Yesterday was 'A' level results day in England. 'A' levels (the 'A' is short for 'Advanced') are national exams taken at age 17 or 18, the results of which being those which universities look at when deciding whether to offer prospective students places. 'A' level results day is traditionally a big event each year - the opportunity for newspaper editors to publish lots of photographs of attractive young women jumping about excitedly at their results, and for newspaper columnists to complain about how the exams are getting easier and grades are becoming inflated.
The grade inflation narrative has been put on steroids this year, because owing to the lockdowns and school closures, there were no actual exams this year - instead, teachers awarded their students' results based on their 'predicted grades'. It turns out that asking teachers what grades their students should get is a bit like asking Nike whether its trainers are any good, or asking a Haagen Dazs spokesman if he thinks his company makes nice ice cream. Teachers - who would have thought it? - are of the opinion that they do a grand job and their students are all set to perform fabulously well. This year, over half of all 'A' level grades were either A or A*, the two highest grade boundaries.
It is a scandal, of course - and one that should be much bigger than it actually is. But it tells us a lot about human beings. When you don't have some kind of neutral mechanism devised to keep people honest, they generally end up behaving dishonestly.
In the case of 'A' levels, that neutral mechanism is the paper exam, externally marked (i.e., anonymously, by markers completely unconnected to the school at which the pupil is studying). An external marker has no skin in the game - it doesn't matter to him or her whether a particular student does well. So his mark is broadly trustworthy. Essentially the opposite is true of teachers marking their own students' work. If students get excellent grades it makes the teacher look good. It's not rocket science to see in which direction the incentives point.
In RPGs, the neutral mechanism in question is generally the dice. The dice, rolled openly so everyone can see the results, keeps everybody - particularly the DM - honest. If they go away, the DM follows his or her own predilections. Being nice to his best friend/girlfriend/person he secretly fancies. Making life awkward for the player he dislikes. Pursuing his own view on what the campaign's 'story' should be. Trying to expedite a scene so he can get the session finished and go to the pub. We're all familiar with those pushes and pulls. To guard against them influencing affairs at the table, we look to the dice - again, rolled openly - which are always unfailingly truthful. And thus we trust what is happening at the table.
Some people reading this will now, I predict, be thinking to themselves: "That's what he says. Whenever I play my games with my wonderful friends, we are all unfailingly honest and work in each other's interests so that we all enjoy the game equally, because we are such fabulous people." To which I can only respond: if it makes you feel better to think of things that way, go ahead and maintain the fantasy.