I hope somebody runs the data in the future to see if there was a statistical difference in outcomes between the companies intended to be accepted vs. not :-)
In the next round if they said they would accept anyone, then presumably there would be a lot of low quality applicants (100's of Facebook clones etc).
They know who they "would" have taken.
And somewhere in the future they know who was successful.
If that are the companies they wanted anyway, all is good.
If not, they can check what was different and what spoke for and against them
I wouldn't worry too much about that angle, personally.
The human mind can follow a community of about 150. This is two orders of magnitude larger.
I think they will be lucky if they can even remember names of folks they are dealing with.
Immediately start tracking all the applicants and what stages they are in, how much funding/traction/etc they have and which segment vertical they are in and then seeing how well they all do.
This will be a very interesting experiment.
Serious question. Is there anything I can start doing/capturing that would help?
/s - because much of the data you'd have on such early stage co's is identity-based, and therefor chock full of bias.
YC has always admirably stretched the scale of vc. This will be an interesting experiment now at this scale--and I'm especially curious how you'll measure progress along the way.