[MENTION=11243]ZoboCamel[/MENTION] stellar analysis per usual. Thank you for it. Follow-up question: what nominal ratio would they have needed for a 50/50 jury/televote split?
Based on how things have played out in the last few years? They'd need a nominal split of about 3-to-1, or 75-25, i.e. tripling all of the televote scores
This year, for example, tripling all of the televote scores would've meant left them as:
1. Benjamin 201
2. Felix 192
3. John 186
4. Mendez 186
5. S&V 180
6. Liamoo 159
7. Renaida 153
8. Rolandz 153
9. Mariette 147
10. Margaret 123
11. Martin 123
12. Jessica 111
Leaving the total combined results as:
Benjamin 315
Felix 286
John 252
S&V 234
Liamoo 211
Mariette 211
Mendez 188
Margaret 185
Renaida 183
Rolandz 177
Martin 166
Jessica 144
Here, there's a standard deviation total of 28.7 among the televotes vs. 29.2 for the jury votes, so about even (with the juries still actually being a little higher there despite the televotes being tripled). The overall results are a little more in line with what you'd expect, with the placings of most acts falling vaguely halfway between their jury and televote rankings, and each of the two groups having some things go their way and others not so much. For example, the televotes succeed in making Samir & Viktor's 4th place safer, lifting Rolandz above Martin and putting Jessica last in line with their preferences, but they still don't completely overrule the juries, who manage to push Mendez from a televoting 3rd down to an overall 7th, bring Mariette up from 9th in the televotes to 6th overall, etc.
In other metrics, this tripling would mean the televotes had 90 points between their first and last places (vs. 112 for the juries), and 42 points between 1st and 5th (vs. 52 points for the juries). Which seems about fair, considering that the televoters didn't seem to have a massive favourite in this year anyway.
The juries perhaps remain more powerful even with the tripling, since they still have the impact of influencing televotes via their own score reveals, but if scores are tripled and voters have more power to overturn jury results then they could potentially have more confidence to vote for their favourites anyway, somewhat nullifying the effect of the jury reveal. This is just about impossible to calculate, though - it's more of a psychology thing - so I'll ignore that and say that, yes, tripling the televotes would have made 2018 just about fair.
Applying the same tripling effect to 2017's results would have made things much closer, with Nano barely scraping out a win: he'd have had 247 points over Robin's 246, with Jon & Aninia on 203. This seems fair and reasonable enough - among the two rankings, Nano had a first (televote) and a second (jury), which sounds like it should take slight precedence over Robin's first (jury) and third (televote). Metrics like standard deviations (28.7 jury vs. 27.3 televote) and the first-to-last split (95 vs. 81) would have also been much fairer under the tripled system, as opposed to those televote figures being a third of that with what actually happened.
And then, for 2016, we'd have had Frans with an easy win, sitting at 292 points over Oscar's second-place 218 and Wiktoria's third-place 204. Once again, that looks about right: the juries had Frans and Oscar essentially tied (89-88), while Frans was a huge televote favourite and just about national phenomenon for that year. And the metrics check out here too, i.e. standard deviations (32.7 jury vs. 29.7 televote) and first-to-last split (88 jury vs. 120 televote).
TL;DR: SVT/Bjorkman, if you're looking for an easy way to fix this imbalance, just have televoters award three times as many points as they currently do. It's a bit of a messy, band-aid solution, but it would've given us three years in a row of a fair and representative system while keeping the app votes etc. in their current form.
Otherwise, there are lots of other, more elegant solutions available, each with their ups and downs. For example, you could:
Split the phone-televotes and app-televotes into two sections, comprising the final score of 50% juries, 25% app and 25% televotes. Spread the app votes out by converting them to another system with the same ranking (e.g. the 12-10-8-7-etc.) and you'd end with a fair score that incorporates the app vote while being equally representative of the televote and the juries.
Remove the app vote from the final entirely while keeping it in the semis, if need be. This would keep the majority of app engagement while returning votes in the finals to their state from 2015 and beforehand.
Rework the app system - rather than score every song out of five, give voters e.g. 10 free votes to award across the entire semi or final.
Limit the number of app votes to, say, one per song, rather than 5. This wouldn't entirely solve the issue, but it'd fix things about 85% of the way off the top of my head, returning the split to something like 45-55 in the favour of juries, which is at least acceptable.
There are a bunch more solutions out there, I'm sure, but that's just what I could come up with on a whim. The entirety of this post took me, a single random fan, something like 40 minutes before having to rush off for dinner; I'd be rather amazed if SVT, with all of its resources, couldn't manage something at least on that level
