You must be registered to see the links
It's a scraper API that looks at steam data to determine how many have been sold or bought. Back in 2018 Steam hid the exact numbers of games sold however so it's more of a guesstimate at this stage but I'd take this 20k sold over your 250k sold simply due to how many current people are playing the game, which is 1,700 people. Since it only came out last week and if it did sell more than 200k copies, there would be more than 1,700 people playing. You don't see a 99% drop off in players inside of a week, especially for a game as highly publicized as this.
Judging by how many steam reviews there are is a false metric since it can be easily wrong. Take Greedfall for example, Steamspy says it sold 500k to 1M copies. Your estimate of 41 copies sold per review at 10k reviews put total sold at 420k.
Link to article saying the company sold over a million copies.
You must be registered to see the links
What you're essentially doing is what Harold Camping did for his end of the world predictions and pulling numbers out of nowhere with figures that makes no sense to the actual value that is way off of what it should be.
As for Subverse itself, 2 million raised in kickstarter, be generous and say 20k bought through steam which means about 700k raised though that. Total budget be generous again and say $3 mil. Certainly more reasonable than your 7 mil budget don't you think? There is no way this project has generated that much of a budget, simply for the fact if it did they would have hired actual fuging devs to work on the game, not make it some space shot em up that a grad student could make for finals.
This game is the Second Ranked Game on the Top Seller Ranklist 
