LIGO et al: Is it possible to estimate the frequency of GW+EM detections? In the past few years LIGO and VIRGO have detected a multitude of gravitational events. As far as I know however only GW170817 had a verifiable EM counterpart. Shouldn't it be possible, based on already available data and models, to estimate the frequency of future GW+EM detections?
 A: I think at this point this hard to do with any degree certainty. While we may be able to give a fairly robust estimate of the rate of GW detections from neutron star mergers (including projections for future detector upgrades). The big unknown is the chance that neutron star merger will produce an observable EM signal. Or differently put, we don't know how lucky we were to see an EM counterpart for GW170817.
The uncertainty comes largely from the messy physics of a neutron star merger being only partially understood, combined with limited knowledge of the equation of state of neutron star matter. The analysis of GW170817 has actually taught us a lot in this respect, but big uncertainties remain.
One of the conclusions of the analysis of GW170817 was that this event was visible because A) it was relatively close to earth B) because we viewed it from the right direction. The kilonova produced by a neutron star merger is much brighter when viewed directly along the jet. In the case of GW170817 the event was actually viewed slightly off this jet, producing a somewhat weaker signal. (We have seen similar gamma ray bursts at much greater distances, that were presumably viewed more "on axis".)
The chance of seeing an EM counterpart to a GW neutron star merger event is strongly dependent on the how the brightness of the EM counter part depends on the viewing angle, which in turn depends on unknown elements of neutron star physics. As we detect more NS merger GW events (with and without EM couterpart), this will actually help us pin down various of this unknown aspects of neutron star physics.
