Stellaris Dev Diary #212: PDXCon Remixed and 3.0.3

[MrGuyPerson]

I won't lie; I was hoping there would be some updates on the longer-term direction for the growth mechanics and other tweaks in the works, but good news that the patch is fully released.

 

[johnty5]

Does this mean there's going to be some kind of Stellaris content announcement at the show?

 

[DrNukeLear]

It's announcements like these which constitute #NotADevelopmentDiary.

The first is that 3.0.3 has moved from beta to live on all platforms, with the full details available here. 3.0.3 includes numerous economic and AI adjustments, bugfixes, and two new sliders in galaxy generation to control pop growth behavior without the need for a mod. There have been a few more fixes added since the last beta update, so check out the patch notes.

That part might potentially be a dev diary.

Will continue to use my mod to remove the new processing overhead of the pop growth behaviour system which the sliders leave in place.

Nice to see the other fixes. Enjoy PDXCon!

 

[Silens]

I was hoping for a "real" Dev Diary that gives me an excuse not to work for the next hour. Oh well, tomorrow then!

 

[Byrios]

The link to the full schedule is broken for me. It just says “(insert link)”.

edit: refreshed and the link is there that sends me to the PDXCON Remixed page. But the “(insert link)” is still there.

 

[Helicopter200]

i really hope for primitive DLC announcement

 

[guiskj]

I'm confused. What was the point of the AI Feedback threads when none of the reported bugs and deficiencies were fixed during the Beta?

 

[sillyrobot]

I'm confused. What was the point of the AI Feedback threads when none of the reported bugs and deficiencies were fixed during the Beta?

Well, it gave those who wanted to donate time to a for-profit company something to do. Also, they can still use those reports as they slowly move forward either redeveloping the AI or getting someone to develop enough expertise in the AI to build out changes. Ultimately, they should at least identify unit tests / performance tests / benchmarks to apply.

 

[johnty5]

I'm confused. What was the point of the AI Feedback threads when none of the reported bugs and deficiencies were fixed during the Beta?

They explained up-thread - to collect information for further AI changes after this patch.

 

[guiskj]

They explained up-thread - to collect information for further AI changes after this patch.

Then why make it a part of the Beta? That directly implies that the feedback will impact said Beta. It was either careless or disingenuous of them to phrase it in terms of the Beta. Either way, it is rather infuriating.

 

[johnty5]

Then why make it a part of the Beta? That directly implies that the feedback will impact said Beta. It was either careless or disingenuous of them to phrase it in terms of the Beta. Either way, it is rather infuriating.

Well clearly the beta was to do three things:

- test the pop mechanic adjustments with a playerbase bigger than they could do internally before a full release
- test the bugfixes (in case they lead to other bugs as sometimes happen) before a full release
- to collect feedback on the AI to inform AI changes in forthcoming patches

 

[Alfray Stryke]

Well clearly the beta was to do three things:

- test the pop mechanic adjustments with a playerbase bigger than they could do internally before a full release
- test the bugfixes (in case they lead to other bugs as sometimes happen) before a full release
- to collect feedback on the AI to inform AI changes in forthcoming patches

Yup, we've got some really good feedback from the open beta and it's informed a lot of decisions we're making. Just need more time to implement changes, test them, refine them, test them again, etc...

 

[guiskj]

[snip]
- to collect feedback on the AI to inform AI changes in forthcoming patches

You don't need a Beta branch for that. All the problems in 3.0.3 were in 3.0.2 and even 3.0.1 when it comes to AI.

So many of us have been waiting for so long for PDX to do something meaningful with the AI, then 3.0 comes and makes the damn thing worse than ever. Pop Growth gets the limelight while all around me I just see the AI collapse worse than ever.

So when PDX makes the most direct effort towards AI in quite some time by asking for feedback and then just moves on and says 3.0.3 is the last of the 3.0.* patches, it makes me enraged.

At the very minimum, they had to jump in those threads and make a few statements or make a meaningful statement in the 3.0.3 release.

 

[johnty5]

You don't need a Beta branch for that. All the problems in 3.0.3 were in 3.0.2 and even 3.0.1 when it comes to AI.

So many of us have been waiting for so long for PDX to do something meaningful with the AI, then 3.0 comes and makes the damn thing worse than ever. Pop Growth gets the limelight while all around me I just see the AI collapse worse than ever.

So when PDX makes the most direct effort towards AI in quite some time by asking for feedback and then just moves on and says 3.0.3 is the last of the 3.0.* patches, it makes me enraged.

At the very minimum, they had to jump in those threads and make a few statements or make a meaningful statement in the 3.0.3 release.

If they're asking a group of players (likely players with a lot of expertise, given those are the ones of the forums/likely to opt-in to a beta) for feedback on pop mechanics and bugfixes, why not also use that period to collect feedback on the AI too? It makes sense to roll those two things together while everyone's in "feedback mode" of running stuff in observer mode, playing games and taking feedback notes etc.

 

[sillyrobot]

If they're asking a group of players (likely players with a lot of expertise, given those are the ones of the forums/likely to opt-in to a beta) for feedback on pop mechanics and bugfixes, why not also use that period to collect feedback on the AI too? It makes sense to roll those two things together while everyone's in "feedback mode" of running stuff in observer mode, playing games and taking feedback notes etc.

Because you distract your testers and dilute feedback that you can use for iterative change in the beta. you also limit your pool of respondents unnecessarily. Not everyone has access to the beta (or wants to use a beta version) and had the call gone out to all users of 3.0.X, a larger pool of respondents would be available. It also sets up expectations by inference that simply were not true.

 

[johnty5]

Because you distract your testers and dilute feedback that you can use for iterative change in the beta. you also limit your pool of respondents unnecessarily. Not everyone has access to the beta (or wants to use a beta version) and had the call gone out to all users of 3.0.X, a larger pool of respondents would be available. It also sets up expectations by inference that simply were not true.

That's certainly a hypothesis about how a beta would work. Though, I don't think it's borne out by how the beta turned out. Two AI feedback/discussion threads - one running to five pages and one running to eight - seems like a decent amount of feedback to me. The QA comments since certainly suggest that they're pleased with the feedback they got.

 

[sillyrobot]

That's certainly a hypothesis about how a beta would work. Though, I don't think it's borne out by how the beta turned out. Two AI feedback/discussion threads - one running to five pages and one running to eight - seems like a decent amount of feedback to me. The QA comments since certainly suggest that they're pleased with the feedback they got.

Not really. Answer to the first set of questions (how is growth doing) effectively stopped when the second set of questions (how is AI doing) came out. Had they performed an iterative change to 3.0.3 and received more feedback, they had another week-long opprtunity for further refinement - and it does need it with a whole bunch of additional edge cases identified.

Also you didn't address (a) that it restricted the pool of respondents (I for one, do not participate in betas for for-profit companies unless they are paying me) or (b) that it set up inferred expectations of follow-up because it was presented in the beta.

 

[johnty5]

Not really. Answer to the first set of questions (how is growth doing) effectively stopped when the second set of questions (how is AI doing) came out. Had they performed an iterative change to 3.0.3 and received more feedback, they had another week-long opprtunity for further refinement - and it does need it with a whole bunch of additional edge cases identified.

What kind of interative change?

Also you didn't address (a) that it restricted the pool of respondents (I for one, do not participate in betas for for-profit companies unless they are paying me) or

I'm confused. Do you think betas are wrong (because beta are companies exploiting players by requesting feedback without offering payment)?

(b) that it set up inferred expectations of follow-up because it was presented in the beta.

People can mistakenly infer things. If they infer something that's not explicitly stated - and which doesn't subsequently turn out to be accurate - they should reflect on how they can make better inferrences in future, rather than lashing out.

 

[Nyrael]

Then why make it a part of the Beta? That directly implies that the feedback will impact said Beta. It was either careless or disingenuous of them to phrase it in terms of the Beta. Either way, it is rather infuriating.

Because it helps identify the problem sooner, which makes it likelier to get a fix sooner rather than later, and may help identifying other issues sooner rather than later. It's naive to expect that all non-critical reported bugs and issues will be solved soon after being reported. No software development works that way, non-critical issues often get fixed in the phase after they were discovered, unless all previous bugs and issues have been taken care of (which is extremely unlikely).