[Labs-l] Down to the millisecond

Platonides platonides at gmail.com
Fri May 27 21:35:24 UTC 2016


El 27/05/16 17:29, Huji Lee escribió:
> Okay let me explain: we want to create a tie-breaker mechanism for ties
> in elections. My idea is that if we have two people who tie, we will
> have our local election representative make a null edit in a designated
> page.

Terminology nitpick, what you need is a dummy edit, not a null edit
https://meta.wikimedia.org/wiki/Help:Dummy_edit#Null_edit


> Then we will use the right-most digit of the timestamp as a
> psuedorandom number, and use it to break the tie (if it is odd, the
> first person wins, if even the second person wins). The problem is you
> can plan your edit such that it'll happen on an even or odd second. It
> is hard, but possible.
>
> Rev ID is a good idea. Are there any other such pseudorandom numbers you
> can think of?

The revision id could be gamed as well by the local election representative.
You could use something like the stock value at closing that day of a 
given exchange.


On 27/05/16 17:56, Merlijn van Deen (valhallasw) wrote:
 >
 > http://cgi.cs.duke.edu/~des/vct/vct.cgi
 > <http://cgi.cs.duke.edu/%7Edes/vct/vct.cgi>
 >
 > There are probably a few other options online as well.

Interesting. random.org offers the same service, but it's a paid one
https://www.random.org/draws/


They wouldn't preclude the rogue organiser to request several coins 
until his desired result is chosen, though,


Additionally, there's the option to implement something like this 
ourselves (this is a trivial service).


Regards



More information about the Labs-l mailing list