Yeah, I forgot to add that idea. IMDB shows the same trend for most recent movies. I have in fact been scraping IMDB ratings for a year now, so I can check.
In my opinion, game quality peaked around 2005-2012: the tech for games was still pretty good and gaming companies hadn't been poisoned with the woke mind virus.
Haven't played many games released post 2016, though I think Overwatch and Nier Automata were decent; Dark Souls III, The Coffin of Andy and Leyley, and Disco Elysium were all great. My prediction is that games are going to get really, really good when AI tools start to get strong.
Alternative hypothesis for the time trend -- it's selection:
- Newer games are played by less discerning and newer gamers, who give more lenient ratings
- Older games that people still play enough to get rated on steam tend to be good.
Yeah, I forgot to add that idea. IMDB shows the same trend for most recent movies. I have in fact been scraping IMDB ratings for a year now, so I can check.
In my opinion, game quality peaked around 2005-2012: the tech for games was still pretty good and gaming companies hadn't been poisoned with the woke mind virus.
Haven't played many games released post 2016, though I think Overwatch and Nier Automata were decent; Dark Souls III, The Coffin of Andy and Leyley, and Disco Elysium were all great. My prediction is that games are going to get really, really good when AI tools start to get strong.
It's quite a disappointment to see games turn up as a topic here. I hope it isn't a taste of things to come.
Sometimes people need to let loose.
Why?
A bit late to the party here but using a model explainability package like SHAP on top of the xgboost would yield very cool visuals and analyses
Very interesting look at the data behind game ratings.