automationgame.com

I don't know if it has already been asked on the forums or anything, because it's quite tough to search for :

We all know that history is made on the butterfly effect.
So why not implement a variable history.

For example : what if there was not oil crisis in the 70's ? Would it stay high-perf gaz-guzzlers 'till the 90's, and then there would be big research on petrol synthesis to replace the vanishing one ?
Or just stay this way until the Presidente Lolcat says in 2019 :

Just joking, but you see my point.

And that would be for a lot of facts like, what if the US didn't enhance their big-blocks but rather tried DOHC and smaller engines, while Europe detuned their race engines for their road cars ?

Of course it would be a real headache to implement, but in the end, it would add a LOT to the gameplay value.
If we stay with history, you could just learn the real facts and predict everything in the tycoon game. We would have a major issue with IA-balancing.
So what about it ? Already thought, DLC... ?

Be prepared to have it as a mod anyway, just like Game Dev Tycoon having a "Platform Randomizer mod".

There will be a choice between randomized history and a realistic one. This has been discussed before. However, I am not sure about the butterfly effect.

As Jakgoe stated, the subject has been spoken of before.

There will be a complete historical game, with real-life events occuring like the Oil Crisis and such.
On the other hand, before starting the campaign you can select to get randomized events. So an oil crisis that lasts 5 years may occur in 1950 or in 2004 for example, or a civil war might happen in France in 1980? That kind of things

Im not sure about implementing things like "What if... the US used a lot of I4 instead of V8?". As soon as you play a Historical game, you could say that 'tags' that go along with AI starting in a particular country (Like US: mainly V8, emissions and efficiency? What are they?) make an AI build more cars like it happend in our history. When you play a randomized game, you could say that these tags disappear, and the AI in the US also just builds in I4 almost as random as a V8.

If the AI is done properly, AI companies will build cars which will be bought by the people in the regions. The devs just have to put the right customer groups into the various countries (historic campaign America: lots of prestige and comfort, not so much sportiness and eco) and the AI has to build cars according to that.

The player can then decide which approach he wants to make: Make something which sells well on the mainstream market (big, cheap V8s) or for example the "import approach" with economical I4s and the potential to sell a lot in the niches of the region. An actually "intelligent" AI would also work a lot better in randomized campaigns.

The best approached for a somewhat random would be in multiple tree-shapes.
Anything can happen at anytime, and this is the beggining of a tree-shape in which there are many other events occuring because of the first one.
And as there could be many very different events, there would be as much tree-shapes.

For example :
Development of the young drivers market since babyboom = pony cars + racing culture.
Pony cars + racing culture = muscle cars.
Muscle cars = no improvement upon push-rods (because of the high quality of the american ones) leading to more costs on change + high performance culture leading to major impact of an oil crisis = Late american car industry.
Japanese cars imports in the US since the 70-80's + late american industry = americans buying more and more economical hatches from foreign makes.

Seems legit.

That seems quite interesting. However, if you do this, some events must be random. Like "racing culture+oversaturation of V8s" could result in "emissions regulations" OR "high demand for high-performance I4s".

Of course. The main problematic is managing the inter-dependencies between each event.
Random laws, nationwide preferences, random interests in sports (like a sudden buggy hype), and all...
That's gonna take quite some time and thinking, but I think we have to have it someday.

I wouldnt recommend putting this is Automation V1.0 though. I think the historical and random events like they ahve planned now will suffice, later your idea might be an interesting tweak to the random event game.
Though, it would mean you can predict some things after a while. When the US uses a lot of V8s like you said, then you can already get ahead with making a very big super sporty car with big HP, and then afterwards start making I4 because you know the oil crisis is coming.