Many ion channels are known to behave as an allosteric protein, coupling environmental stimuli captured by specialized sensing domains to the opening of a central pore. The classic Monod-Wyman-Changeux (MWC) model, originally proposed to describe binding of gas molecules to hemoglobin, has been widely used for analyzing ion channel gating. Here we address the issue of how accurate the MWC model predicts activation of the capsaicin receptor TRPV1 by vanilloids. Taking advantage of a concatemeric design that makes it possible to lock TRPV1 in states with zero-to-four bound vanilloid molecules, we showed quantitatively that the overall gating behavior is satisfactorily predicted by the MWC model. There is however a small yet detectable subunit position effect: ligand binding to two kitty-corner subunits is 0.4-to-0.6 kcal/mol more effective in inducing opening than binding to two neighbor subunits. This difference, less than 10% of the overall energetic contribution from ligand binding, is not expected in hemoglobin, in which each subunit is related equivalently to all the other subunits.