In our study (Bauville & Yamato, 2021), we derived simple formulas for pressure-to-depth conversion that allow (or not) to take into account deviatoric stresses. We applied these models to a data set of peak pressure (P p ) and retrograde pressure (P r ) of high-pressure metamorphic rocks. Because the equations have many free parameters, we presented end-member models to illustrate the influence of individual parameters on the estimation of depth. For clarity, we summarize the main results of our study (see Table 1 and Figure 1). First, we presented models where we estimated the peak and retrograde depth independently, assuming varying magnitude of deviatoric stress, from zero (i.e., lithostatic) to maximum shear stress (see Introductory model and L-model in Table 1). We noticed that the range of depth estimated for P p and P r overlapped. Therefore, we tested the hypothesis that P p and P r were recorded at the same depth. At constant depth, a change in stress state may lead to the transition from P p and P r . We calculated the combinations of stress states consistent with this hypothesis assuming either stress rotation or a change in horizontal stress magnitude (see S-model1, S-model2, and YB-model in Table 1). Figure 1a summarizes the depth estimate at which P p was recorded for the various models. We also presented the results of the models that assume constant-depth with scatter plots of pressure data on top of theoretical model results (Figures 1b-1e). The stress states are limited by the Mohr-Coulomb equation, where we assumed a coefficient of friction of 0.65 and a cohesion of 0. After this initial choice of parameter we did not fit the model to the data (data and models are independent in our study).We concluded that all data points could be equally well explained by models where (a) P p and P r are independent (L-model), or (b) P p and P r are recorded at constant depth (S-model, that is, all points fit in the model domain in Figure 1d). The model proposed by Yamato and Brun is a more constrained version of the S-model and can explain all data points except the ones we termed "outliers" (Figure 1e). The maximum depth for the P p event is ∼160 km according to the L-model and ∼80 km for the S-model. Thus, the depth predictions of the two models imply very different geodynamic scenarios. These models are only end-members, and exhumation and stress change are likely to co-occur.