultaneous operation of the 3.6 m and the 2.2 m telescopes at the Observatory. This article describes the telescope configuration used, as well as the data reduction process carried out in order to estimate the scientific object tilt. The technique performances are discussed in terms of the residual tilt error variance and related correlation coefficient. The analysis shows that, despite the low SNR of our measurements, the atmospheric tilt variance is reduced to 80% of its initial value corresponding to a correlation coefficient of about 0.6. To get a better estimate of the performance achievable using this technique, the tilt error variance due to photon noise in the laser measurement is estimated and removed from the obtained tilt error variance. When this correction is done, this variance is reduced to about 50% of its initial value, showing that the use of this technique can give rise to a significant reduction of the scientific object image motion.