How to win a won game?

I, as a 2060 rated player, often get into +3 to +5 positions against 2100-2300 USCF rated players and lose my advantage, leading to a draw or loss. It’s not like I blunder the whole game away (although I’ve had my fair share of these), it’s more like not playing the critical line and making several inaccuracies. I wouldn’t hesitate to play the critical moves against lower rateds, but I seem to play too conservatively against higher rateds. So how should I prevent these inaccuracies and keep up the pressure?

If you lose a +3 position, then that is no inaccuracy, that is a blunder (??).
"The hardest game to win is a won game" - Lasker
Always check your intended move is no blunder before you play it.
"He who has a slight disadvantage plays more attentively, inventively and more boldly than his antagonist who either takes it easy or aspires after too much. " - Lasker
"When you sit down to play a game you should think only about the position, but not about the opponent." - Capablanca

I said that I make several inaccuracies, not a blunder. Thanks for the advice though. I think the second and third quotes appeal to me more.

#3
Inaccuracies do not exist.
If a move does not change the game state from "won" to "draw", then the move is not inaccurate.
If a move changes the game state from "won" to "draw", then it is a mistake (?), not an inaccuracy.
Whenever you have a won position and later it becomes a draw or even a loss, it is always possible to pinpoint a single move as the culprit: the mistake (?) that threw away the win.

Maybe that's true in a higher chess sense, but for actual OTB play it is not good advice. To kleong, I would recommend trying to understand why you making the mistakes. If it is because you played in the wrong manner, then maybe try to think of why you are better during the game. If you understand why you are better and are failing to find the right move at critical moments, then you could try to train calculation.

He's talking about a series of moves, such as a wrong plan, that change the game to a draw. Often times there is no one culprit, but rather one move that makes the win harder, then another that makes the position holdable, then another that makes it a dead draw, and it's hard to pinpoint exactly what went wrong until you check with an engine and it tells you that a series of 3 moves each lost you 1 point of advantage. In that case, it was not one blunder, but a couple of moves that would not have blown the advantage on their own, and that's what makes some wins tricky at higher levels.

Lasker should think again. It's a darn sight harder to win lost games.

Lol, unless you're a triple-digit player. 

Show a couple examples. That will make it much easier to give relevant advice.

Separately, I have a little room in my schedule if you are interested in professional coaching. Your level and types of problems are very much in my specialization. 

When you fail to win a won game, it's often because you have not adjusted properly, to the situation of having a won game. Sometimes people over-adjust and try to simplify to a won ending but they try to do it too fast and wonder why and how they got it wrong. At other times, they fail to adjust their thinking correctly.

When you can see that a game is won, then there's no more need to take a risk. If there may be a need to take a risk, it isn't won.

One more point is that you shouldn't be talking in terms of computer assessments because, for one reason, they often aren't realistic. Secondly, very often, you will not be thinking in terms of such an assessment while you're playing. Sometimes, you'll think you're doing better than you really are and at others, you won't see the winning route clearly. If it's the latter, you can't consider it to be won. If it's the former, you need to play especially diligently and always think in terms of the minimum requirement for a forced win, bear that in mind and play safe.

I'll add to that: Recently I got a +2 position out of the opening in an OTB game against an IM, then threw it away by making a questionable exchange sac without giving it too much thought, and then playing poorly after that. The moral of the story is to put time into conversion, if you find the right moves things will be easier.

Here's a couple examples in the National Online Quick this weekend against NM's:

Against NM Vladyslav Shevkunov

https://lichess.org/BUtPMu71/white#100

Against NM Joseph Levine

https://lichess.org/4q3Z5f8RKxpG

Can you explain, in your own words, why you were winning in those games and what you missed? The engine evals make it look simpler than it really is. Have you analyzed the games yourself?

I know I'm a patzer/CM (Choke Master) but I think you made it too hard on yourself in the first game. You had a chance to take on f6 and force favorable trades where you would have been up 1-2 pawns, but you tried a complex bisho maneuver instead, and after that the moves didn't seem so natural.

Game 1: Froze the opponent’s queenside, got a dominant knight on c4 and a strong pawn on e5. Eventually won a pawn and killed the opponent’s dark-squared bishop’s activity.

I missed exf6 followed by Nd6 and Re4, netting a huge advantage. I also missed that my a5 pawn was hanging later in the game.

Game 2: 

Attained a space advantage on the kingside as the opponent’s pieces were paralyzed on the queenside. With fewer pieces on the board, this is decisive.

I missed f5 (after Kf7) followed by Bf4 and Re7+, and I missed the amazing ...Ne5! from my opponent. I also hesitated to play f5 later, drawing the game because his knight is rendered ineffective.

Ne5 was a cool move, that looks easy to miss.

kleong just sucks lol imagine not winning +3 positions -_=

if u can't tell im joking

You should be aware that the engine position evaluation tends to be heavily inflated, towards the attacking side or the side with an advantage, tygxc. Indeed, if you look at any normal game, the computer evaluation tends to mark "engine moves" such as an early h4 against a fianchettoed position as "correct", when quite simply, it isn't.

Much as I repect him, I regard Capablanca as wrong, regarding not playing the player. You should be using all information and not just the board position. It's a mark of arrogance not to wish to use all the information. Of course, how well you use information about the player is another matter.

I may be much lower rated than you but I normally try to trade off material.