In recent times I have encountered many opponents, whose ranking is in the range of 2200-2500. Playing games with these strong players is a daunting task and I am still learning, how to cope with this based on my (not always satisfactory) experience.
A few things I found out meanwhile:
It is useless to play for advantage during the opening. Players, whose ranking is 200 points stronger than mine, are bound to know the moves better. So using sharp variations or gambits will only increase my sufferings. Instead I have to try for equalizing openings and variations, this in addition to me monitoring my development and the safety of my king.
With White I try often: 1. Knf3 - 2. g3 - 3. Bg2 - 4. 0-0 -
This brings my king out off the middle in four moves and gives me three tempo too, though it is not very aggressive.
With Black against e4 I try Russian: 1. e4 e5 2. Knf3 Knf6, yet I also need to be prepared in Four Knights, Vienna, Middle Gambit and Game and Kings Gambit. Still a task to have variations ready in all these openings.
With Black against other attempts of White I try the Indian systems, I am quite glad, if I end in a Queens Indian Defence, which plays more smoothly.
The task is always not more, than to get an equal middle game position, in which I am not behind in development, not cramped and have no major weaknesses. Unfortunately this is not as easy as it sounds, considering my humble abilities.
Again and again I stumble across opening variations, which are unknown to me or in which I lack experience and so I am often bound to make the first and often deadly mistake. Every fourth game against stronger players is already lost in the opening.
Next come miscalculations, when I try yet again to calculate more than fives moves ahead in tactical variations or in judgements of strategical positions. I feel, that a variation or position is in my favor only to find, that after reaching it I have a bad or even lost game.
For all my woes and worries I will now, after so much talk, give a practical example with the above game: my opponent here had White and a rating close to 2500.
Here are my comments:
11. -Bd7 I saw trouble ahead against my double pawn in the half open c-line and a knight snuggling on c5, so I tried to get counter play along the b-line with my Queen
14. Knb3 here he is coming already! I decided to advance my a-pawn, even driving the knight to c5, yet preventing the other of supporting the first one. Also I hoped for play against b2.
17. Kng5 creating a weakness in my castle
18. - Qxb2 the pawn comes at a cost!
21. e6! I overlooked this move: taking the rock is deadly as I saw NOW: 21. - Bxc2 22. ef+ Kg7 23. Qe5+ Kh6 24. Kne6 Bf5 25. Qg7+ Kh5 26. QXh7+ Kg4 27. Qh3++ (Mate in SEVEN)
So I contemplated the whole stuff again and thought, that I might be able to lure my opponent into a positional trap based on f2. Yet I had to calculate so many moves ahead, that my judgement was more based on mere intuition instead of logical calculation.
27. Rxd5 So, this was the position I strove for, beginning with 21. e6 and I was convinced, that after 27. - Rxf2 due to my "threats" I would certainly win. Yet NOW I suddenly saw, that I am just going to loose a pawn on a4 and after a exchange of rocks will end in a bad endgame. 
28. Qe7+ It was this check I had overlooked!
31. Rxa4 And here goes the a-pawn! Fortunately the white rock is in front of the pawn and I could get my rock behind it...
Computer Analysis Of The Game:
Inaccuracies: 5%
Up to move 23 the game is equal.
23. - Rf7 (Inaccuracy 0.52) better Kh8(?)
Yet the given variation after Kh8 leads to a position, which is even worse than that in the game!
31. - Kg7 (0.8) With that last move you went from roughly equal to having a worse position. Your best continuation was Rc7.
32. - Re2 (0.97 = so the endgame should be a win for my opponent?)
38. - Ra1 (Inaccuracy 1.22) better Kf6
39. - Ra2 (Inaccuracy 1.45) better Kh6
Yet the final variation in the computer analysis also ends in a repetition of moves!? So there seems to be no win in the endgame despite the evaluation of more than a pawn down (2.04 at move 44) which is a little strange.
61. - Rb1+ 62. Kc4Ta1 63. Kb4 Ta2 64. Kc5 Rc2+ 65. Kd4 Ta2 66. Ke4 Te2+ 67. Kd4... Nothing gained for White!
Does anyone see a winning method in this position for White???
Then please write a comment about it for the enlightenment of us all!!!
The end result of the "combination" was based on a grave error of judgement by me, striving for a position, in which I felt, that I am winning. I was lucky though, that the error was not grave enough for loosing the game in this case.
Well, I am still trying to improve and to learn, yet with half a century in my bones it's not an easy task anymore!