Should you trade queens or keep queens in this position? Move 22 for white.

I was expecting Qh3, so I was planning on trading queens. I figured white had a good position with 2 rooks vs. black's rook and knight, as well as an e pawn advancing. However, Stockfish is showing that keeping queens on the board is better. What would you do in this position? I have included 2 possible lines.

I would trade, but only when I could do it without ruining my pawn position. I don't like gxh3. So agree that Rd2 looks best.

It does not matter. White is winning anyway: up an exchange.

I like Qc3 the most but I would prefer Rd2 over Qd8+.

A line that's evaluated +2.1 is only very slightly better that a line of +1.8. You should look at the numbers, and not worry about tiny differences such as this one. Those are not significant.

Your replies to my posts seem to be too quick. Please look more carefully at what I am asking.

First, you say, "You should look at the numbers, and not worry about tiny differences such as this one."

Your example of tiny numbers is .3 difference. But Stockfish is showing this as what seems to be two significant differences.

With e5: 3, 4.7, 4.7

With Rd2: 4.7, 4.9, 5.15

This is not .3 difference. This is 1.7 difference with e5, and 1.8 with Rd2. If you compare e5 first result with Rd2 first result you get 1.7 difference. Are you claiming that .3 difference and 1.7 difference is too small to worry about? Please look more carefully into this.

It looks like I shouldn't trust chess.com's weak engine. Running a real Stockfish on the position gives even higher differences, an the top choices (Rd2 and Rd3) are over +6.3. Meanwhile, the e5 line is just 3.7, presumably because white's pawn structure is so bad. I would never have thought a difference of this magnitude. Sorry for being too fast.

@6

"With e5: 3, 4.7, 4.7

With Rd2: 4.7, 4.9, 5.15"

++ There is no difference at all: those numbers just say white is completely winning either way.

And the 1200 wins!! "Great" minds think alike.

Going strictly by the engine, we can say Rd2 is the best move. However, with human play, my argument is piece activity in the endgame can dominate the chances of winning. Is there a straight forward win without opening the g file for the king? The instinct is to say it is a weakness for the side who has the g file opened (on that move).

I suggest comparing and contrasting this game that finished in a draw. Black had the opportunity and had their g file opened, but they did not get their king into play until later in the game. This is what the engine is showing. It appears to me, that the downfall was allowing the a pawn to be traded. On move 24, white didn't go for the trade and played Re3 instead. But then on move 29 black gave white the opportunity to trade it again.