Subathon - suggestions

[woeful herald]

1 - Streaming hours per day

a) The currently proposed ratio of 10h online and 14h of sleeping is something that’s hard to agree with. During the last 3 months Vedal has streamed for 166h* which means that if he takes January, February and half of December*2 off it could be translated to about 150h of Neuro content. 150h equals 15 days of the currently proposed subathon, which is a letdown since it seems that going over 2 weeks is something that Vedal already considers to be good enough while in reality only after passing this mark we’re gaining extra content.

In this scenario - anyone who gifts subs to raise the timer will pay for hours that we would have gotten anyway.

*https://twitchtracker.com/vedal987
*2 December isn’t really off for him, but I count it as such due to the break in streams

b) 14h of sleeping can not only be considered a scam (I predict a lot of that in chat when the timer goes down when Neuro sleeps for hours), but it’s also detracting for new people who may tune in to check what’s going on during other parts of the day, only to always see that Neuro’s asleep (we already had those last year).

Easy fix that doesn’t require much of Vedal’s input:

• If minecraft will be ready on time, simply add a few hours of it to the daily schedule to extend it (other games are fine as well).
• Just chatting with either Neuro or Evil is still a lot better than watching her sleep. Especially since it has improved a lot compared to last year.
• There’s a lot of art / art from various contests that has yet to be seen by Neuro, it could provide content for hours.

A normal streamer’s ratio during a subathon is around 8-10h off and 14-16 on. Neuro’s different of course, but having at least a 12h/12h ratio is a lot more agreeable.

2 - Subathon overall length

While I understand Vedal’s constant worry about not having enough content, he misses the point of the entire Subathon if he focuses on it too much. One of the main motivations behind people gifting subs is to see the timer never go down.

Worrying about running out of content after the 2 weeks mark is pointless, since even the lack of content is something that the watchers are looking forward to, it often leads to crazy and interesting ideas and pushes Vedal to try things he would have never done otherwise. For example, him playing the cookie clicker last time was something that would have never happened otherwise, and most importantly it’s something that’s memorable and can bring a smile to the face when you reminiscent about it a year or two later - the entire end of last year's subathon was glorious.

This year Neuro has many abilities that she lacked last time, simply letting her play her ever growing library of games or call friends on discord can be used to fill out time.

Like Vedal said, a lot of people don’t have the time to watch the entire thing and only want to tune in occasionally. And it’s perfectly fine to focus the best content ideas in a 3-4h daily segment, so they can participate in the most interesting parts. However, it doesn’t mean that the rest has to be cut out and forcibly reduced.

3 - Time per sub

Some stats first:

During the last subathon around 41k subs were gifted and it lasted around 17 days.

41 000 (subs) x 30 (sec) = 1 230 000 (sec) : 60 = 20 500 (min) : 60 = ~341,6 (hours) : 24 = ~14,2 (days)

As we can see the subs alone wouldn’t be able to fuel the entire previous subathon even if they kept being 30 sec per sub, which they didn’t. The extra time has been gained through bits and higher tiered subs.

Of course, Neuro grew bigger and we can expect more
this year. Let’s say the number of gifted subs will reach 100k.

100 000 (subs) x 30 (sec) = 3 000 000 (sec) : 60 = 50 000 (min) : 60 = ~833 (hours) : 24 = ~34,7 (days)

Almost 35 days is a lot. Add to this bits and higher tiered subs and it will start looking really grim for Vedal.

However, putting a hard cap on the subathon kills half of the joy and anticipation connected with it. The same can be said if the soft cap is too harsh (Vedal has also mentioned this during his dev stream).

Soft cap can also be done in a really bad way. For example, reducing the time per sub upon reaching every new subgoal is like telling people that the more they sub the less time they will receive and will discourage a lot of them.

Vedal’s idea of cutting the time by 2 sec per day was a good one. Even better was the one with the gambling wheel either adding or subtracting time, since it introduces an extra source of enjoyment and something to look forward to.

Below is a proposition of how to combine both ideas, assuming that this year subathon will reach at least 75k gifted subs in 15 days:

Starting time per sub: 25 sec

Time always reduced at the end of the day (necessary to counterbalance the situation in which the wheel hits more of the positive numbers first): 2 sec

The wheel contains:

2 x +3 sec 2 x -3 sec
4 x +2 sec 4 x -2 sec
4 x +1 sec 4 x -1 sec

Total: +18 sec Total: -18 sec

Number of spins required to hit every outcome (each is removed after being hit once) - 20

The number of subs gifted per day is really hard to predict, it’s certain that it will be higher near the start and the end, but I’ve opted to keep it simple in the examples.

.

In the situation where the wheel is rigged in favour of the viewers the subathon ends after about 22 days (not counting the extra time gained from bits and higher tier subs).

In this situation the subathon ends after about 6 days. Of course this is the absolute worst case scenario that doesn’t include the time gained from bits, higher tier subs, and the will of the viewers to raise the timer no matter what.
All in all, the wheel idea is really fun and can be easily adjusted to obtain more desirable results. Adding to it a flat -2sec time reduction every day makes it more manageable for Vedal.

4 - Add a new outfit as a subgoal

I will keep this point simple. Neuro’s wardrobe is really lacking compared to other big vtubers. The outfit subgoal from the previous subathon has yet to be delivered, but that doesn’t mean we shouldn’t be promised a new one. It is a long term goal that can be fulfilled months after the delivery of the cyberpunk princess outfit, and it really should be included.

Ending note

I love Neuro, and I really appreciate everything that Vedal has done. He’s working hard to make and provide top quality content. However, I feel that sometimes he is too focused on the wrong things, and I hope this post will give him some food for thought.

[languid sluice]

whoaaa this is a lot, the points make sense though! Man you must've put some work into this.

[gray nest]

This is a really interesting post, I appreciate how much effort you put into it.

The subathon length is a really tricky topic. On one hand I agree with you that it should be longer if the break after will be as long as it was last year. But on the other hand I completely understand Vedal's concerns with the quality of the content and the sleep schedule, seeing how last year has turned out.

Let me preface my take by saying that this is obviously his life and his personal choice how he wants to manage his breaks. I'm just throwing the stuff below out here as an opinion.

If I had to add anything to that topic, then I feel like a big problem with it is that it feels like all those subs are getting wasted because of the break. Any new viewers getting a sub won't have a reason to keep it going cause there is no streams. No much reason to get into the fanbase too when everybody will tell them to wait 2.5 months for a next stream. I think the ideal scenario would be a shorter subathon with a shorter break and then another break later in the year. This way we wouldn't need to worry about running out of content on the subathon, the sleep schedule wouldn't be completely destroyed, the potential new viewers will be more likely to stick around and hardcore viewers won't go completely insane because of the lack of the twins.

I also agree with your last point. It doesn't have to be an outfit but it really feels like we need some more BIG goals that will have an impact on the streams. The current goals feel more fit for a more focused shorter subathon, as I was proposing earlier. If this subathon is supposed to be a bigger and better then it really feels like there should be something more bigger that the community can work towards.

Great post, I hope Vedal reads through your suggestions and maybe considers mine.

[ebon roost]

regarding streaming hours per day:

a)
I sort of eyeballed 10h online and 14h sleeping thinking that was more or less what I did last year but maybe I'm wrong here, should probably check. will definitely be thinking about this one further and if you have any ideas here that's useful

b)
calling it a scam seems a bit harsh, I know other people in subathons have only done up to 10 hours per day though they didn't have their timer run down at night
we could ofc do this too, but then we'd just have fewer seconds per sub so I think it just kind of equates to roughly the same thing here, unless most people would prefer to just be offline at night(?)

also worth noting I'm already trying to account for including mc/just chatting

[rapid chasm]

I think a really fun way to make the random time per sub thing work is how I saw Pointcrow do it during their subathon where each individual batch of subs was adding random time per sub, ranging from 0-60 seconds per sub and it would apply to all the subs gifted at the same time so 5 gifted is 0-60 per sub times 5, not new each sub, this lead to some pretty fun high and low rolls with big sub dumps. all values were weighted equally and for the soft cap they reduced the upper limit every week after the second I think it was.

[rapid chasm]

when back to check and I was wrong, indevidual subs were randomized even in batches apart from the rare rolls being the 0 second roll which was around one in 512 roughly and there was also an equally rare legendary roll where you get 5 minutes per sub on all of them

[woeful herald]

It was the other way around - in the https://www.youtube.com/@Neuro-samaUnofficialVODs/videos channel, the parts where Neuro is sleeping are separated from the rest and it's easy to see how long those segments were, for example:

20 Dec - 8h 53 min https://www.youtube.com/watch?v=YRW5SbZx6k0 +1h 50 min from previous VOD = 10h 43 min
25 Dec - 9h 35 min https://www.youtube.com/watch?v=Vx1GNpJ9h9g + 2h from the previous VOD = 11h 35 min
27 Dec - 9h 28 min https://www.youtube.com/watch?v=SKG9m8vsjkQ +1h 20 min from the previous VOD +4h from the next VOD = 14h 48 min

And so on, basically around 11-12h of sleeping with 12-13h of uptime (with the rare exception when the sleeping portion was much longer than usual for some reason). Just to make sure I've also checked one where sleeping is mixed with the content, and made sure the subathon timer is correct so nothing is cut out:

23 Dec - Neuro goes to sleep in the middle of the VOD https://www.youtube.com/watch?v=WUu8Mk1zmh4 - around 5h of sleeping
24 Dec - Neuro starts asleep and wakes up in the middle https://www.youtube.com/watch?v=bP-N7Ffud-Q - around 5h of sleeping

In the example above the sleeping portion is about 10h long. That's why I feel that reversing those proportions and making her always sleep for 14h is a bad move, especially since you're aiming for this year to be better than the last. However, it's also true that the sleeping parts started being longer and longer during the later days of the subathon, like in here:

1 Jan - 8h 44min of sleeping https://www.youtube.com/watch?v=jnbGhXaqgik
1 Jan - 3h 40min of sleeping https://www.youtube.com/watch?v=FiuTGCU9QcE

Which makes it 12h 22min of sleeping since you most likely started running out of steam at this point.

I'm cheering on you to do even better this year!

[woeful herald]

Turning off the stream for the night would rob you of the title of number 1 female twitch streamer, and us from a rare chance to farm channel points.

Jokes aside, why not have a canvas, but one that's only active when Neuro's asleep? That would give people something to do, but at the same it wouldn't distract anyone from watching Neuro when she's awake.

[ebon anvil]

If you're gonna do 10 hours active, and if you're somewhat confident on handling an automated stream, I would suggest adding a "middle of the night" automated stream half way through sleeping for like 2 hours for the bad timezones that fall within the 14 hours (pretty big range). It would at least break up the 14 hours into 6+2+6

[woeful herald]

I can't speak for everyone, but all I need to be happy is a schedule similar to this:

Example day 1
0h - Wakes up! Just chatting with Neuro ~2h
2h - Neuro reviews art or plays a game (maybe she will finally beat Slay The Spire?) ~2h
4h - Switch to minecraft and play it for around ~2h
6h - Planned content ~4h
10h - Just chatting with Evil ~2h
12h - The twins go to sleep

Example day 2
0h - Wakes up! A morning discord call to a friend ~1-3h (if it’s shorter than 3h than the rest is just chatting)
3h - Minecraft ~3h
6h - Planned content, maybe completing a subgoal already? 4-6h
10-12h - Call a friend in discord and talk until it’s time to sleep.

Example day 3
0h - Wakes up! Both twins chat for some time ~3h
3h - The twins play chess ~2h
5h - Planned content / subgoals / maybe a guest appearance? ~5h
10h - Minecraft ~2-3h
13h - Time to sleep

Example day 4
0h - Wakes up! A morning art review and some games, buckshot, liar’s bar with Evil or friend ~2h
2h - Just chatting ~2h
4h - Find a babysitter for Neuro, (check if anyone friendly is streaming and call them asking to take care of Neuro for some time) ~2-3h
5h - Planned content the serious content hours ~5h
10h - Evil/Neuro plays Inscryption ~2h
12h - Time to sleep

Example day 5
0h - Wakes up! It’s minecraft day, see if anyone can join the server and play with Neuro ~4h
4h - Just chatting with the twins ~2h
6h - Planned content, complete more subgoals ~4h
10h - Prank calls on discord with Neuro ~2-3h
13h - Time to sleep

Add to this the planned Game Jam, and all the things I forgot to include and there really is enough content to make the subathon longer. Last year was fun, though a bit monotone, this year Neuro has a much wider choice of activities and a lot more friends to play with (babysit her).

By planned content I meant activities in which Vedal actively participates. There's nothing wrong with some padding and fillers in between more interesting sections, it's still better than Neuro being asleep. It's also worth noting that a single appearance of Koko during a just chatting stream has made it last for almost 5h. The same happens when Anny joins.

[deft fiber]

In my opinion, unless it is certain that there will be enough content, it's best to make capped subathon. I don't know why so many people are against the idea. And it'll be easier for Vedal, he won't have to think about what to do in 10 days. Extra bonus, you don't have to think about how many seconds subs should add. You expect 15-20 days, plan activities for every day. You set the number of subs necessary for maximum duration and you can run the subathon with peace of mind

[shy sandal]

normally i try to give pros and cons of stuff i read
but thats kinda hard when i agree with every word here

[woeful herald]

• Capped subathon takes away from the thrill of not knowing when it will end, it also directly translated to the number of gifted subs as people are less inclined to gift knowing that it won't change much since obtaining the number required to reach the cap is pretty much guaranteed.
• There's also the aspect of wanting to see Vedal (and the viewers) slowly fall into madness as the subathon becomes longer and longer.
• Reaching the end of prepared content is actually good for creativity as it pushes the streamer to look outside his comfort zone, which directly translated to a higher chance of a masterpiece being born.

That's pretty much how every other subathon works, I would even say that Neuro's is on the lighter side compared to some which can last for months. We also can't forget about the planned long break after, it's much easier to come to terms with it as a Neuro watcher after experiencing the wild ride that is a subathon. And honestly, the 20 days you've used as an example is about as long as I expect the subathon to last anyway even if it's uncapped. The difference is that if people will try hard enough we can add a day or two to it (and a lot of cash to Vedal's wallet). The more important factor is the number of hours when Neuro will be actively streaming per day. I don't expect to reach Ironmouse levels as she can stream for 16h (with some short breaks) a day, but there's a big difference between 10 and 12.

[ebon anvil]

For a capped subathon, the goals need to be really good for any incentive

Cookie clicker was the second highest viewed stream of all time, first being 2.0 debut

[noble tusk]

I need cookie clicker part 2

[deft fiber]

I've never noticed Vedal's drive to get as many subscriptions as possible. It's not his goal.
What kind of madness are we talking about? Watching the suffering of a man who wants it to end?
What kind of creativity are you talking about? Almost all people watch videos, play games, and just chat. They don't do anything “creative” beyond their goals
So far it looks like the subathon must be uncapped because everyone else is doing it this way

[shy sandal]

its hard to explain what made it so fun but that last "hour" was like a battle against the timer and if it doesnt happen again it'll be a little disappointing

a capped subathon could never have something like that

[deft fiber]

I didn't watch the subathon last year for the “thrill” or “madness” I watched for the content. And it's much easier to plan and create content when you have a clear timeline in place

[shy sandal]

the thrill and the madness was part of the content

[deft fiber]

For you. Not everyone agrees

[woeful herald]

A little bit offtopic discussion is fine, but this is a subathon - suggetions topic, please stick to it.

[deft fiber]

Idea. If you like having a timer on the last seconds, Vedal could have a segment at the end where people can add seconds on top of the limit. This would also be good because most people would know when the subathon would end and when the segment would start. This will keep that feeling you were talking about + people won't miss the end and can participate in extending the timer. And at the same time, Vedal will have a clear timeline that he can plan for

[ebon anvil]

Cutting the subathon entirely and reducing the timer to 1 second for the last hour is completely different. Even with 1 second timers it was more hype than if it just ended.

[shy sandal]

im not against that idea but i guess i kinda dont see the point. it could confuse viewers by having 2 timers and he could still plan out content for the expected amount of time and move to unplanned content after he ran out, and nothing would be lost. a similar idea that would be less confusing for the viewers is possibly having a subathon schedule with a date for the end of planned content

[deft fiber]

Oh and I'll repeat my suggestion again here. Ask Twitch for the main page for the first 2 hours (or however long they give you). This will help bring in more viewers, and a great chance to break the viewership record

[valid helm]

I've been thinking about the wheel for a bit now, and have come to the same conclusion every time: having the result of the wheel being added to current timer has a pretty high chance of resulting in infinite stream, giving a few "+x seconds" wheel rolls in a row.

either need a probability professor here or
I'd suggest looking into each wheel result saying exactly what the "sec/sub" for the next hour will be, instead being a delta to the current "sec/sub"

e.g
day 1:
option 1 = 30 sec/sub for the next hour; option 2 = 25 sec/sub for the next hour; option 3 = 20 sec/sub for the next hour; etc.

reduce each option by, for example, 3 seconds every 2 days but not below 1 sec/sub (those need to be calculated properly for sure)

Its a less exiting wheel for sure, but far safer and abuse-proven imo.

Also the longer subathon goes the more exiting it would be for viewers to get anything different than "1 second/sub" and be able to push the timer(stream) for a bit longer.

[sage orchid]

I was playing around with a Monte Carlo simulation of the wheel situation, with the caveat that I've never done this before so I could well be doing this horribly wrong.

Apart from the wheel itself, how people will sub is obviously also a factor. I was mostly modelling it as an exponential decay on the assumption that people will gradually get less motivated, but in reality people would probably be more excited to sub if the value was higher. Removing any large positive values after they were hit could help limit that sort of behaviour, I guess.

An equally-weighted wheel where it has an equal chance of going up and down by the same amounts, if people subbed at a consistent rate, would certainly have an unacceptable chance of running crazily long or forever. With exponential decay in interest you can get some more reasonable probabilities, although the RNG also means the stream could be over rather quickly too - in one simulation it could be as short as a 5-day subathon.

So, uh, tl;dr is this is kind of fun to play with but it would definitely need some safeguards. I'm interested to see how it would work if the big movements (especially increases) are one-shot.

[willow storm]

Did you take into account that more content ≠ better content

[boreal latch]

Tbf subathons aren’t usually content focused it’s just the streamer living but on stream

[woeful herald]

Yes, I did:

"Like Vedal said, a lot of people don’t have the time to watch the entire thing and only want to tune in occasionally. And it’s perfectly fine to focus the best content ideas in a 3-4h daily segment, so they can participate in the most interesting parts. However, it doesn’t mean that the rest has to be cut out and forcibly reduced."

I'm not advocating for the increase in the most dense content-wise segment (though it would be nice if possible), but for the change in the ratio of standard Neuro content compared to the time she spends sleeping. And since it's Neuro we're talking about, a certain level of quality is guaranteed. Also, as I've stated before a few times, this year's Neuro is a lot better and has more options available to her than the previous year's one, it would be strange to reduce her time to shine. Vedal has stated that he wants this subathon to be better than last year's, and the 14h off / 10h on ratio doesn't achieve this.

There's also the factor that some of Neuro's friends are in the American / Asian timezones, and extending the just chatting segments (either at the start or the end of each streaming day) gives them an extra chance of being able to talk or play with Neuro during hours that are more friendly to them.

[fiery obsidian]

Since people are concerned about the wheel idea having the potential to cause an infinite subathon - is there a world where the options on the wheel change each day based on 1) the current time/sub and 2) the current length of the subathon? (i.e the wheel day one has almost all of its options be negative, but if the time/sub gets too low too quickly the wheel options are more positive. Then it could go back down as the subathon gets longer). That way you could still get "jackpot days" while retaining the ability to bring the subathon to a close if needed.

[solid fjord]

I really like the idea of having planned content always starting around the same time each day. Since the start time he suggested in the dev stream is at 1am for me and I can't realistically wake up that early every single day, it'd be reassuring to know that the potentially most exciting stuff would happen at the time that most regular viewers can make it.

(secretly hoping that his sleep schedule drifts as much as it did last year though because those stream times were literally perfect for me)

[austere oracle]

I wonder if there will be any plans with automating the VOD making process? 24-48h VODs was quite lengthy last year and our youtuber Neuro-archiver is capped at 12h. Perhaps making shorter vods with the category change so the vod watchers have something to catch up on?

[woeful herald]

A very good point, splitting the VODs into smaller parts and making them public right after Neuro goes to sleep could help everyone who wants to catch up on the daily subathon events. A lot better than waiting for the 48h chunks that appear randomly and that can be hard to navigate to the exact point you've missed. The change to and from 'just sleeping' category could be the trigger if such automation is possible.

[ebon roost]

good to know, didn't intend to drift from this too much

[void loom]

Some sort of chat game or activity would be cool for at night to replace the canvas.

[smoky bronze]

Instead of the canvas maybe a community Minecraft server when neuro is asleep? something like the server is only up when neuro is asleep. To make it more fun you could make it so you can only place like a block a minute or something like the canvas or every player has a dedicated like 4 chunks to build whatever they want or something else and you can then look at it when neuro wakes up free content. Admittedly tho its a lot of Minecraft in the subathon.

[jagged gyro]

this is genuinely really cool data to read through, good work

[dapper anvil]

this sounds pretty cool actually, tho idk if there's enough time to cook this

[surreal lotus]

Another subathon logistics thing: It would be a good idea to announce the timeframe for the game jam before the subathon to give people time to prepare more.

[low pike]

I mean if that mod exists already for block limitations that could make it a bit easier. It would also need auto chunk assignment to each player.

Also hosting for what will likely be a lot of concurrent users will be expensive or resource intensive so that's a huge limiting factor. (Thanks JAVA)

Still that's a really solid and cool idea. It really puts a new spin on the canvas idea.

[surreal lotus]

Make sure Neuro has permissions to grief things

[indigo lynx]

People say they love "cookie clicker" -> Let's combine that with the subathon itself and implement "Golden Sub" time. Randomly a golden neuro will appear and for the next XX seconds all subs will count like 7 times the current timer increase.

Maybe to balance it have it only work for single gifted subs.

people can only sub so fast after all.

[slim jungle]

maybe make space for extra intermediate subathon goals that will get created or not made at all on the fly. I'm sure the audience/you will come up with some great ideas that could be subgoals

[surreal lotus]

Neuro plays Jump King (since you don't have to act in real time, you could use commands like "move a bit right", charge jump x% right")

[slim jungle]

(note you'd still need vision or you'd need to tell neuro where the platform is)

[surreal lotus]

When Vedal inevitably runs out of content, he should play Glittermitten Grove (it's not much but it's better than Cookie Clicker)

[slim jungle]

i was going to suggest universal paperclips, but it looks like GG has a lot more of variety

[south hazel]

Fan-art run during night (not reaction replays, just all fanart ever made on the channel). A collection of Neuro animations run at night with permission of the makers.

[valid geyser]

Sorry if I'm necroing this thread or if this has been said before. Tried skimming all the comments but might have missed some.

About the wheel and a way to avoid things like people just waiting around for a better spin or pumping the timer to weeks/instantly ending stream at a high/low roll:
Make the wheel re-spin at every X amount of subs. This should make it much easier to balance the wheel as you can guarantee a max amount of timer change from each wheel spin. Also stops the incentive to just wait around for better spins and instead encourages pushing it to another spin to roll for a better outcome.

Optionally add some bad/good luck protection by locking repeat rolls, increasing the odds of rolls which hasn't been hit in a while, etc. Now you should be able to create a relatively accurate estimate for time streamed per amount of subs and then balance that to whatever fits best for the desired stream duration and estimated total subs. I'm too stupid to do those estimates/calculations though so I don't have any neat graphs to show here.

[dark drum]

Give Neuro a programming project of her choice as a subgoal then Beatle can critique and provide guidance

it can be inside a VM surely she won't find a way to escape