Comparison

The results are almost identical to that of the Jansen et al. paper and shows that they can be reproduced and that the inference can be slightly improved with more information. The databases were not the same, as specific games that were included in Jansen et al.'s paper were not recorded and hence only the best available databases on the Internet were able to be used [7]. The compression is not high but, at around 5%, it is significant in that it clearly shows that the simple model, to a small extent, is predicting the moves of the grandmasters. A key item to note is that the inferred values differ by very little from Jansen et. al.'s and this would indicate that there is a clear difference between the playing strategies.