Primary Source: Cognitive Lode
Giving users an opportunity to achieve closure after making a choice will make them feel better about any choice they make.

Closing a Laptop (source example) | Closing my laptop at work to go to lunch feels so good. (A random picture of my desk at work. Just looking at it now I want to shut that laptop and go to lunch.) I also typically only close my laptop after I have made a decision and will often leave my laptop open if I haven’t made up my mind yet.

Amazon Post-decision Ads (Internet example) | This is a bad example of a situation that doesn’t provide choice closure. I find myself second guessing myself and feeling the need to browse other similar items.
[ source: https://www.amazon.com/gp/huc/view.html?ie=UTF8&newItems=C3AUQK6K1GR1AF%2C1 ]

Jira ticket status (non-Internet example) | One of my favorite things at work is setting “UX Done” or “Design Complete” on a Jira ticket. This is not strictly an intentional design, this has to be done so that the team can track its progress, but it makes me feel really good to close out a design and make it final for developers who are going build it into a real interface.

In your online example you said that the reason why you don’t feel closer from Amazon is because you end up second guessing yourself. Do you think this is because of the multitude of other options out there or perhaps it is because of addictive behavior patterns? In the Civilization game series, there is closure once you’ve won the game by completing one of the major objectives. However after you’ve won or lost, Civ asks you whether you want to quit or play for one more turn. The game still continues after it’s completed as you can still play and mess around. Like your example with Amazon, I find it very hard to convince myself to click on quit for closure. I get an itch to keep on playing…just…one…more…turn. Unlike Amazon, there are only two options in Civ which makes me think closure is ineffective if the user is already addicted?
Well, I think this principle is less about allowing the user to make a choice and more about whether or not the design provides closure for whatever choice the user has chosen to make. For example, if you do choose to quit the game in Civ, if there is a screen that appears after you choose that option featuring a graphic with say a door closing, you would probably feel better about your decision than if that screen wasn’t shown.
For Amazon, users might be more satisfied if they were maybe able to add similar items in a pool for comparison and then shown a better closure screen once they made their final selection out of a narrowed down field of options.