Just wanted to say this is a great market, good question

The "Top Sellers" page seems like it shows the games selling the most in the last week rather in the year overall. It's more likely to include random flavor of the month indie games as opposed to AAA mainstays.

I think Best of [year] - New Releases better represents sales over the entire year.

@Multicore Maybe so. Seems late to change the criteria now, but the title was clearly trying for the second thing, so everyone's thoughts? Like Multicore's comment to endorse switching. Like this comment to endorse not switching. If there is clear consensus I will do the switch.

@Multicore By vote of 8-1 and I agree that this is a better metric, so we'll switch to it.

@ZviMowshowitz Actually, this only tells us the top 12? If we can't figure out what the actual top 10 is, I don't think I'm ok with switching. If we can tell which of the 12 are the top 10, we can switch, if we do it relatively soon (e.g. by Feb 1). But I don't want there to a random excluding of 2 games at the end...

@ZviMowshowitz Ah, yeah, from bottom of that page:

Revenue specifics are not disclosed for the games on any of these lists, so we group them together into randomly sorted buckets to give an idea of how they all stacked up:

Platinum: 1st - 12th
[...]

Seems unlikely we'd be able to get a reliable ordering by the revenue metric.

@jskf Yep. Given that we are going to stick with the original metric here unless some someone can come up with a way around this. I may change the title to clarify.

@ZviMowshowitz As someone with a NO position I would prefer the yearly metric even if that means we look at the top 12 instead of a top 10, but bias towards changing as little as possible seems reasonable.

@jskf Yeah I think the yearly favors YES structurally if anything, so it might actually be a fair trade, but I think it's basically too late to do that.

Some relevant info: The Finals was recently top 10 and it used AI to generate some voice lines, though it was released 2023 so it doesn't count towards this question. It also doesn't have any AI disclosure on the steam page.

I anticipate
- more AI tools for making game assets (also other steps, but afaict those wouldn't need to be disclosed)
- that those tools will have good tutorials & user interfaces, so that many people can adopt them even under moderate time pressure to finish their product
- that those tools will save developers/artists time, more so than existing tools

Already I think
- AI generated textures are of highest quality, comparable to the work of professionals who draw textures for distant houses that few players ever look at etc.

uncertainties
- how late in the game development are assets still made? are all high-end candidates for the top 10 list already in such a late stage right now that they won't get AI assets?
- will transitioning to AI tools really be easy enough, or will all startups screw up their UIs again?
- will the disclosure policy be revoked or changed to not include AI generated art assets?
- will companies / artists refuse for legal or comfort reasons not to use AI?

Edit 24-1-11: Note for future self P=75%

I'm unsure about this one. OOH, a game could add such a disclosure due to a trivial feature, with the purpose of stirring up controversy or making itself more notable. Including AI can have positive or negative valence depending on the audience and how it's marketed. The state of AI tools has little to do with any of this.

OTOH, top 10 is a really high bar. I think I'm on no due to boring base rate reasons.

games such as the planned remake of Persona 3 would count

would not count?

@BenjaminCosman yes I meant would not count

@ZviMowshowitz now it says "would be not count" :)