The rule is that the game makes the list immediately if it has no ads or IAP, and requires manual review if it does. Games with unclear IAP/way to remove ads usually get a cursory dismissal, I err on the side of removal a lot.
This may be a winnable battle, or at least winnable for reasonable investment of effort.
What kills things like spam filtering is that A: we have no oracle of spamminess; if we did, there wouldn't be a problem in the first place and B: effort mismatch; fighting spam often takes more effort than sending it.
If I rate an app 1 or 5 stars, there's no objective way to say it's wrong. There are objective facts here, though; anyone who says Raid Shadow Legends is a "no bullshit" game is objectively wrong, by the definition given for what that sort of game is. This can be leveraged to build defenses.
I would build a system where I can pin a game as being either NoBS or BS in my database, attached to a time range. (Time ranges account for the fact that the situation can change, and also allows me to retroactively change determinations without having to instantly catch BS-level changes.) Any account that incorrectly labels a game within the given time range is shadow banned. Full stop. No mercy. (Any mercy becomes an exploit, so, sadly, there can be none, or at least, it would take a lot of human interaction.) Since we don't really care why an account is rating Raid Shadow Legends as NoBS, whether it's a spammer or a sincerely incorrect human, this approach purges either source of problems.
This flips the usual content moderation process cost/benefits on its head. Someone who wants to systematically review bomb Raid Shadow Legends as a NoBS game has to build out a lot of accounts and send in lots of reviews. Maybe you change things up a little bit to keep them on their toes like by changing out the CAPTCHA every so often or something. They burn non-trivial effort; you simply set a flag in the DB that anyone who rates it as NoBS is shadow banned. You wake up in the morning and check out the most controversial ratings and/or the games with the biggest surge in ratings, spot check them in the App Store (where it is usually obvious whether the game is NoBS pretty quickly), and set the relevant flags on that game. Not only do you prevent them from getting the satisfaction of an incorrect rating, you end up automatically purging out any other corruption they may have added while trying to create fake good looking accounts.
Since there is simply nothing the spammers can do to get Raid Shadow Legends marked as NoBS, this isn't the usual arms race. There's no exploit they find, then you have to close, then they find a new one... indeed, most of their "exploit finding" would work to your advantage as they still end up with accounts that simply get ignored.
A bit more gaming out, like, don't show literal uncooked numbers for ratings and a few other things like that, and I think you can set up a system where for just a bit of effort (not zero, yes, I'm aware of that, there is some ongoing effort going on here) you can defeat them, and they'll fairly quickly stop trying because their successes will be shortlived and ultimately self-defeating.
Spammers won't target every single game; they'll be paid to try to get specific ones mislabeled. By anchoring the DB on a relatively small set of objective truths, you can really take a bite out of their efforts to fake the data by basically setting the system up with a large number of high-powered landmines that have an extremely good chance of flushing these out, plus if you do the time-based locking, time is basically on your side rather than theirs... even if they did push something through last month, once you spot check their game their accounts get retroactively nuked once the objective truth is added to the DB again.
The really neat thing is, if you build a high-quality information source that can't be polluted through something like this, you could start building an audience, and if your audience becomes large enough, there would eventually come a tipping point where the mere existence of this app starts encouraging NoBS games to be built again.