@yebellz:
When I say theoretical maximum, here's what I really mean: In gunboat, there is an upper bound induced by the nature of the game; if several of your neighbors attack you in the opening and keep at it for a few years, you die, and you can't prevent it. Having experimented with AIs a bit, I believe that the maximum win rate for a good player against a selection of the weakest players on webDip would be about 70% in gunboat, and against the average set of gunboat players I've seen in live games, I suspect closer to 30-40%, lower when you include games where CDs play a major role. The problem is, 35% is low enough that it leads to a meaningful cap. So, I believe it is impossible for a mostly-gunboat player to break into, say, the top 10 without only playing carefully-selected, high-ranking opponents. Hence my feeling that I need to choose between aiming for a high GR and playing a lot of live gunboats (and picking the latter, since it's more fun for me). I suspect that if I were allowed to read all of my opponents' orders, I could maybe win 70% of gunboat games against GR 100 players; even at that point, I would barely become the top-rated player.
On the other hand, a diplomat has many more tools at their disposal in a press game; win rates against the same quality opposition are much higher. There is still, presumably, an effective cap, but it's higher than the highest current GRs.