There is no equation to this ranking. While I have looked at some mathematical models such as F1metrics and based it around them, this is still a subjective ranking. Yes, some of the minor placings are less meaningful, most drivers outside the top 100 could move forward or back around ten places. The general answer of what it is based on is what positions I think the drivers would finish in a championship where they were all in equal cars, assuming it could somehow be standardised so that Hamilton's Mercedes now is equal to Fangio's Mercedes then. There are a few exceptions to this, as I believe Michael Schumacher would win that championship, but have dropped him to second because of Jerez 1997 (and others). I don't really want someone who crashed into another driver deliberately as the 'greatest', but on pure driving ability, I think he was. As I said, the top four are all extremely close and I am undecided on the order of them, so I might have Schumacher number one anyway. Also, the likes of Mark Donohue are penalised because he only drove in fourteen races, so he has a small sample size in terms of how good he really was. Fagioli is much higher even though he only raced in seven, but he was teammate to two known quantities in the best car in that time, so there is a clearer picture of his ability. Had Donohue survived to race for a full career, he likely would have been 100 places higher, but as it stands, he didn't drive in enough races. I could have put him a bit higher. And the list is a best F1 drivers list, meaning the Formula 1 championship between 1950 and 2021. It doesn't look the same as a general 'drivers' list, but I don't know enough about other series to make a general drivers list. It does not make it pointless to not include other series. In fact, I would say it is arguably better because then it is very clear which races from history are taken into account, and there are still over 1000 of them. I should note that the Indy500 from the 50s was not included.
I will now justify my placings for some of the other drivers in the list. Remember that in a list of 200 drivers, there is almost always something you will disagree with. That doesn't make it laughably wrong.
Giancarlo Fisichella/Juan Pablo Montoya - the simple answer here is that I rate Alonso higher than Raikkonen. Also, remember Fisichella was often called 'a good driver of bad cars and a bad driver of good cars' so his time at Renault were not his best seasons.
Graham Hill - most mathematical models rate him a lot lower than Schumacher. I've reduced the gap between the two because I don't think these are perfect, but I think they are accurate enough to mean that Hill cannot be on par with Stewart. Stewart was a rookie when they were teammates, while Hill was already an experienced driver.
Romain Grosjean/Lando Norris/Lance Stroll - here, it is just that I think the current crop of drivers are the best there have ever been in F1. There are far more other series and junior formulae now, which means to make it to the top level you have to be better than you had to be in the past. Norris, I think, could one day be one of the greats and I actually would now move him higher than that. Grosjean made a lot of mistakes but was extremely quick, look at his time at Lotus. Stroll may not really be good enough for today's F1, but he is still better than certain drivers from the past. Rodriguez could have been a bit higher, maybe, but I don't think he was better than Norris. And it is only a F1 driver list. Mark Donohue I already explained, but he could have been a bit higher. Phil Hill is rated lowly because F1metrics calculates the Ferrari in 1961 to be the most dominant car of all time, so he should have beaten the likes of Clark by a greater amount.
Enrique Bernoldi - matched up well to Verstappen as teammates, despite being a rookie.
Alex Rossi - to be fair, I'd forgotten he was on the list. With only five races, he probably shouldn't have been there.
Luigi Villoresi - he is at position 103.