This is super interesting! I like the choices you’ve provided but I’m not entirely sure how they map to the roll and how would audience actually take the roll. In particular if people simply shout/message 1-10, it’s statistically unlikely that you’ll get 10 given that at least some people in the audience are going to experiment and potentially select 1, some people will genuinely be interested in low scores (and select one) and people are bound to not be so polarized to just select 10s and 1s (which could save the end-result) and pick something inbetween.
So I suspect that unless you sell the participation as something that has to be pro/against players, you’ll always end up in 6-7 range. If that’s what you want, its totally fine, but with time participation may withe. I also don’t get who/how picks one of the choices for Journey Forth (I admit I’m not sure how much the Embark and Journey Forth is different in your proposed setup either). If GM picks one of the choices from the list above then participation seems superficial to me (same goes for Pokedex one). Given that Embark and Journey Forth would probably be the first participation audience will be involved in, you may end up discouraging them or showing that this participation is meaningless.
Conversely Wild Encounter participation seems like a great idea. Unless there are more than a few dozens of people in the audience (in which case you’re probably bound to hit “my vote doesn’t count” attitude) this is going to be a high stakes, engaging and meaningful mechanic.
So to answer your questions directly, here’s my take:
- Convention panel that’s sub 100 in the audience and smaller Twitch (i.e. sub 100) streams should work great with raffle tickets. It feels important, high stakes and direct.
- Larger audience is probably better suited for a poll-like experience with predefined set of outcomes. So, say, voting on choices should work (think: Who Wants to Be a Millionaire? Ask the Audience lifeline) but using people as a dice probably wouldn’t.
- For smaller audience you could use counting randomizer (that’s probably used by children around the world) as a dice but I’m not sure it’s superior to raffle.
- As an audience I’d be upset if my choice was ultimately up to GM’s interpretation.
- Raffle could probably work for large audiences if it were triggered many times over a span of a game (more often than once per 100 audience members per hour would be my initial estimate) and this was known ahead of time.
And no, I haven’t seen anything like this before but I love it.
 In Poland we have this counting game: everyone picks a number from 1 to 5 and participants show a number of their choosing using their fingers when triggered (typically ref shouting “now”). Then you spell out, syllable by syllable, a short rhyme, counting syllables over a randomized set (typically participants). As long as there are participants around or below the range one selects (1-5) it feels like you have some control over outcome. Probably simpler to show on an example:
Rhyme: “(1)ci-(2)sza (3)w^koś-(4)cie-(5)le (6)ci-(7)sza (8)w^kom-(9)na-(10)cie, (11)kto (12)się (13)o-(14)dez-(15)wie (16)zdej-(17)mu-(18)je (19)ga-(20)cie” (silence in church, silence in hall, who says a word takes off it all – I’m paraphrasing)
So by spelling out a rhyme you basically count to 20. Most kids know that and if the purpose of this is to select someone who does something dangerous, they try to 1) guess what other would select, 2) pick something that puts 20 as far from them as possible. So if you have 4 participants and you’re the first one you probably assume 2.5 as average and want to either go low (1) or high (5) so id it comes back to you it’s either sub-10 or high above 10 and you can evade the responsibility. There’s a variant where you start counting from zero (heavens), exchange church for sea and call this a sailor game (not to be confused with another sailor game which is about drawing matches). I ramble, I apologize.