The Duke of Ben
Sz is correct about this specific game, but I don't think his analysis goes far enough into why you find yourself losing more against lower rated players than higher rated.
I've noticed that HHH seems to have the most trouble with people rated between 1600-1700, and I think for the same reason you lost this game. People at that rating level know how to expand well, can fight all right, and have the other necessary skills to play their position, but they pick "weird."
To have a consistently high rating, you need to pick consistently good locations. Siberia would have been a wasted pick if Summer had gotten his #3 in Central Russia. That makes it a lackluster pick against tons of other picking strategies. That's a good bit of why he's rated 1800 instead of 1900+. Had you picked SA/WA, then he would be sitting on two picks to expand into Central Russia which you counter from West Russia, sealing the game for you in most scenarios.
A higher rated player would never have picked #3-#4 to double-pick/counterpick Central Russia, making your choices fine against, say, Dunga or Zibik (they would still be very likely to pick Pakistan IMO, and make West Russia and Central Russia both a hard thing to defend).
I think you have in your mind a set of areas that are "safe" against good players, and then try to use them against players who have different (frequently less good) picking strategies.
The correct strategy, I think, is to pick significantly safer bonuses against 1800ish rating players, even if you don't expand as well from those. West Africa in your game was a fairly poor expansion spot, but very safe, especially if you could get South America to go with it. You can use your knowledge advantage or their poorer picks to make up for picking sub-optimal picks on your side.
Since players rated less than 1600 can and do pick really randomly, I find it best to simply pick the best spots possible and out-expand them, so that I can win the first encounter and control the momentum for the rest of the game. Trying to pick safe spots or counterpicks against that group, even as high as 1700 rating, can be futile.