(Abridged) OJ287 is a BL Lac object that has shown double-peaked bursts at regular intervals of ~12 yr during the last ~40 yr. We analyse optical photopolarimetric monitoring data from 2005-2009, during which the latest double-peaked outburst occurred. The aim of this study is twofold: firstly, we aim to analyse variability patterns and statistical properties of the optical polarization light-curve. We find a strong preferred position angle in optical polarization. The preferred position angle can be explained by separating the jet emission into two components: an optical polarization core and chaotic jet emission. The optical polarization core is stable on time scales of years and can be explained as emission from an underlying quiescent jet component. The chaotic jet emission sometimes exhibits a circular movement in the Stokes plane. We interpret these events as a shock front moving forwards and backwards in the jet, swiping through a helical magnetic field. Secondly, we use our data to assess different binary black hole models proposed to explain the regularly appearing double-peaked bursts in OJ287. We compose a list of requirements a model has to fulfil. The list includes not only characteristics of the light-curve but also other properties of OJ287, such as the black hole mass and restrictions on accretion flow properties. We rate all existing models using this list and conclude that none of the models is able to explain all observations. We discuss possible new explanations and propose a new approach to understanding OJ287. We suggest that both the double-peaked bursts and the evolution of the optical polarization position angle could be explained as a sign of resonant accretion of magnetic field lines, a 'magnetic breathing' of the disc.

We report a coordinated multi-band photometry of the RS Oph 2006 outburst and highlight the emission line free y-band photometry that shows a mid-plateau phase at y ~ 10.2 mag from day 40 to day 75 after the discovery followed by a sharp drop of the final decline. Such mid-plateau phases are observed in other two recurrent novae, U Sco and CI Aql, and are interpreted as a bright disk irradiated by the white dwarf. We have calculated theoretical light curves based on the optically thick wind theory and have reproduced the early decline, mid-plateau phase, and final decline. The final decline is identified with the end of steady hydrogen shell-burning, which turned out at about day 80. This turnoff date is consistent with the end of a supersoft X-ray phase observed with Swift. Our model suggests a white dwarf mass of 1.35 \pm 0.01 M_\sun, which indicates that RS Oph is a progenitor of Type Ia supernovae. We strongly recommend the y-filter observation of novae to detect both the presence of a disk and the hydrogen burning turnoff. Observational data of y magnitudes are provided here together with other multi-wavelength light curve data.

We report a coordinated multi-band photometry of the RS Oph 2006 outburst and highlight the emission line free y-band photometry that shows a mid-plateau phase at y ~ 10.2 mag from day 40 to day 75 after the discovery followed by a sharp drop of the final decline. Such mid-plateau phases are observed in other two recurrent novae, U Sco and CI Aql, and are interpreted as a bright disk irradiated by the white dwarf. We have calculated theoretical light curves based on the optically thick wind theory and have reproduced the observed light curves including the mid-plateau phase and the final sharp decline. This final decline is identified with the end of steady hydrogen shell-burning, which turned out the day ~80. This turnoff date is consistent with the end of a supersoft X-ray phase observed with Swift. Our model suggests a white dwarf mass of 1.35 \pm 0.01 M_\sun, which indicates that RS Oph is a progenitor of Type Ia supernovae. We strongly recommend the y-filter observation of novae to detect both the presence of a disk and the hydrogen burning turn-off.