Singlet fission promises to surpass the Shockley-Queisser limit for single-junction solar cell efficiency through the production of two electron-hole pairs per incident photon. However, this promise has not been fulfilled because singlet fission produces two low-energy triplet excitons that have been unexpectedly difficult to dissociate into free charges. To understand this phenomenon, we study charge separation from triplet excitons in polycrystalline pentacene using an electrochemical series of 12 different guest electron-acceptor molecules with varied reduction potentials. We observe separate optima in the charge yield as a function of driving force for singlet and triplet excitons, including inverted regimes for the dissociation of both states. Molecular acceptors can thus provide a strategic advantage to singlet fission solar cells by suppressing singlet dissociation at optimal driving forces for triplet dissociation. However, even at the optimal driving force, the rate constant for charge transfer from the triplet state is surprisingly small, ~10 s, presenting a previously unidentified obstacle to the design of efficient singlet fission solar cells.