Why no two lists agree
Run the same search twice and you will get two different winners. That is not because the evidence is contested. It is because the orderings are not derived from evidence at all.
A roundup is assembled from products that operate affiliate programmes, which silently excludes several reputable ones, and sorted substantially by what each pays. Product descriptions come from the manufacturer. The science paragraph is usually written by paraphrasing last year's roundup rather than by opening a journal.
This site earns a commission on one product, so treat that paragraph as self-interested and check it. It is also true.
The substitution that makes the question answerable
Stop asking which product is best. Ask instead which ingredients have been given to real people under randomisation, at amounts a capsule can actually deliver, and then whether the capsule in front of you delivers them.
That version has an answer. It does not depend on anybody's opinion, and it empties most of a shelf inside a few minutes.
Twenty minutes, in order
- Photograph the back of every bottle on your list. The front is advertising. The back is the product.
- Discard any with a proprietary blend. A combined figure covering several ingredients means no individual dose can be established, by you or by a pharmacist. This usually removes half the shortlist.
- Note which evidenced ingredients appear, and at what amount. Then compare each against the dose used in the study people cite for it. A fashionable ingredient at a tenth of the studied quantity is decoration.
- Check the form, not only the name. A standardised extract commits to a concentration of named compounds. A plain weight commits to a weight of plant material whose potency moves with the harvest.
- Read the inactive ingredients line. Capsule shells, allergens and anything that rules the product out for you personally live here.
- Count the capsules and look at their size. A three-a-day regimen tends to lapse around week five, and ninety capsules at three a day is a month's supply rather than a quarter's.
- Take the survivors to a pharmacist with your prescription list. Two questions, four minutes: does anything interact, and is anything here present at an amount worth having.
- Price the quarter and read the refund window. Covered below.
The two numbers everybody compares wrong
Price. Per-bottle figures compare the wrong unit. This is a category measured in seasons, so convert everything to a ninety-day total before putting two products side by side.
The return window. Read it against how long the ingredients need. Fourteen days to decide, on a product whose quickest supporting study ran six weeks, is a deadline built to close before an answer could exist. It tells you how seriously a company treats its own research.
What the method cannot do
It will not identify a product that works. It identifies products that could plausibly work, which is a smaller and more honest result. Even for the survivors, the effects in the underlying research are modest, arrive over weeks and months, and were absent entirely for a meaningful share of participants.
And the comparison that beats all of it is not between bottles. A hearing test, a blood pressure reading and a daily walk each sit on firmer ground for cognition than anything on this shelf, and two of them are free.