We carry out a Bayesian model selection analysis of different dark energy parametrizations using the recent luminosity distance data of high redshift supernovae from Riess et al .
2007 and from the new ESSENCE Supernova Survey .
Including complementary cosmological datasets , we found substantial evidence ( \Delta \ln ( E ) \sim 1 ) against a time-varying dark energy equation of state parameter , and against phantom dark energy models .
We find a small preference for a standard cosmological constant over accelerating non-phantom models where w is constant , but allowed to vary in the range -1 to -0.33 .