Blocking ?page= in robots.txt while linking to it. How do you handle pagination?

[slotbonus]

Found this on my own site last week and I suspect it’s common. I had Disallow: /*?page= in robots.txt to keep faceted crawl space down — but my listing pages link to ?page=2..8 in the pagination control. So I was linking to URLs I’d told Google not to crawl, which is what generates “Indexed, though blocked by robots.txt”, and the deeper pages of my main game listing had no crawl path at all.

First fix was to just delete the disallow. That was worse: measured 116 non-listing URLs then answered ?page=2 with a 200, so every game/article/legal page suddenly had an unbounded parameter space, and it exposed an existing 500 on ?page=-1 to Googlebot.
What I ended up with: block ?page= by default, then re-allow it per paginating route using Google’s longest-match rule (Allow: /games?page= at 14 chars beats Disallow: /*?page= at . Plus a redirect for anything malformed. Curious whether others handle pagination with robots at all, or purely with canonicals.

[allfreechips]

Stop using exposed url params a rewrite to something useable such as page2 ext, and then if its s bogus page then 404 it..

[slotbonus]

Already 404 anything past the last real page, and 301 malformed values back to the clean listing. page=-1 was throwing a 500 before I looked, hidden behind the robots block.
Paths would sidestep the robots side, agreed. But my sort/filter states are still query params, so /games/page/2?volatility=high mixes both anyway. Did you move facets to paths too?

[allfreechips]

really you want to avoid ? in the url. there are some cases it makes sense but usually on completely non indexed items.

[topcaz]

Look on the URL structure, it's just giving a name to your pages https://topcaz.com/compare-slots-arc...no_arch_page=3

how this will help