TL;DR Update on my thoughts: I’ve updated significantly downwards on the probability that trying to make the Meta Giving Season match go very well for EAs is worthwhile and I am not investigating it further, pending the EA GT Team’s reply to me (I emailed Philip some more information).
More details:
After getting some data on recurring Facebook donation timestamps, it appears to me that getting matched will likely be a random lottery for anyone who sets up their recurring donation within the first few hours of the match, only slightly weighted towards those who set up their donations early.
Specifically, the data suggests that the second donation that goes through on December 15th will go through not at the exact same time as the first donation on November 15th, but at a random time in a ~7 hour window (based on 11 data points). That’s quite a bit of variation, which means a donor who donates in the first second on Nov 15th can get beaten by someone who donates a few hours later.
(This assumes that 70,000 recurring $100/mo donations will be set up within the first ~7 hours of the match. Given that ~$150M was donated in a single day on Giving Tuesday in years past, and that $7M in donations was made in the first couple seconds last year, this seems quite plausible to me, though not guaranteed. If the matching funds actually last for much longer (e.g. a full day or longer), then a donor probably can get matched with high probability by donating right at the beginning of the match.)
So I don’t think I can be confident that a bunch of EA donors setting up their donations right when the match begins will almost all get their second donations matched, and because of that I think it’s probably not worthwhile to put in the effort to try to get ~$100k-$1M matched by EAs. The strategy to do so would involve a lot of donation trades and would take a lot of organizer time, plus be asking a lot from donors, so I wouldn’t want to do it unless there was a high chance that a high fraction of EAs’ donations would actually get matched.
TL;DR Update on my thoughts: I’ve updated significantly downwards on the probability that trying to make the Meta Giving Season match go very well for EAs is worthwhile and I am not investigating it further, pending the EA GT Team’s reply to me (I emailed Philip some more information).
More details:
After getting some data on recurring Facebook donation timestamps, it appears to me that getting matched will likely be a random lottery for anyone who sets up their recurring donation within the first few hours of the match, only slightly weighted towards those who set up their donations early.
Specifically, the data suggests that the second donation that goes through on December 15th will go through not at the exact same time as the first donation on November 15th, but at a random time in a ~7 hour window (based on 11 data points). That’s quite a bit of variation, which means a donor who donates in the first second on Nov 15th can get beaten by someone who donates a few hours later.
(This assumes that 70,000 recurring $100/mo donations will be set up within the first ~7 hours of the match. Given that ~$150M was donated in a single day on Giving Tuesday in years past, and that $7M in donations was made in the first couple seconds last year, this seems quite plausible to me, though not guaranteed. If the matching funds actually last for much longer (e.g. a full day or longer), then a donor probably can get matched with high probability by donating right at the beginning of the match.)
So I don’t think I can be confident that a bunch of EA donors setting up their donations right when the match begins will almost all get their second donations matched, and because of that I think it’s probably not worthwhile to put in the effort to try to get ~$100k-$1M matched by EAs. The strategy to do so would involve a lot of donation trades and would take a lot of organizer time, plus be asking a lot from donors, so I wouldn’t want to do it unless there was a high chance that a high fraction of EAs’ donations would actually get matched.