mikee wrote:Well I don't want to go into statistics very deeply, but it might be worth noting that the sample size of 0.14-a is roughly 10 times that of 0.9-a and 5 times that of 0.11-a, so I'm not sure how telling those are.
What's important in terms of our interpretation is not the sample size nor even the size of the population involved (which may be what you mean here*) but the characteristics of that population in relation to likelihood of winning - skill and experience, most likely. Since we have no real way of knowing those it's not possible - as has already been suggested in this thread, by duvessa and others - to assess whether the difficulty of the game is changing or not.
Put another way, there are two major unknowns here: the difficulty of the game, and the average level of skillfulness of (all) the players. We might reasonably expect a change in either one of those, or in both, to lead to a change in the proportion of games won. The proportion of games won does change but we can't tell which of the putative explanatory factors has changed so we can't comment on whether the game has become harder or non - or at least, not on the basis of those numbers. (Qualitative judgements are of course possible.)
tldr: we can't tell.
* by which I mean that if you say "the sample size is bigger" is implies to me a larger sample drawn from the same population, which would enable more precise estimates to be made. I think you mean, and forgive me if I'm wrong, that more people are playing - i.e. that the population itself has grown larger.