Wellington Scoop

Andy vs Justin: Part Two (with AI)

by Ian Apperley
If you are reading these articles, then you are a political junkie on some level, have an interest in local politics, and are far more likely to vote on what you research. You’re more likely to educate yourself on candidates, what they stand for, and then make an informed decision to vote.

Unfortunately, you’re probably in the less than one per cent of people who do that. Voters generally do no research, have little idea of what is going on, and make a millisecond snap judgement to vote on a face, and not much else.

Over the last week I’ve been asking people, almost randomly, “who are you going to vote for mayor?” The majority of people can’t name a candidate, and quite a few had no idea there was even an election underway. Older people, who read the newspaper, had more precise ideas about who to vote for but often it was based on incorrect facts. For example: “I’m not voting for that candidate because he’s stopped Shelly Bay from going ahead” or “I’m not voting for that candidate because they stuffed the buses up.”

Democracy in action… Do I even need to roll the old quote out? Yes, yes, I do.

“The best argument against democracy is a five-minute conversation with the average voter.”
– Winston Churchill

Numerous studies have shown that a candidate’s face is what wins them votes. That a person will make a snap decision in milliseconds when they see the face, no doubt a leftover part of our primal brain that would need to make a threat assessment each time they saw another human.

We look for a couple of things. In general, the first is trustworthiness, and the second is competence. Now, you can tell your brain it’s illogical to look at someone’s face and be able to determine that in a second, without research, but it makes no difference, your unconscious brain does it every time.

I spoke to a lot of people about Andy and Justin’s faces. The general consensus from the political junkies was that Justin had a “better face.” Couple that with name recognition, also another lever in how we vote, and he’d get back in as mayor.

I spent quite a lot of time looking at various images of Andy and Justin and my gut response, in the end, was that I quite liked Andy’s face over Justin’s. So then I got to thinking. Could there be a more scientific way of analysing both candidates’ faces to understand how people saw them?

Absolutely not, but in the interests of an opinion piece I’ve never let that stop me before, so I used Artificial Intelligence to analyse them.

I took the first five images of Andy and Justin, and I put them through this website that claims to have an algorithm that examines the face and labels it accordingly with a series of words that it thinks are related.

Here are a couple of examples.


AI thinks Andy is a Sociologist.


AI picks Justin straight away…Isn’t that interesting?

Honestly, there are hours of fun in this. At one stage, the AI labelled one of the candidates “first offender”, which caused me great entertainment. No, I’m not saying which, that’s not cricket.

Seriously, try it for yourself. It’s not infallible. I put my picture in, and it declared me a “beard.”

Based on my unscientific analysis of feeding images to a notional “AI,” I reached the following conclusion: Andy’s face is more attractive to a voter than Justin, by a slim margin. Generally, AI labelled Justin a little more harshly while Andy ended up with more “professional” tags like “cardiac surgeon”, for example.

So I conclude that Andy’s face is likely to score more votes than Justin’s. And given that his face is plastered super-size in some key locations – god, you can’t miss him – then that bodes well for his chances.

Both candidates have decent name recognition though Justin’s is higher, given he’s the Mayor and has been for some time, an upside of being the incumbent. Andy has had a significant boost via the Peter Jackson effect, which could be a double-edged sword, but will lift his votes across the city. Peter Jackson, who has more than two million social media followers, has undoubtedly boosted Andy’s chances.

Right now, I think that Andy could just tip Justin out, and that’s an instinctual call. A call for battening down the hatches a bit, more focus on short term issues, a little more city focus, which feeling of austerity, thumbing his nose at central government somewhat, could just put him in the lead.

With the campaign still running for a few more weeks, it leaves both candidates needing a strategy.

Justin has seemed relatively comfortable sitting back and going on the campaign trail ride without much fanfare. It may be that he has a “set and forget” approach to the rest of the campaign hoping that his record will get him over the line again. It’s not a bad strategy.

Andy has been quite active and is still regularly starting policy debate through media releases. His billboards are certainly the most prominent around the city, and that will be making a difference. He’s the chaser in this trail and needs to work hard to stay above the press parapet.

Frankly, there is little either candidate could do to spoil their chances. Most of the mud was thrown in the early days, and none of it stuck, We are well past that point now. Short of some major disaster, both candidates are likely to run to the end.

Both candidates will lose some votes respectively to the other mayoral candidates. At the most risk of this is Justin because he has a couple of candidates sitting on his political spectrum that people may jump to. Those being Conor Hill and Jenny Condie.

Andy is likely to pick up the votes that fell last time to Nick Leggett and Jo Coughlan; he’s the logical replacement for those voters. He might get a few taken away from him by Diane Calvert, but I doubt that it will make a difference. Those centre-right votes are quite substantial, and they are looking for a home.

The other, risky strategy that Andy could take is dragging Justin into debates he doesn’t want to be in. What really is going on at Shelly Bay? What about that airport extension? How much is the Convention Centre really going to cost? What is happening with Civic Square? What really happened with LGWM?

But that’s pointless, for two reasons. First, unless that spills over into mainstream media, then it’s a valueless exercise. Second, Andy has been a Councillor for many a long year and is involved in all of those contentious areas.

Justin needs to focus on the message that if we stick to the plan, keep moving forward, and cherry-pick opportunities, then we’ll be ok on the other side. The long term plan took time to develop, went through consultation, and is two years into execution.

Andy needs to disrupt that plan with alternatives. Forgoing LGWM to some extent in favour of something else, more life in the CBD including green space, more focus on the environment, and more of a, dare I say, “let’s make Wellington great again” style of approach, playing off the fact we are a little rusty right now.

Neither candidate (nor any of the others) has come up with a killer policy or slogan. There is still time, about a week, to do that. So I’m hedging my bets right now but tipping Andy to get over the line before Justin, if nothing changes.

I should warn you. I’m terrible at betting. I am explicitly forbidden to wager on New Zealand in the Rugby World Cup. Indeed, the tournament will be forefront in people’s minds.

Freakishly, rugby does affect elections; however, in this case, voting will be done and dusted before the final unless New Zealand go out in the first pool. Then Andy will definitely win.

First published by Inside Wellington.


  1. Stephen Todd, 21. September 2019, 9:05

    “Both candidates will lose some votes respectively to the other mayoral candidates that are running.”

    Both candidates will lose some *first-preference* votes, respectively, to other candidates. Assuming that, as you predict, Andy and Justin will be the last two candidates remaining in the count, they will get most of those votes back again as the other candidates are progressively eliminated.

    Generally speaking, the votes of the Left(-leaning) voters will accumulate upon the most-preferred candidate of the Left, and the votes of the Right(-leaning) voters will accumulate upon the most-preferred candidate of the Right, to elect one of them.

    At the end of the day, what matters is how many single votes cast by individual Wellingtonians end up with Justin, compared to how many such votes end up with Andy (or, if you’re wrong and Andy is eliminated during the count, with Diane), and *that* will depend on who votes.

    Having said all that, just a reminder to voters: second- and later preferences can never be used to defeat your first- or earlier preferences, so it is perfectly safe to rank-order some or all the candidates, as you see fit, in your own personal order of preference.

  2. Ian Apperley, 21. September 2019, 11:17

    As always, excellent advice Mr Todd. Thank you.

  3. Andy Circus, 21. September 2019, 13:41

    Peter Jackson could get WETA to fashion a prosthetic face for Andy based on the ideal for voters!

  4. Andy Foster, 23. September 2019, 9:10

    My policies would require a 3 to 4% annual rates increase. That would still mean cuts in some projects.
    The Mt Vic tunnel is State highway #1 and the Government would pay for the second tunnel. I expect some negotiation with Govt would be required to put it up the priority list. [via twitter]

  5. BHS, 23. September 2019, 9:31

    I’ve posted my vote in and it was for Don Newt McDonald who is the only candidate wanting to clean up our dirty streets and stop growth ‘Growth is Bad 4 WGT’. Go NEWT!

  6. Donald T., 23. September 2019, 9:42

    Just toll the tunnel and make users pay. I don’t want to pay anymore for accommodating more people in Wellington. Wellington is full!

  7. Andrew S, 23. September 2019, 12:19

    Doesn’t he who count the votes count?
    The whole thing smells like it has already been decided (by those who choose the agenda and who gets to debate on RNZ and who doesn’t) who gets in.

  8. Dave B, 23. September 2019, 12:45

    We don’t need another Mt Vic Tunnel unless it’s a rail tunnel. We need to extend the regional rail system to the eastern suburbs to provide unbroken rail-connectivity to the city and rest-of-the-region, such that traffic through the Mt Vic Tunnel reduces to only that which needs to be there. By duplicating the road tunnel and failing to extend the rail spine that the rest of the region benefits from, we will simply encourage more traffic that shouldn’t be there, and worsen the city’s traffic-problem overall.

    The present government has freed-up NZTA’s previous funding-strictures such that government-funding of public-transport projects such as this can now be considered.

    Andy Foster, Diane Calvert and other 4-lanes-to-the planers, can’t you see this? Don’t even think about more roads. Get us out of the car-dependency pit, don’t try to dig us even further into it.

  9. Henry Filth, 23. September 2019, 17:30

    Is extending rail to the airport possible, or just too insanely expensive? You could get off the plane from Sydney and be in Taihape before you knew it!

  10. Roy Kutel, 23. September 2019, 18:39

    Yes Henry – it is possible to extend rail to the airport – Paraparaumu airport, that is. Forget Wellington Airport, it will be under water before LGWM has even finished its d(r)aft report.

  11. Dave B, 23. September 2019, 20:58

    @ Henry Filth. Extending rail to the airport is no more difficult than building a motorway to the airport, except that the latter was given a big head-start in the 1970s when the existing motorway was built whereas nothing more was done with rail. Rail was also proposed for extension at the time but the plan was foolishly canned. We should be very wary of continuing to rule-out what really needs to be done, just because a dumb decision to take rail no further was made in the 1970s.
    The same situation pertained in Auckland, but at last they are now biting the bullet and building the city rail link.

  12. Guy M, 23. September 2019, 21:01

    Dave B – you’ve got it spot on – 100% absolutely right with your comment: “We don’t need another Mt Vic Tunnel unless it’s a rail tunnel.”

    That’s the plain and simple truth.

  13. Henry Filth, 25. September 2019, 21:26

    Thanks for the history Dave.
    Roy, could the existing Kapiti rail infrastructure cope with a Stockholm/Arlana or Oslo/Gardemoen scenario?

  14. Roy Kutel, 26. September 2019, 22:51

    Henry F – yes.

  15. Mike Mellor, 27. September 2019, 8:27

    Roy K: Wellington airport is 8-13m above sea level and Paraparaumu 7m, so which one will be underwater first?

    Kapiti’s rail infrastructure would find it hard to cope with the increase in frequency implied by the addition of a dedicated airport service. Apart from the obvious need for a link between the station and the airport, the steep single-track multi-tunnel stretch of line between Pukerua Bay and Paekakariki has long been identified as a significant restriction on capacity.

  16. Roy Kutel, 27. September 2019, 9:45

    Err Mike – Wellington’s runway is in the sea!

  17. Dave B, 27. September 2019, 14:01

    But Mike, if rail could get its hands on Transmission-Gully-type funding then the single-track section could be eliminated and fares could be reduced to zero (in keeping with Transmission Gully being offered toll-free to users).
    Let’s break out of the thinking that says roads take the whole cake while rail just gets the crumbs.

  18. Mike Mellor, 27. September 2019, 21:23

    Roy K: the north end of Wellington’s runway is at the top of a substantial embankment, the south end on top of a vehicle tunnel. In both cases it’s higher above sea level than Paraparaumu’s, so if, as you say, “it will be under water before LGWM has even finished its d(r)aft report”, your preferred Paraparaumu option would be even further underwater.

    Dave B: agreed. I was just responding to Roy K’s assertion that Kapiti’s existing rail infrastructure could cope with scenarios like the rail links to Oslo and Stockholm airports. This is not the case: these links handle six trains per hour at 200km/h, in both respects roughly double the existing Kapiti infrastructure’s capabilities.

  19. Jono, 27. September 2019, 22:51

    Wellington runway would be an island surrounded by water, so instead of rail or ‘highway’ to the Wellington airport (both of which are idiotic non starter ideas anyway), we had better start planning for a ferry service.

  20. Madeleine Simpson, 28. September 2019, 7:00

    Yeah Jono if they are saying the sea is rising they can’t act, build and plan as if it’s not.

  21. Guy M, 28. September 2019, 7:33

    It’s worth noting also that Auckland Airport’s runway is just over 2m above sea level, and so if there is any airport that is most likely to be affected by sea level rise, it will be Auckland – long before Wellington…

    You’ve seen those pictures of Tamaki Drive with waves washing across it in a storm? Now imagine that at Auckland Airport… (mind you, if we get to that stage, Lambton Quay will be looking like Venice at Aqua Alto…. gumboots on for your next visit to David Jones…)