There's one more sad thing about Steam achievement rates: most of them are using readily available achievement unlocker cheats. Jamestown in particular had a lot of people use an unlocker, likely because of a holiday-exclusive achievement that gave a bonus to one of Steam's sale events (it wasn't hard in the slightest by anyone's standards).
You can easily see this by taking a good look at the achievement clear rates for things that require a good chunk of effort versus the things that require an insane level of effort. In Jamestown's case, that'd be 1CC on Divine vs no-hit on Judgment. In CC:WI, the ordering of the achievements actually tells that there's more Unlimited 1CCs than Boost 1CCs; the leaderboards of course tell a way different story with a hundred or so Boost 1CCs but only like a dozen unique players with any Unlimited 1CC across all ship types (there can of course be valid 1CCs that don't make it to the leaderboards, but the gap between modes is disproportionately large and I'd expect the mode built around skyrocketing rank to be more affected by that anyhow, seeing as you can get a relatively easy but low scoring 1CC by intentionally bombing every 29 seconds).
Steam also filters out players that never even played the game, so it'd be even lower than that if you look at full ownership rates. SteamSpy reveals a
~44% player rate for CC:WI and a
~57% player rate for Jamestown (compare "Owners" and "Players total"). Not surprising due to them being bundled and put on sale for a few dollars, just saying it's not a factor.
Ikaruga seems to have a much lower rate of full-unlockers and I think Eschatos and XIIZEAL are mostly legit as well. Those are games that haven't been in bundles or thrown on sale for less than a dollar. The latter two in particular are names that nobody outside the shmup world is likely to recognize, you pretty much have to go hunting for them and I'm not expecting much impulse-buys either. SteamSpy's player rates on those are also vastly higher;
Ikaruga is at ~86% while
Eschatos is at ~100%.
It's hard to get real stats, but if you were to outright throw everyone with a CC:WI Unlimited 1CC out of the stats for that game (noting that the dozen or so people that have legit clears on the leaderboards are an anomaly compared to the skill gap between Original and Unlimited), you'd only end up with 0.1% for Arcade-Original. Likewise, Jamestown Divine would be 0.1% after you throw out the 1.1% of people that have done something that there exists no video evidence of (I'm sure someone's no-missed Judgement but if 4000 people did it then there'd at least be a run on YouTube, right?).
The "aim for an 1CC" arcade-game playstyle just doesn't seem to be a thing at all for the mainstream population, and thanks to people cheating for no particular reason, the completion stats for shmups that gained popularity in the mainstream are
even worse than they look at first glance. They're not just bad, they're outright abysmal to the point where clears can be easily traced back to the shmup community.