The origin of ultra-high energy cosmic rays (UHECR) has been widely regarded as one ofthe major questions in the frontiers of particle astrophysics. Gamma ray bursts (GRB), themost violent explosions in the universe second only to the Big Bang, have been a popularcandidate site for UHECR productions. The recent IceCube report on the non-observation ofGRB induced neutrinos therefore attracts wide attention. This dilemma requires aresolution: either the assumption of GRB as UHECR accelerator is to be abandoned or theexpected GRB induced neutrino yield was wrong. It has been pointed out that IceCube hasoverestimated the neutrino flux at GRB site by a factor of ~5. In this paper we pointout that, in addition to the issue of neutrino production at source, the neutrinooscillation and the possible neutrino decay during their flight from GRB to Earth shouldfurther reduce the detectability of IceCube, which is most sensitive to the muon-neutrinoflavor as far as point-source identification is concerned. Specifically, neutrinooscillation will reduce the muon-neutrino flavor ratio from 2/3 per neutrino at GRB sourceto 1/3 on Earth, while neutrino decay, if exists and under the assumption of normalhierarchy of mass eigenstates, would result in a further reduction of muon-neutrino ratioto 1/8. With these in mind, we note that there have been efforts in recent years inpursuing other type of neutrino telescopes based on Askaryan effect, which can inprinciple observe and distinguish all three flavors with comparable sensitivities. Suchnew approach may therefore be complementary to IceCube in shedding more lights on thiscosmic accelerator question.