01-06-2015, 12:44 PM
So between last night and this morning I have switched cartypes from the two variable system to all variables. What do I mean by that?
GC 1.17 and older uses what I call the "Two Variable" system. Every car type has two primary variables, such as Performance and Driveability. The consumers would look at these two variables as the most important factor when buying a vehicle. Say 10 points for Primary and 5 points for secondary. And then other values were hardcoded so if primary was Performance than Fuel would be worth 0.25 points where as Power would be 1.5 points.
The AI also had two variables, using those two variables they would select which vehicle to design.
In 1.18 this has changed.
I've taken all the primary values of vehicles, such as performance and fuel etc and applied values between 0-1 for each car type. This allows for more flexable cartypes, it allows for regional differences using the same car types. It also allows for dynamic changes in car types based on yours and the AI's actions. It improves moddablity of the game, makes the turn times faster, and most importantly it makes the consumers much more picky, realistic, fluid.
In theory of course. I still have to do testing and tweaking.
I've also converted the Top 10 AI file to handle this data. Each AI company now list the primary vehicle values in their behavior code. Each variable has a range between 0-1 that is the percentage of what values it's looking for. So for example if Lumbergini is a super car company. Then they will have high values for Performance and Driveability, but low values for Fuel Economy, Safety, Cargo, and Dependability. When Lumbergini goes to parse the car type list it should be much easier for it pick out vehicle types that meet it's behavior. I've also added a "GenericDeveloper" flag for companies such as Generic Motors, to give it a more broad focus than it's behavior when developing new vehicles. This should force niche makers such as Lumbergini into very few car types with high performance/driveability, and allow Generic Companies like Forward and Toyoda to produce a broad range. The new code should be faster (again not tested yet.) which is one of my primary goals over the next few builds so that I can get us up to 600AI.
The new behavior values could also make its way into the AI's car design sliders. Not sure if I'll do that. But if so, that would make for example any SUV's designed by Catalac have higher luxury ratings, even if SUV don't need much luxury.
Anyhoo, currently working on the AI's selection of Car Types code. Should be able to get it working by tonight, I'll then run a few hundred test and tweaks to the new code using the Top 10 companies. Then I'll move on to adding these values to the top-25 AI and then the full AI xml file.
GC 1.17 and older uses what I call the "Two Variable" system. Every car type has two primary variables, such as Performance and Driveability. The consumers would look at these two variables as the most important factor when buying a vehicle. Say 10 points for Primary and 5 points for secondary. And then other values were hardcoded so if primary was Performance than Fuel would be worth 0.25 points where as Power would be 1.5 points.
The AI also had two variables, using those two variables they would select which vehicle to design.
In 1.18 this has changed.
I've taken all the primary values of vehicles, such as performance and fuel etc and applied values between 0-1 for each car type. This allows for more flexable cartypes, it allows for regional differences using the same car types. It also allows for dynamic changes in car types based on yours and the AI's actions. It improves moddablity of the game, makes the turn times faster, and most importantly it makes the consumers much more picky, realistic, fluid.
In theory of course. I still have to do testing and tweaking.
I've also converted the Top 10 AI file to handle this data. Each AI company now list the primary vehicle values in their behavior code. Each variable has a range between 0-1 that is the percentage of what values it's looking for. So for example if Lumbergini is a super car company. Then they will have high values for Performance and Driveability, but low values for Fuel Economy, Safety, Cargo, and Dependability. When Lumbergini goes to parse the car type list it should be much easier for it pick out vehicle types that meet it's behavior. I've also added a "GenericDeveloper" flag for companies such as Generic Motors, to give it a more broad focus than it's behavior when developing new vehicles. This should force niche makers such as Lumbergini into very few car types with high performance/driveability, and allow Generic Companies like Forward and Toyoda to produce a broad range. The new code should be faster (again not tested yet.) which is one of my primary goals over the next few builds so that I can get us up to 600AI.
The new behavior values could also make its way into the AI's car design sliders. Not sure if I'll do that. But if so, that would make for example any SUV's designed by Catalac have higher luxury ratings, even if SUV don't need much luxury.
Anyhoo, currently working on the AI's selection of Car Types code. Should be able to get it working by tonight, I'll then run a few hundred test and tweaks to the new code using the Top 10 companies. Then I'll move on to adding these values to the top-25 AI and then the full AI xml file.
"great writers are indecent people, they live unfairly, saving the best part for paper.
good human beings save the world, so that bastards like me can keep creating art, become immortal.
if you read this after I am dead it means I made it." ― Charles Bukowski
good human beings save the world, so that bastards like me can keep creating art, become immortal.
if you read this after I am dead it means I made it." ― Charles Bukowski