• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Stellaris Dev Diary #161 - Development Update

Hello everyone!

For this week’s dev diary we chose to switch the order of a couple of dev diaries to be able to give you some updates earlier rather than later.

The Stellaris brand has understandably been under more scrutiny than usual for the last few months, and we want to address situations related to work-in-progress art in Federations. As an example, we had some UI design mockups (shown during PDXCon) that contained placeholder art. We want to make clear this is not how the game will appear in its final version.

Moving on to Federations: During PDXCON 2019 we said that we would give more information on the expansion later during the year – and today we want to share some news that Federations is targeted for release in early 2020. Although we understand that some of you might be disappointed that Federations will not be released in December, we want you to know that we are taking more time to make sure that the next update is going to be amazing.

In addition, to give us the best chance of improving some of the pain points you’ve shared with us, we have assigned some of our team members to focus solely on trying to improve performance and AI. It is very important to us that 2.6 does not compound any of the current issues with the game, and that we can take the time we need to address some of the issues remaining from 2.2. It’s important to remember, however, that working on these kinds of issues is not a sprint, but a marathon – it's something that is constantly being worked on over longer periods of time.

If you want to read more about performance, and how we work to maintain it over time, we shared some more information on this topic in Dev Diary #149.

While we have been unable to give concrete information or specifics related to these issues, we can say that it is very important to us. With that said, it's important for us that you know that your feedback is not being ignored, even if we have no news to share.

We want to thank you for being such a dedicated community and helping us by providing feedback and reporting issues with the game. We appreciate this to no end and encourage you to continue voicing your thoughts to us.

From the beginning of next year, we’ll be doing a series of dev diaries dedicated entirely to answering questions related to specific topics each week. The schedule for those dev diaries will be released later in December when we’ll summarize and round up the year.

---

That is it for this week! Dev diaries will resume their regular schedule, and as promised last week, next week we will be talking about some of the new things affecting diplomacy, such as Envoys.

P.S:
Since this dev diary had no pictures I felt it was necessary to add something, so here's a picture of the premade Lithoid-empire that some of you have been asking us to add to the Lithoids Species Pack! (Will also be updated with 2.6)

upload_2019-11-21_15-28-57.png
 
  • 1Like
Reactions:
I was kind of hoping for a minor patch before Thanksgiving.
I am just mentioning that us Canadians celebrate Thanksgiving in October, which probably confused me upon reading your post.
 
I am just mentioning that us Canadians celebrate Thanksgiving in October, which probably confused me upon reading your post.
I’m British. No idea about when it was, it’s not something we celebrate.
 
It’s probably easier (and therefore faster) if they incorporate the fixes into the main update, rather than working on 2.5.2 and 2.6 alongside each other. You also don’t want to fix 2.5 and then replace the fix with superseding 2.6 code or even worse fix 2.5 and not fix the version of 2.6 which is behind closed doors.
As someone who has to maintain this exact scenario: Don't. The overhead it eats is enormous (to the tune of single merges taking half a day). And even then, it produces plenty of messes.
 
It’s probably easier (and therefore faster) if they incorporate the fixes into the main update, rather than working on 2.5.2 and 2.6 alongside each other. You also don’t want to fix 2.5 and then replace the fix with superseding 2.6 code or even worse fix 2.5 and not fix the version of 2.6 which is behind closed doors.

I’d like a few small bugs patched soon, but this is probably the best way forwards. Hopefully the new mini performance and bugs team will continue beyond 3.0 while other content is produced.

Also I’d like to take a moment to point out that 2.6 will still be buggy, just less buggy and (hopefully) performing better than it would be otherwise, but yes - a definite step in the right direction.
With 2.6 not being released until 2020 (and we're not sure when "early 2020" is, we're assuming January, but it could be February) that would mean not even a beta patch to fix any outstanding issues between 2.5.1 and February.

That would be a nearly unprecedented gap between patches.
 
I am fully support the idea to wait and roll out update in more polished state.
But anybody knows why devs do not do beta-tests for the updates? It can help a lot. Some other games(like anno, that is also strategy) do close and open betas for the dlc and updates. Is there any reason why paradox do not do that?
 
I am fully support the idea to wait and roll out update in more polished state.
But anybody knows why devs do not do beta-tests for the updates? It can help a lot. Some other games(like anno, that is also strategy) do close and open betas for the dlc and updates. Is there any reason why paradox do not do that?

After the main patch comes out you can opt in to the open beta which handles the fixes, re-balancing, and updates.

I don’t know why they don’t do closed betas, maybe the code is changing too much to be practical?
 
After the main patch comes out you can opt in to the open beta which handles the fixes, re-balancing, and updates.
I remember that they are doing open betas for minor versions, but can't recall that they did it with major version update.

I don’t know why they don’t do closed betas, maybe the code is changing too much to be practical?
Also they might not have resources to run the process and handle all the errors =(

They playtest this all in-house.
200 beta-testers on close beta will find much more than 20 devs playing it one-two hours every day.
 
A beta requires that all simple bugs are found and fixed first - because fix time for betas are a lot longer than the fix time for an internal version.
1-2 QAs playing full time, aware of which bugs are being fixed and what has not been reported yet, and being able to chat in during lunch time with a developer is a lot more helpful than 100 betas playing on their free time on last weeks version of the game.

Once internal QA no longer is finding any significant issues, the game can be opened for more public consumption. For a major release, with significant changes in the patch releasing it before the DLC (as a beta or not) would distance it from the DLC, which would reduce complaints about the DLC breaking the game, but probably also reduce sales of the DLC as the hype of the DLC is linked to the hype of the patch.
 
For a major release, with significant changes in the patch releasing it before the DLC (as a beta or not) would distance it from the DLC, which would reduce complaints about the DLC breaking the game, but probably also reduce sales of the DLC as the hype of the DLC is linked to the hype of the patch.
Yea, I am also thinking that hype about DLC is the case here. But surprisingly for me, other games also do close and open betas and it did not harm the hype, it is actually helps.

A beta requires that all simple bugs are found and fixed first - because fix time for betas are a lot longer than the fix time for an internal version.
1-2 QAs playing full time, aware of which bugs are being fixed and what has not been reported yet, and being able to chat in during lunch time with a developer is a lot more helpful than 100 betas playing on their free time on last weeks version of the game.
Of course for beta you need stable beta build, it does not make sense to do beta with pre-alpha version ;)
One of the pluses for beta I see is that stellaris have too many options and variants. Internal QA just does not have time to test all the possible combinations(and I don't think they are doing real full play-test. It takes too much time.).
Also, one more plus for betas is that devs and QA quite often see the game from different perspective than regular users. And it is really important.
Beta-testing should not replace internal QA. It is totally different thing, but can be really helpfull in polishing game.
 
Yea, I am also thinking that hype about DLC is the case here. But surprisingly for me, other games also do close and open betas and it did not harm the hype, it is actually helps.
Those games don't have a free patch with more features than the DLC though. For them it is a open/closed beta of the DLC in question.
They could of course have a beta of the DLC, but then it becomes a question of paying to beta (which of course is what some feel we are at now...)
 
200 beta-testers on close beta will find much more than 20 devs playing it one-two hours every day.

Find them, yes.

But will they take the effort to duplicate the bug, test for all variables, and create a clear set of repo steps that will allow the devs to quickly center in on the section of code that needs fixing?

Most "Open Beta" testers aren't aware of what a good or even usuable bug report looks like, much less willing to take the effort to do one, so it usually takes someone whose job is collecting all the bug reports about X issue and try to cobble together some repo steps from all the random bits and pieces. (That's why most devs ask for a save/in-game screenshot/some form of metadata when you report a bug, because they can use that to fill in the blanks that an amateur tester might not have any idea to look for)

For some bugs (especially rare bugs or bugs that might have weird variables) that's fine, but you can also end up with too much noise and figuring out what the bug actually is could take more man-hours than if you have a few professional QAs who know how to properly investigate and report on the bugs they come across.
 
I've been in several closed betas for updates, for a half-dozen games, including MechWarrior 4 and more recently, KSP. When properly organised and managed, which they were in all those I participated in, they proved very successful in contributing to the QA of the associated update.

I'd suggest getting some experience of the process, and participating in a closed beta of an update, before assessing their worth.
 
In addition, to give us the best chance of improving some of the pain points you’ve shared with us, we have assigned some of our team members to focus solely on trying to improve performance and AI. It is very important to us that 2.6 does not compound any of the current issues with the game, and that we can take the time we need to address some of the issues remaining from 2.2. It’s important to remember, however, that working on these kinds of issues is not a sprint, but a marathon – it's something that is constantly being worked on over longer periods of time.

If you want to read more about performance, and how we work to maintain it over time, we shared some more information on this topic in Dev Diary #149.

While we have been unable to give concrete information or specifics related to these issues, we can say that it is very important to us. With that said, it's important for us that you know that your feedback is not being ignored, even if we have no news to share.
I think the point is, the AI was never great, but maybe at some point could actually play. That changed with 2.2, when core gameplay mechanics changed drastically.
As much as I want to agree with you, that writing an AI is a marathon, I can't really. Just like you can't reasonalby expect to finish a marathon if you keep calling a cab every half an hour to take you back to the starting line, you won't get far with an AI if core game mechanics keep changing too fast to keep up, as it happened with 2.2. And that was a year ago.

I know, creating an AI is hard. And maybe it's intentional that all other non-player empires are more like decoration and less like actualy opponents, I just hope they could be more, eventually. If you can't be bothered to do it yourself, give the modders access to more powerful tools than that awful scripting language. even the most dedicated modders can't do much with that.

Also, words are cheap. The new improvements you are hinting at will only be available for the next big update, which will change core game mechanics yet again. How big of an improvement can we hope for, when it should alraedy be quite challenging to even adapt the existing, pretty incompetent AI to the new changes?

As much as I want to be more optimistic, as long as the business model stays at the current scheme to change core gameplay to make room for more content which can then be sold as a dlc, I can't realisically expect the AI to improve in the long run.

Please prove me wrong.
 
I'm SO happy to see this outreach. And am 100% on the take the time to push out a product everyone can feel proud of / excited for train.

Federations marks the final major loop I've been waiting for for Stellaris to really feel completely immersive for me. (minus maybe espionage gameplay), hearing that you're working to enhance AI and late game performance issues is very exciting; and the fact that you guys reached out to communicate that is also extremely re-assuring. Thanks for all the hard work.