Art? It's not a science? No, game selection is not a science. If game selection was a science then I would have to stick exclusively to objective measures, and that would make a very lopsided games list. Here are some examples:
1. The games that raise the most money.
The series that raise the most money are Mario, Zelda, Metroid, and Final Fantasy/Square-Enix RPGs. There are a couple of other games that raise a good amount of money too (such as ID and Valve first-person shooters and Super Meat Boy), but selecting games based on how much they raise would still cut out a lot of variety. You wouldn't see many 8-bit nes games (RIP Sunsoft), any obscure underrated gems and/or cult classics, and you would never see other popular series such as Metal Gear Solid, Sonic, or Pokémon (Pokémon is popular with viewers but generally doesn't raise as much money as the four series I mentioned).
Now there is no denying that the four series I mentioned see lot of representation in the Games Done Quick (GDQ) marathons, but what if they were the only games being represented at the marathons? That would kill a lot of the variety we have in marathons, and would prevent unexpected successes like Mega Mari from happening. I like Final Fantasy games and even I don't want to see more than one per GDQ.
2. The games that have the highest viewer counts.
Mario, Zelda, and Metroid would still be on this list, but you would take out Final Fantasy and put in Pokémon and 3D Grand Theft Auto games instead. This runs into the same problem focusing on the games that raise the most money does, which is lack of variety.
Also, while people love watching Pokémon and 3D Grand Theft Auto games, I think watching more than one would wear on people because the games in the respective series are similar.
This is not to say I don't want games that raise money or get lots of viewers, but if those were my only two metrics for determining games in the GDQs, then that would be a very narrow vision that only includes a few fanbases. Imagine if I denied a game because it had less than 20,000 viewers, ignoring factors such as the day the game was played (generally games later in the marathon get more viewers), and time of day (there's generally a viewer dip during graveyard shifts).
But let's say I use these two metrics and the next AGDQ is a roaring success, it raised way more money than we ever expected, record viewer counts, etc. That's all well and good, right? Actually, there's a new problem, how would I follow up?
I have now cornered myself by only including our biggest titles, and while I can switch around the categories for each run, that only works for so long. Do I repeat all of the successful games that worked before and risk stagnation? Or do I add in something else for variety's sake that might alienate the fanbase?
I personally think if I deviated from the variety currently in Games Done Quick marathons, that would be a big risk. Is Awful Games Done Quick ever going to raise the most money? No, but it's a fun way to kill a graveyard shift and people in the community enjoy it. Do Zelda games raise some of the most money in marathons? Yes, but making it the sole focus would shut out a lot of other games. While I cannot let in every game each speedrunning community wants, if I only focused on a select few series or genres, then I think I would end up alienating the majority of speedrunning communities.
Now you might be wondering, well if game selection is not solely about the most popular and lucrative games, then what is it about? That's what I'm going to discuss in future blog posts.
P.S. By the way, for those curious, here the donation statistics for SGDQ 2013 and AGDQ 2013. In case if you want to see how games fared in terms of donations.