We constructed graphs for all the countries where the number of infected depends on the number of days since the beginning of the epidemic. From this research we found that Gamma distribution works well for the prediction of the countries which increase their growth rates, but as soon as the country reaches plato condition, Gamma distribution works worse for predicting a trend.
From the graphs we constructed for all 6 countries we can see, that usually exponential curve works well at the beginning of the function, but as soon as the time passes exponential curve starts to grow too fast, and logistic curve becomes much better approximation. Good example of this is shown on the second graph of prediction curves of Russia. For exponential curve it is possible to find coefficients that work better than logistic graphs for some time, but later the situation will be similar the second graph.