Elements of kinematical and dynamical modeling of elliptical galaxies (Es) are presented. In projection, NFW models resemble Sersic models, but with a very narrow range of shapes (m=3+/-1). The total density profile of Es cannot be NFW-like because the predicted local M/L and aperture velocity dispersion within an effective radius (Re) are much lower than observed. Stars must then dominate Es out to a few Re. Fitting an NFW model to the total density profile of Sersic+NFW (stars+dark matter [DM]) Es results in very high concentration parameters, as found by X-ray observers. Kinematical modeling of Es assuming an isotropic NFW DM model underestimates M/L at the virial radius by a factor of 1.6 to 2.4, because dissipationless LCDM halos have slightly different density profiles and slightly radial velocity anisotropy. In N-body+gas simulations of Es as merger remnants of spirals embedded in DM halos, the slope of the DM density profile is steeper when the initial spiral galaxies are gas-rich. The Hansen & Moore (2006) relation between anisotropy and the slope of the density profile breaks down for gas and DM, but the stars follow an analogous relation with slightly less radial anisotropies for a given density slope. Using kurtosis (h_4) to infer anisotropy in Es is dangerous, as h_4 is also sensitive to small levels of rotation. The stationary Jeans equation provides accurate masses out to 8 Re. The discrepancy between the modeling of Romanowsky et al. (2003), indicating a dearth of DM in Es, and the simulations analyzed by Dekel et al. (2005), which match the spectroscopic observations of Es, is partly due to radial anisotropy and to observing oblate Es face-on. However, one of the 15 solutions to the orbit modeling of Romanowsky et al. is found to have an amount and concentration of DM consistent with LCDM predictions.