Why are these moves inaccurate?

Hi

I recently played a game.

https://www.chess.com/analysis/game/live/14656625637

Why are moves no. 17 and 27 inaccurate? It would be great if someone can help. Thanks.

Bharat

I wouldn't worry about inaccurate moves too much. Sometimes it is because of a slightly better move possible (although the game move wasn't bad). Sometimes it is because of some deeper positional concept a 2000+ player might see, but most chess players won't. The computer can't "tell" us the positional idea was missed because the computer just calculates and doesn't understand "position" the same way humans do. Sometimes the innaccuracy is actually wrong. Well, not objectively per se, but wrong because the computer chooses the lesser of two evils objectively, but the other route was actually tougher to convert and therefore better chance the opponent will slip up (but the computer assumes the opponent will find all the best moves).

Move 17 was innaccurate because the computer liked ...Ne4 slightly move, but the game move was good too. 

Move 27 is worse than this, but not so easy to see why. 27...Rb7?? is checkmate in 5 exactly how the game went. 27...Nd2! was very tough to find but it saved a draw and this is better than checkmated!

27...Nd2! and one line might be:

28. Rxd7 Rb1+ 29. Kh2 Nf1+ 30. Kg1 Nd2+ and Black can repeat for threefold repetition and a forced draw via perpetual check

To add onto @KeSetoKaiba's point, 17...Bd6 was an inaccuracy because it retreats your bishop from a strong square, losing space on your opponent's side of the board. Additionally, the suggested move, 17...Ne4, develops your knight further into the center where it controls more squares. e4 is a really good square because knights in the center control much more of your opponent's key squares than on f6 or c6, although those squares are decent as well.

Thanks, @KeSetoKaiba  @K_Simonson I am white here .. But yes now I can see why my 27th move was suboptimal as a draw could have been forced through Rb1+ continuation. 

For move 17, I played Na4(-2.03) whereas Engine suggests moves like Ra2 or Be1! I think with the help of the engine, now I can see playing Ra2 does enable doubling of rooks on a-file faster!

Here's something important to remember: Unless you are at least a FIDE Master, "inaccuracies" simply don't matter! Chess games are won and lost because of big mistakes. In any game below master level, and even in many above that level, there are plenty of small mistakes, but that isn't what determines the outcome. Avoid big mistakes, and take advantage of your opponents big mistakes, and you will be a very strong player

Thanks @KeSetoKaiba @Laskernephew  for your feedback. Yes, I understood .. I actually felt that in this game, that I was kind of stuck( Where I could not see much, what to do next). So, thought of analyzing this game in greater depth. 

There is nothing wrong with move 17.

Move 12 b4? of white was a mistake as it loses a pawn which gave you a won position.

Move 20...Ne4? was a mistake that throws away the win. You had to preserve your pawn with 20...a4

Move 27...Rb7? of black is a mistake, that loses.

Thanks @tygxc