Partial password usability sucks

Partial passwords are a terrible idea for a multitude of reasons but one in particular stands out to me. We have come to the general conclusion that humans shouldn't remember their passwords, they should remember one complicated pass key that is linked to a number of otherwise inaccessible secret keys used to interact with servers but, let's step back for a moment and assume people are remembering their passwords (which most people not on HN still do)...

A password has an entropy n, a higher n yields more security but comes at a cost in terms of memory power, I think it's reasonable to state that very high entropy passwords exhaust our memory quite quickly which is why password reuse is so common, this particular authentication is requiring the user to have a high entropy password n, and then is reducing it to an entropy m where m is at most as large as n (usually much smaller) and using that for authentication.

We, users and tech people, don't care about the cost to transmit bits over the wire or the cost to verify those bits on the far end (in this scenario) we care about the cost on the user's memory more than anything so why are we optimizing this puzzle to be least efficient for our most valuable resource.

When I first read this article I thought this practice was silly, now it strikes me as stupid... it is a transformation that simply discards some of the potential security of the system for no reason.

(Also, banks love ease of use, which is why up here in Canada they love to send out password dongles to customers that make it even easier to log in to a site than this while still requiring possession of a physical thing - and many of those dongles require additional authentication to activate whenever using, though not all)

> If you think I'm wrong, then write it up in a spec and show us.

Sure. It's fun.

On password creation:

0.0. demand 25% longer passwords from users than considered safe

0.1. along with the hash of the whole password remember a randomly choosen 25% of the characters of the password and their positions (in clear text or reversibly encrypted)

On each login:

1.0. ask only for the characters in positions you haven't remembered

1.1. server-side fill the positions you remembered

1.2. hash the result and see if it matches

If hash matches:

2.0. randomly choose 1/3rd of the freshly provided positions and remember them in clear text on server replacing the previous ones

2.1. give access

If hash doesn't match:

3.0. deny access, don't change anything on server

If I understand correctly - to get the whole password keylogger must be specific to this bank site (to understand which characters are provided), and must be active on at least 4 logins. In my book that's a big improvement.

And I can imagine writing the login page in a way that makes parsing it in keylogger to understand which characters are masked.

So they are storing length and hashes of subsets of the password?

You could brute force a subset of 4-5 characters way easier than a 20 characters password.

How is this secure? Anyone care to explain how this adds security?

With a collection gutter at the bottom, from which your yearly interest will be derived.

For the context: partial passwords have been the default and a de facto standard for Polish banks since early 2000s. They are not "cutting edge novel idea" but more a common legacy.

I think people got used to them, so changing this now for everyone would feel weird (and would probably annoy many).

In some banks you can opt-out in settings and have regular password instead.

I guess it's an equivalent of swipe cards in USA.