Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

How else would you define there to be only two winning numbers then? If you define the winning number afterwards, there’s no way telling how many numbers of those would’ve been in circulation.


Make the winner number outside the normal printed range.


This is so obvious - what are the downsides? I mean, it could be easily guessable that the bottlecap is valuable, ruining the surprise.

Ultimately they didn't test. Or confirm the vendor's test. It's all Pepsi's fault for a mission critical bug.


I mean...that's what they tried to do...


Perhaps they should have instructed the manufacturers to not include numbers which were multiples of, say, 15, but to use all other numbers uniformly.

Random audits (with expensive penalties) could have detected whether the manufacturers had complied with this, without making it too obvious to contestants that the winning number would be a multiple of 15.

The winning bottles could then be manufactured to contain some specific multiple of 15, while still allowing the organisers some freedom to choose which one.


Make it a non-number?


Use a CSPRNG to produce the tokens and keep a list of the outputs. Select uniformly/randomly from this list for payouts.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: