Larian Studios Forums Baldur's Gate III Baldur's Gate III - General New interview with Sven by Gamespot

I think you are over interpreting. Their "starting point" might have been RAW DnD, but it doesn't mean that they implemented whole system as in PnP and then changed it. As some of the original impomentations (like reactions) would require their own UI and systems, I find it more likely that they made the decision on how to impliment certain mechanics, before they started putting them in game. They are still putting stuff into the game, so certainy at no point was there a comprehensive, pure DnD BG3 build.

And if anything, EA is the perfect opportunity to test their homebrew ideas on the audience. There is no point on giving players raw systems, if you intend to change them.

Community doesn't get to design the game. That never happens. At best we can be guinea pigs for Larian's mid development experiements.

Let's be honest, what that means was, they flipped through the Source books for 10 minutes and said "BORING" and started with DOS2 instead.



also, this. WotC is not quite as bad as Games Workshop, but give them time, they'll get there

No these are not margins of error. They are ranges. And saying a sales figure falls somewhere between 2 million and 5 million is the equivalent of saying nothing at all. It is completely useless information and does not qualify as data. You cannot input a range into a dataset. It makes no sense. In the case of BG3, is it 2 m? Is it 5 m? Is it something else? Who knows. And it makes a HUGE difference whether it is 2 m versus 5 m. And if Steamspy is saying "we don't know if it is 2 m or 5 m," then they are literally saying there is a margin of error of something like 95% to what they're saying, and anything with that huge a margin of error is useless information.

Yes, exactly. My way of interpreting any numerical information provided to me in the form of a range is to use the lowest end of the range as my number. That is what is most honest. So for Solasta, a game I backed and love, that number (per Steamspy) is 200k.

There was no 100% faithful adaptation which was also in end-user-playable form. That much is clear from what HAS been released. There are only a handful of classes and fewer sub-classes, many missing spells and mechanics, missing races, missing features like Expertise, and so on and so forth. Whatever they did have might have been rules-faithful, but it was only a partial amount of the possible game space. Even then, it was pre-alpha and thus unstable and not correctly working properly.


Also, games can't be developed by committee. Larian discovered this when trying to write the D:OS 2 Origin stories. Everyone at Larian could contribute ideas, but too many cooks spoil the broth and it had to be pared down to just writers in order to actually get them finished. Democracy doesn't work for development, a direction must be chosen and that enforced, otherwise it spins its wheels and gets nowhere.

From Steamspy: About.
So its even more accurate than I stated, using a 98% confidence interval rather than a 95% confidence interval. As for how accurate it is exactly, well, here are a few anecdotal testimonies from developers.

Whilst the exact margins of error are not provided to you unless you subscribe to the service, you can be pretty damn absolutely sure the margins of error are within the given ranges they provide there. Ergo, the statistic has some value. How about you actually learn something about Steamspy first before commenting on it?

Hey thanks for the link. The original interview of Chris Cocks in the gamesindustry.biz article, which this article quotes, is even more informative. I for one am thrilled we will be getting so many other D&D games. At least this way, given how bitterly disappointed I am with BG3, I have those other games to look forward to as POSSIBLY being good D&D videogames I will enjoy playing (though definitely NOT Dark Alliance). The game being developed by Hidden Path Entertainment in particular (AAA open-world third-person game) has my strong interest.

That's an article from 2017. BEFORE the change to the default Steam profiles (used to gather Steamspy's data) that I was mentioning in my reply and that severely compromised Steamspy's reliability as a source.

It was also from a time where the "range" used as margin of error was measured in few thousands units at most (as you'll notice if you read the article you linked). Not hundreds of thousands or even millions as it is today.

Yeah, don't be an ass. I'm quite confident my knowledge of statistics is way beyond yours.

All this is saying is that Steamspy has 98% confidence the true sales number falls somewhere between 2 million and 5 million. It is easy as pie to have that kind of confidence when you place your value between such a humongous range (i.e. margin of error). The confidence is not the issue. The range is the issue. In any other walk of life, if you gave someone a margin of error of 3 million and then said, "Don't worry; I have 98% confidence the true number is somewhere within that 3 million margin of error," they would laugh you out of the room.

It didn't compromise the reliability, it should still have a confidence interval of 98%, its just given the fact that you have less useful data to extrapolate from, the size of the range has become much larger. Where as before they could estimate within say ~20,000 on a game with 10,0000,0000 sales, now they need to provide a much larger band (say 1,000,000) for example. Whilst the 2nd figure is less useful, it will still hit the mark just as often as the previous statistic did.

Exactly! Thanks for pointing this out.

Do I need to quote you again?
I was responding to this. We are in agreement now, they are margins of error, correct? If you are going to so confidently state something which is incorrect, be prepared to be smartassed out of the room. Furthermore, Steamspy collects many other analytics which are exceedingly useful. The analysis of user reviews as well as looking at trends like the day 1 concurrent player count followed by the decline in players are both useful points of data which can be used by someone who is new to the industry and potentially looking to make a new game to help them evaluate the market case for their potential game.

The range 200k to 500k may not be as useful as 210k to 220k... but to say that it is "useless" is a stretch. Obviously, the new data is much less precise, but it does allow you to conclude that Solasta outsold the game "My Friend Stalin" (0 to 20k owners) at a very high confidence level. Granted, you would probably guess that correctly too.

We are realistically not going to know the ultimate impact of what effect BG3 and Solasta will have on cRPGs moving forward, let alone the wider gaming industry, for years. It appears to me that we are finally seeing evidence of what impact DOS2's combat design is having on other games, with Project Triangle Strategy and Lost Eidolons incorporating field effect elements into their games that are obviously inspired by DOS2. Zoria: Age of Shattering appears to have adopted the AP system instead, and the devs for that game have been on record saying that they've been inspired by DOS2 (and it plays much closer to DOS1 compared to DOS2 otherwise).

That said, 200,000 minimum sales for Solasta isn't bad, and actually pretty decent given the context. I don't think anyone was expecting earth-shattering numbers, but I also don't think anyone was expecting Pillars of Eternity II levels of bomba either, which was the bigger worry. I don't think POE2 ever made it past 500k despite it having a much higher budget than most other cRPGs. I think Pathfinder: Kingmaker has gone past 1 million at this point, but it had an extremely rocky start due to being an unfinished buggy mess at the time, but the base game was so strong that it flipped its reputation around completely once it got fixed and now enjoys insane word of mouth.

That's very possible, and I think there is a lot of good to learn from D:OSs design, especially when it comes to PC RPG design. Externalizing status effects, buffs and debuffs into visible hazards on the tactical maps is great, and It feels silly that D:OS had to remind people that AP is good.

It is very much modern XCOM situation - I very much like the presentation, but underlying systems need work.

DOS2 has good systems, I just wouldn't try applying them so strongly to DND.

My post had nothing to do with Solasta vs BG 3, just with the assertion that Steamspy is mostly useless now - its not. It is still exceptionally useful to potential developers, even if it is less useful than it was before. There is a reason it still exists and is maintained, because there is a very good business case for it and because the data has an audience that wants it.


Regardless of whether Solasta sold 200,000 or 500,000 I still think it did pretty well. Even if we knew exactly how many copies it sold however, we could not properly estimate its gross revenue, let alone its profit, because the game is sold at different prices in different regions and we don't have access to a breakdown of the sales figures by region. Once the game has been out for longer, discounts will also complicate the matter. For those curious, here is the price of BG 3 and Solasta within different regions.

Solasta.
BG 3.

DOS had plenty of lousy systems even ignoring the comparison with D&D, frankly.
It's a game that I liked despise some of its design choices, rather than because of it.

There's an underlying problem, though: it's a game that thanks to its high production value (nice enough graphics, fully voiced, pretty to look at, reasonably intuitive to approach, etc) popularized a former "niche genre" among people who never paid attention to it before, so these "new fans of the genre" now seem to think everything about its design was pristine and it can't be topped.

People are always afraid of changes. Almost like those despisable BG2 fans

I can't even tell what you are trying to say, frankly.

Nor I'm sure where the "fear of changes" even comes into play here.

If you mean does Steam consider this to be a margin of error? Sure. But I don't, which is my whole point and what I was saying. No respectable statistician would consider claiming a range of 3 million as a margin of error (unless you're dealing with numbers that are astronomical).

But feel free to believe what you will.