[D]Manual TG Queue suggestion

[Wymondham]

This was mentioned in an NSGP VC as a potential solution to the large region's dominance of the manual TG queue, and Elu suggested I make a thread here for it. The idea was to prioritise newer manual messages in the queue, while still retaining the the current system for messages the stamps & API queue, feedback would be appreciated.

[Comfed]

Could you elaborate on how exactly would this work?

[Eluvatar]

Not a detailed explanation but an attempt at a basic summary.

Currently, not only does the site randomly select from manual, api, or stamp queues to draw recruitment messages from with a 1/3 chance of processing a particular queue (or 50% if one of the others is empty, etc) but within each such queue it's equally likely to select any of the queued telegrams. Whether you just messaged a brand new nation or you queued a manual telegram yesterday it will be treated the same.

The concept is to prefer more recently queued telegrams in the case of manual telegrams. This would mean that sending a few telegrams once in a while would be more effective than sending the same number of telegrams in one go, and if you manually messaged a brand new nation you'd have a better chance of reaching their inbox quickly.

As I write this reply, I realize there's a very simple way to implement this. I'd just want to make sure to collect some statistics to do a good before/after comparison and to get some input from the community, mediated as necessary by our development manager Katganistan.

PS: Wymondham could you please also post a link to this topic in the Communications tech development thread?

[Wymondham]

Not a detailed explanation but an attempt at a basic summary.

Currently, not only does the site randomly select from manual, api, or stamp queues to draw recruitment messages from with a 1/3 chance of processing a particular queue (or 50% if one of the others is empty, etc) but within each such queue it's equally likely to select any of the queued telegrams. Whether you just messaged a brand new nation or you queued a manual telegram yesterday it will be treated the same.

The concept is to prefer more recently queued telegrams in the case of manual telegrams. This would mean that sending a few telegrams once in a while would be more effective than sending the same number of telegrams in one go, and if you manually messaged a brand new nation you'd have a better chance of reaching their inbox quickly.

As I write this reply, I realize there's a very simple way to implement this. I'd just want to make sure to collect some statistics to do a good before/after comparison and to get some input from the community, mediated as necessary by our development manager Katganistan.

PS: Wymondham could you please also post a link to this topic in the Communications tech development thread?

Done. Also, thank you for expanding more on this than my tired brain could manage at 11PM at night.

[Comfed]

That sounds smart. If it's so easy to implement maybe try it for a little while and see if it makes a difference?

[Sedgistan]

Full support from me.

I'd also like to see manual TGs get more of the first few slots (and API ones less), if that were simple to implement at the same time.

[Wymondham]

Full support from me.

I'd also like to see manual TGs get more of the first few slots (and API ones less), if that were simple to implement at the same time.

I think this is an excellent idea, maybe change it to 2/5s Manual, 1/3 stamps and 4/15s API?

[Bassiliya]

I think that stamps should be given lesser priority than API, like 2/5s manual, 1/3 API, 4/15 stamps. But either way, great idea if it could be implemented!

[Wymondham]

I think that stamps should be given lesser priority than API, like 2/5s manual, 1/3 API, 4/15 stamps. But either way, great idea if it could be implemented!

The issue around deprioritising stamps is that they need to be a realistic recruitment option so as to ensure people use them so as to yknow, actually fund the site. A solution that may be worth discussing in the long term is changing the API rate limit of recruitment TGs to say, 150 seconds, or 120 seconds. Although idk what Elu and Kat think about that?

If that is a proposal people are interested in, then I'm more than willing to make a new thread about it so we can move forward with this change while also discussing a change to API recruitment?

[Unibot III]

Do you want priority though?

Just noting that telegrams are read in reverse chronological order. Like is this going to result in manual TGs being buried?

You might want to consider a system where manual tgs are temporarily “pinned” to the top of the tg box until you check off the tg’s pin.

[All Wild Things]

Can you confirm that this would only impact TGs tagged as recruitment? It would be incredibly frustrating to have a regular TG conversation held up.

[Haganham]

Can you confirm that this would only impact TGs tagged as recruitment? It would be incredibly frustrating to have a regular TG conversation held up.

how many non recruitment telegrams are people receiving that it would be an issue?

[Eluvatar]

Can you confirm that this would only impact TGs tagged as recruitment? It would be incredibly frustrating to have a regular TG conversation held up.

Only recruitment telegrams get delayed in a queue system, yes.

Do you want priority though?

Just noting that telegrams are read in reverse chronological order. Like is this going to result in manual TGs being buried?

You might want to consider a system where manual tgs are temporarily “pinned” to the top of the tg box until you check off the tg’s pin.

I would prefer discussion of how telegrams are displayed go in a different topic. I wouldn't want to combine it with this change.

[The Seeker of Power]

I personally like and support this idea!

[All Wild Things]

Can you confirm that this would only impact TGs tagged as recruitment? It would be incredibly frustrating to have a regular TG conversation held up.

Only recruitment telegrams get delayed in a queue system, yes.

Thanks!

[Flanderlion]

Agree manual should be prioritised over the other two types. Stamps should always go ahead of API though, as one has no benefit to the site/isn't work and is just leaving a script running, while the other directly financially benefits the site.

Rate limit for API, happy with status quo, if you want to send faster via API it should cost stamps/put it in stamp queue. Also stamp price should be upped. But both are for another topic.

[Unibot III]

I would prefer discussion of how telegrams are displayed go in a different topic. I wouldn't want to combine it with this change.

I can understand that Elu but isn’t the motivation of this change to ensure players see manual tgs first? That’s the goal right?

I was just wondering whether this change might accidentally bury manual tgs? Because you’re reading the tg you receive last, first. Does that make sense?

I suppose it depends on how much time passes for an average new user before they open their tgs.

[Eluvatar]

The goal isn't necessarily to make nations see a manually sent TG first, but to level the manual TG playing field a little between large regions and smaller ones.

[Katganistan]

Seems like an idea that's got some support, so I will take it to the devs for sure.

[Galiantus III]

Do you want priority though?

Yes. 100%. When I was recruiting, I've seen telegrams sent within 5 seconds after founding spend 30+ minutes in queue, only to be blocked and never get delivered. If the options are: A) be the first telegram in the inbox and only be seen if the player opens their telegrams, or B) never make it into the inbox due to RNG; I think manual recruiters would choose option A.

[Wymondham]

Was this ever implemented?