In this article, we introduce a new model derived from Polya–Eggenberger urn model. This model is defined by considering a delay in collecting information. The mathematical formulation of this model is done through four parameters; the number of balls in the first structure (N0), the number of white balls in the first structure (W0), the number of rewarded balls of the color of the ball withdrawn (a), and the length of the delay (i). As the first attempt to deal with point estimation in this model, we consider at any time one of the parameters separately as unknown conditioned to knowing the other three parameters, and find its estimation. Accordingly, we introduce a sufficient estimator for this model, and found on it, obtain the maximum likelihood estimators for each of the four parameters. In addition, moment estimators for N0 and W0 are calculated. Also, for the other parameters, we obtain estimators based on the correlation coefficient of consecutive withdrawals. To evaluate the performance of the obtained estimators and compare their accuracy, we apply five simulations of the delayed Polya urn model. The simulations have been done with the software Matlab R2015b. According to the simulation study, the estimators obtained from the method of moments are preferable to maximum likelihood estimators.
