1. e4 e5 2. Nf3 d6 3. d4 exd4 4. Nxd4 Nf6 5. Nc3 Be7 6. g3 d5 7. e5 Ng4 8. Bg2 O-O 9. Nxd5 Bc5 10. c3 c6 11. Ne3 Nxe5 12. O-O Re8 13. b4 Bb6 14. a4 a5 15. Bb2

<Immediate problems should have been caused with 15. b5.>

15...Na6 16. bxa5

<16. b5 is optically more pleasant, once again.>

16...Bxa5 17. Qc2 Qf6 18. Rad1 Qg6 19. Be4 Qh5 20. c4 Nc5 21. Bg2 Bh3 22. f4?

<22. Bxh3 is obvious. Even 22. f3 should have been considered.>

22...Bxg2?

<I don't see why 22...Ng4 was avoided.>

23. Kxg2 Ng4 24. Nxg4 Qxg4 25. h3 Qg6 26. Nf5 Ne4?!

<OK, I'm not a Grandmaster, so it's a little bit odd that I write like this, but I would have gone for the much more natural 26...Re6 instantly.>

27. g4 h5

<27...Nf6 seems to be more natural.>

28. Nxg7 hxg4 29. Nxe8 Rxe8 30. h4 Qh6 31. Qd3

<31. h5! was better.>

31...Qxh4?!

<Objectively not precise, however, in a blitz game this is a very good practical move. The computer wants 31...Nc5, but this is irrelevant, given the time control.>

32. Be5?

<I think White missed Black's more or less obvious plan to occupy the g1-a7 diagonal with the bishop, suffocating the king. 32. Qe3 or 32. Bd4 come instantly into consideration, with edge for White.>

32...Bb6 33. c5?

<Why? I don't see the problem with 33. Qb3 Re6! 34.c5! Bxc5 35.Rd7 with some pull for Black, although the position is a tightrope for both players, certainly.>

33...Bxc5 34. Rde1??

<Gross miscalculation, making the position completely equal. 34...Rd8 was required, and there are checkmates everywhere.>

35. Rxe4

<35. Qb3+ is also OK.>

35...Qh3+ 36. Qxh3 gxh3+ 37. Kxh3 fxe4 38. Kg4 Kf7 39. Re1 e3 40. Rc1 Bb4 41. Kf3 Bd2 42. Rb1 Re7 43. Ke2?

<43.Rh1 Ke6 44.Rh5 with equality.>

43...Ke6 44. Kd3??

<With 44. Rg1 c5 45. Rg7 counterplay and/or simplifications should have been sought for.>

44...Rd7+ 45. Ke2 Kf5 46. a5

<Interesting trick, that utimately leads nowhere.>

46...Rh7?

<46...Ke4 was the option to win quickly. Later we will see how devastating are the ideas of Bxa5 and Rd2+. Capturing the pawn is unnecessary: it leaves Black in a better position, but guarantees nothing more.>

47. Kd3

<Maybe 47. Ra1 offered better practical chances, but White is still dead lost, of course.>

47...Kg4??

<A blunder that is very difficult to understand, even if it was made in a terrific time trouble. Throws away the easy win. After 47...Rd7+ 48. Ke2 Ke4 White cannot simultaneously defend the a6 pawn, and discourage Rd2+ ideas.>

48. Rg1+ Kf3 49. Rf1+ Kg2 50. Rb1??

<Whether in a time trouble or not, this is a very serious, almost amateurish blunder, being extremely rare for 2700+ players. The king should have been moved away from the open line, and the f-pawn should have been protected. I checked it with a computer (it is totally clear that deep analyzing a blitz position is stupid and meaningless, but for its theoretical value I have done it anyway), and wasn't able to come up with anything better than 50. Ke2 Bxa5 51. f5 Re7 52. Bf4 Rf7 53. Bxe3 Bc7 54. Rf3 Be5 55. Bg5 b5 56. f6, which is equality.>

50...Kf2 51. f5 Rd7+ 52. Bd4

<52. Kc2 is not good, either, as the a5 pawn will fall without compensation.>

52...c5

<Ouch! :)>

Magnus Carlsen was the world's highest rated classical player at the time of this game, but not yet the 16th world champion. Carlsen has always been a fine blitz player. He was world blitz champ in 2009, and various years later.

Shakhriyar Mamedyarov has been one of the world's best players this century. Mamedyarov scores better than 50% w/the traditional 1...e5 Philidor Defense, although he usually heads for the Spanish/Scotch/Italian Game with 2...Nc6 instead of 2...d6.

Stockfish assigns plenty of <?>s here. In blitz, there is no time to seek the absolute best move. Instead, quickly play a reasonable move that generates threats, ideas and doesn't leave the backdoor open. For example, 17...Qf6? became rather useful for Black's kingside pressure, even though Stockfish does not approve of how the assault was conducted. Make trouble for your opponent (without getting into trouble)!