The question of significant deviations of protein folding times simulated using molecular dynamics from experimental values is investigated. It is shown that, in the framework of Markov State Model describing the conformational dynamics of peptides and proteins, the folding time is very sensitive to the simulation model parameters, such as forcefield and temperature. Using two peptides as examples we show that the deviations in the folding times can reach an order of magnitude for modest variations of the molecular model. We, therefore, conclude that the folding rate values obtained in molecular dynamics simulations have to be treated with care.Modern computational power is enough to simulate small proteins up to the times when they fold into their native conformations [1][2][3][4]. This is a remarkable achievement because phenomenological, relatively simple interatomic interactions built into the model lead to the molecular structures that essentially coincide with the crystallographically determined native conformations.In contrast to the structure of proteins, the results on folding times are not as optimistic. For the majority of successfully folded proteins there are significant discrepancies between simulated and experimental folding times [5][6][7]. This is taking into account that only the results when the trajectories approach the folded conformations sufficiently close are published. In few cases even complete failures to reach the folded state in silico in simulations significantly exceeding the experimental folding times are reported [8]. Indeed it is well known in the modelling community how difficult it is no fold a protein ab initio, that is without introducing any information on the intermediates.By analysing the MD trajectories of peptides in explicit water we suggest an explanation for these discrepancies. We show that the folding rates are very sensitive to the details of the simulation model. The sensitivity is so high that the obtained values of folding times are meaningless and can not be compared between each other and with the experiment.We use the Markov State Model (MSM) [9-12] to describe the folding process. The configurational states are defined by clustering the MD simulated trajectories. This is done by analysing the Ramachandran plots of the residues of the peptide, Fig. 1. Each Ramachandran plot is clustered independently and the molecule's configurations are defined by the cluster indices from each plot. Not all possible combinations of index values are realised in the trajectory. For example, for the peptide from Fig. 1 the conformation B 1 C 2 was very scarcely populated and was, therefore, joined with A 1 C 2 into one conformation, thus resulting in 5 total configurations of the molecule.In the MSM framework the model is described by a state vector v, which holds probabilities of all the configurations at a given moment of time, and a transition matrix T . The total probability of the state vector has to sum to 100% since the peptide has to be in some configuration at any tim...