Interpreting an engine correctly

I'm trying to better understand how to interpret an engine analysis so I can learn more effectively from key positions in games that I've lost. To help you help me, I'll ask 3 questions.

While Stockfish does its "thinking" it produces new rows, one move per row and the depth and number of nodes increases every row. Is the last row, the one produced the latest, always what the engine thinks was the best move in the position?

So if there's various moves with a score around -0.5 for white at a depth of 6-9 but at a depth 25 to infinity the engine only suggests one move let's say Bc5 that's -4, would that simply mean that white has a losing game and Bc5 is still actually White's best move?

If that is how it works, are some of the lower depth moves with better scores actually better against opponents at my level - the highest rating I've had is 1241 - because you can't expect the opponent to play perfect chess and calculate 30 moves ahead like an engine can?

Thanks, regards,
Roenie

If you're looking at a single line, the current depth is going to be the best move it finds. If you're looking at the position for white and the move suggested is your move, then that suggestion is the best move and everything else is worse.

Yes, there will be positions where you may actually need to be an engine to play it successfully, so going with a different move, that may be slightly worse, would be fine, especially if you understand that position better. However, if the evaluation is of a more tactical nature, that will resolve in a couple of ply or moves, just hoping your opponent won't see it is probably not the best line.

An assumption here is you're using the engine to study and not offer suggestions in ongoing games, which isn't allowed.

I probably should have shared what I'm actually looking at. I just realized the engine analysis on chess.com is presented in a very different way. After a game ends on chess.com there's a button that lets you copy the game (FEN). I hit that button, paste it into the BabasChess client, then tell Stockfish to analyze whatever position I want to learn about. Example (right click the image and open it in a new tab to enlarge it):

It's not about this specific position/game it's just an example of how it presents the data. It's similar to what you'd get from Fritz and such. In this example though, is Bd3 better than Rc5? At depth 25 it came back to analyzing Bd3 again and scored it worse for whatever reason without even adding the rest of the line but let's pretend that one lower score into negative 4 at depth 25 didn't exist and all the Bd3 scores were the ones at depths 13 to 20 that are in the -3 range, then would Bd3 be better than Rc5 which it presents at deeper depths and scores below -4?

"An assumption here is you're using the engine to study and not offer suggestions in ongoing games, which isn't allowed."

Yeah obviously- it ruins the game for yourself and the opponent both. So unless a person is pathetic enough to equate their self worth to a chess rating (I mean, really?) even when it's a lie there is zero point to ever pulling a Hans.

The very first number is how many ply the engine is looking. So at a depth of 27, Rc5 is considered the best move. Two ply earlier Bd3 was considered best. It's very likely the two moves are really close in eval.

Now it makes sense to me, thank you.

If a move gets a higher rating at deeper depths, this simply is the best move. You should play it.

Assuming your opponent is at a low level to never grasp the suggested move's meaning is what we call 'hope chess.' Learn to play the best moves in the position is the only way to play actual chess.

Remember Bobby Fischer saying that even though chess contains billions of moves, actual chess contains just thousands - since the best moves are the only playable ones. The position simply falls apart otherwise - and ends up a mess, not chess.

The only reason to deviate from a top engine suggestion is if you want to explore more dynamic / tactical / thematic ways to enter the next phases of the game at a slight positional cost. But yes, aiming to train that computer-like precision is the only way to move forward in chess. It's the hard truth that indeed, human thought isn't at all special and starts meeting that of our silicon-brained accomplices.