BTW @lukeflegg I don't think Nationwide have ever done business banking, but they are clearly considering it - I've had a couple of surveys from them recently that would suggest as much anyway...

@jdaviescoates we can introduce the ability to configure the min and max, but wanted to wait until it was requested. Even better, until someone submits a PR for it.

Hi Rob, does "submit a PR" in this context mean creating an Issue here? https://github.com/loomio/loomio/issues

[Edit: I realized later that you probably mean pull request. I need to slow down sometimes in my commenting.]

That's exciting. Yes I think they used to offer one but stopped like 7 years ago. I can't remember why. We went with Metrobank in the end for Dignity platform

Well... no, sorry. There are a lot of myths and misinformation out there about voting systems. You seem to have run into quite a few of them. Since there are others reading this I'll take some time to put context around the links in my response, but you should familiarize yourself with their content if you'd like to discuss further.

Range ('score') voting is as good or better than both IRV (what folks here have been calling "ranked choice") and Plurality in every case in terms of the quality of results. It's both superior to IRV as well as superior to Plurality. (In fact, the plurality system is so bad that (computer simulations indicate) replacing it with range voting would improve society by a comparable or greater amount than the entire invention of democracy in the first place.)

this is precisely why ranked voting exists, and I think that ranked voting is deeply preferable to plurality voting

You are generally correct that IRV is intended as an improvement over Plurality, although in fact it's frequently not - since they share many of the same pathologies, as the city of Burlington learned the hard way when they were among the first to try it. (They have since repealed IRV.) Unfortunately, IRV's problems are hidden under its complexity, whereas Plurality's flaws are simpler, familiar, and obvious to voters who can take them into account - meaning socially at least, IRV is in fact still worse than plurality.

On the other hand, Range Voting exhibits no pathological behavior, ever. The results may be less than perfect - but never completely wrong (as in the Burlington case, as well as others such as Peru 2006). Plurality and IRV can't even make it past Arrow's theorem of impossibility, making it mathematically impossible for them to always have sensible results.

Hopefully that covers your comment of

I'd disagree that ranked choice is inherently pathological, although it has a limited domain of effective usability.

Yes, I'd say that domain is quite limited. Namely, limited to elections you don't care about. You seem to want to do the opposite:

I'd certainly use ranked choice instead of score polling to make any official collective decisions.

Yikes. Why use the broken system precisely where good results are needed most? Why introduce the spoiler effect and pick the wrong person for your next secretary? Why allow the chance that you could be making the results worse for yourself just by voting?

However, it holds immense potentials for general co-sensing and co-design process, so I'm really glad that Loomio has added it.

Loomio is a perfect place to pick the right one and stick with it, focusing developer efforts and minimizing complexity. Not only is range voting simpler than IRV, but it can in fact be used as if it were other systems - want to vote Approval-style? Simply vote maximum and minimum ratings only. Want Plurality for some reason? Vote minimum for all but one.

Score polling is deeply vulnerable to insincere tactical voting if it's used in official decision process. However, it holds immense potentials for[..]

Your "however" implies that Range has somehow been disqualified from serious voting due to your previous statement. This is a fallacy. Indeed, more expressive systems (range being the most expressive) allow more room for dishonest voting (by nature). If you really want to limit dishonest voting, you could simply take away some expressiveness - try flattening range to approval voting, and viola, goodbye 'rampant' dishonest voting. The cost? results quality! (though still better than plurality and IRV!) See the graphic attached. Why forfeit that quality prematurely though? You could allow the dishonest votes on the more accurate system and the output will suffer, yet will still be better than any other systems. Do we give a higher priority to making it hard to lie, or to getting the most representative results regardless? It's good to have your priorities straight when filing objections.

(To see the above visually, see the image.)

That said, it's understandable to look for voter honesty. But the Gibbard-Satterthwaite theorem shows that no rank-order ballot voting method – not IRV, not Borda, not Condorcet, not any of the infinite number of rank-order voting methods that nobody has even invented yet – can give you full honesty - for all of them, that feat is impossible. With IRV or any such voting method, there are elections in which the only way for you to get a good election result, is to flat-out lie by pretending A > B when your honest belief about that candidate pair is A < B.

Indeed, the better way to achieve a gamed voting system is... to use anything worse than Range. For more on this, check out the book "Gaming the Vote" (NYT review) by William Poundstone.

This is a tremendously important subject – it is not often-enough realized how important. In that vein, I therefore re-emphasize that we should remove "Ranked Choice" from the decision type options. We just might leave our future users quite pleasantly surprised.

I just noticed this implementation is missing one critical feature - the ability to opt-out of rating a particular option at all. At least, it's not clear that this is possible. There are a ton of reasons why this is a necessary aspect of range voting, and forcing zeroes in the place of "blanks" is giving us biased results. Hope to see this fixed soon.

In the future, it would also be a nice feature to allow voters to rank their choices - for those who prefer to just drag-and-drop in order of preference. This would then be reflected automatically in the ratings - distributing their scores equally across the range in the declared order. :)

Yeah, there a lots of numbering schemes with built-in biases, even the 0-100 scale (since it's used so heavily in school, where more than the whole bottom half of the scale was simply "fail") or the 5-star concept (online product ratings). Using negative minimums is fun, but it turns out you can actually solve this problem while also increasing the speed at which people vote by simply not using numbers at all.

This has the added benefit of allowing infinite precision. You don't think about numbers - just position the cursor where it feels right, and click. To see it in action, try out the smartphone app UpVote. It's pretty cool!

Thanks for responding in such depth, @adroit . I have mixed feelings. All the hyperlinks above go to the same rangevoting website, which presumably exists purely to promote range voting. You're clearly an enthusiastic advocate and you're welcome to advocate it. However, that website looks to me like lots of theory with thought experiment examples, but practically no use in real serious elections. If it's deeply sound theory which you want people to trust enough to commit serious elections to, then I'd like to see links to it in peer-reviewed professional journals-- probably mathematics journals-- or heck, on Quora, or at least edit the wikipedia page for Score Voting and see how well that holds up. Better yet, of course, get it used in small but serious elections, and build evidence of satisfied users.

I get that range voting is simpler and more expressive than alternatives, and I actually base some of my key decision process models on practically analogous ratings systems. (By contrast, typical political elections-- and ranked voting-- are tertiary subjects for me.) I don't currently understand why you suggest to focus intensely on the strategic voting risk in ranked voting, without discussing how strategies influence range voting. But hey I'm not here to argue against anyone trying ranged voting in any ways they desire! Loomio has it now (as Score Voting) and I hope you explore its performance in varied groups with varied goals. If its effective use-cases exceed my expectations, it'll make my work easier.

Anyway, this conversation reminds me that my main problem with range voting-- and ratings systems in general-- is not the simple and deeply expressive method of rating subjects according to one or more finely graduated dimensions such as Approve/Disapprove. My problem is that ranged results-- and official decisions, when applicable-- are usually determined by (1) 'averaging' by arithmetic mean and (2) ranking the averages. I'll try to make time to focus on alternatives to that procedure.