Audience Prize - Discussion

Lessons learned from the Wallace

The voting procedure must be dead simple. As in:

1) Go to audienceprize.com. 2) Click on your favourites. 3) Click "submit".

So, this app must be modified so that if there is only one active competition, we skip the opening screen and make audienceprize.com open directly on the vote page for that active competition. Having repertoire and bios available is great, but it's a "nice to have". The first thing is to get the vote. Move along from there, if the voter wants more information.

Ranking candidates and a Condorcet vote calculation is far superior to single vote. This should be clear to anyone that lives in a democracy. The beauty of single vote lies only in its simplicity, and in this case I think the extra complication (choose more than 1, ranked in order) is justified. Differences between condorcet rankings and single-vote rankings can be dramatic.

Implicit with the simplicity principle is the need to not have any sort of login or further data collection criteria. If you want to collect more data, they must be very, very few data points, and answerable by a tap/click (no drop down select boxes or type in responses such as name and email). My questions about "How many competitors did you hear in Round 1?" needed tap buttons rather than dropdowns.

This app allows for many votes to be configured: Overall Audience Prize, Best NZ Composition, Best Classical Sonata, etc. That's great, but violates the simplicity principle unfortunately. Private votes (restricted to an invited pool of voters) can use them, but public ones probably should not.

If you feel the need to "filter" votes (say, to prohibit "bots" from voting), it needs to be done on the back end. It will be far too complicated to verify an email address, or try to stop people from voting multiple times from the same device or IP address. You won't get votes otherwise. Furthermore, you can't prohibit multiple votes from the same device or IP: Multiple people in the same household (ie on the same wifi hub) present from the same IP address. Parents who bring young children to view the competition (something we should encourage) hand their phone to the child to vote, and this will appear from the same device/IP address. All of which need to be allowed. The basic principle is: Collect as many votes if you can, and quietly omit some from the analysis if you need to.

To simplify things, submit the vote, then clear the vote and get it ready for the next vote (the app now does this). This will allow a volunteer (or volunteers) with an iPad at the venue to wander around afterwards collecting votes. That allows this web app to be used as an electronic collection method for "in hall only" votes too. Storing the rankings locally up until the vote occurs is good, however, since it allows voters the ability to rank as they go along.

You cannot make the analysis public. You do not want to show which candidates are at the bottom of a Condorcet ranking, for example, without removing the names of candidates and replacing with B, C, D, etc. The winner's name can be public, and possibly up to the number of candidates that advanced to the next round, but even then I would not put them in order or give any indication as to the strength of the win. And I would only make public N top names if N is small relative to the number of competitors (as in less than 40%).

If you wished to restrict it to in-hall voting, but still wanted to allow Internet voting for ease of collection (mobile phones), the website could advise the voter that the site needs access to their location to confirm that they're actually on-site. The voter would tap on "Allow", at which point the device would send its GPS coordinates to the server, where they could be denied a vote if they weren't within 500m of the event, let's say. This may lose a few people that think that's too invasive, but could be a valid filtering mechanism. It's getting more and more common nowadays. However, in my opinion, this violates the "simplicity" priniciple.