Quick Knowledge Share Post:
I was just revisiting the greenpill episode that I did with Professor Bryan Ford (who has been working on sybil resistence for 18 years).
At this timestamp I asked him about anti sybil aggregators (Passport). In his answer, Professor Bryan Ford makes a couple great points about designing aggregating scoring algorithms.
TLDR:
- Professor Bryan Ford is against relying whollly on mechanisms that are fungible with money (eg stake X tokens get $X in cost of forgery) bc they are exclusionary to ppl without $$$.
- a holy grail might be to find mechanisms that have a high cost of forgery for attackers but are marginally difficult for actual ppl
- Professor Bryan Ford advises that hybrid mechanisms can have problematic attributes depending on whether you use an AND/OR gate on them.
- (fungible with money mechanism) OR (non fungible with money mechanism) = (fungible with money mechanism)
- (fungible with money mechanism) AND (non fungible with money mechanism) = ( NON fungible with money mechanism)
A few recommendations:
- In the docs for the scoring algorithm, it might be worth expanding on these best practices.
- The whole episode is worth its weight in gold for ppl designing passport IMO. But this timestamp covers the point above. If you’re working on Passport, it might be worth listening!
Disclaimer: This post is for informative purposes only and is not financial advice. This post reflects my personal views, not a decision or commitment by Gitcoin governance. Forward-looking items in it are targets, not commitments. The information in these posts is subject to change as we continue learning. This post may contain estimates, may contain errors, and is provided on a best-effort basis. DYOR, do not make any financial decisions based on these posts.