"Just canonical the duplicate pages so they don't compete."

"Just canonical the duplicate pages so they don't compete."

Depends what you actually want — and people reach for canonical when they need a different tool.

Three instruments, three jobs. rel=canonical is a hint that says "these are the same page, credit this one" — Google can and does ignore it. noindex is a directive that says "keep this out of the index" — obeyed, but the page still gets crawled. robots.txt Disallow says "don't crawl this" — but a disallowed URL can still get indexed if it's linked.

The classic blunder: noindex a page AND block it in robots.txt. Now Google can't crawl it, so it never sees the noindex, so the URL lingers in the index forever. The two cancel out.

So: near-duplicates that should consolidate equity — canonical. Pages that must vanish from results — noindex, and let them be crawled. Crawl-budget waste you never want fetched — robots.txt, and don't expect it to deindex anything.

(Canonical to a noindexed page is the other trap. You're pointing equity at a door you nailed shut.)