Something that I've been working on, enabled by the recent "Remove Rating Limits" bounty:
![[Image: attachment.php?aid=1541]](https://www.ventdev.com/forums/attachment.php?aid=1541)
A hypothetical picture of what the worldwide auto fleet might look like in a GearCity game. The game, of course, focuses on selling new cars. But as anyone who's been behind a late-'80s Mercedes 420 SEL at a traffic light recently knows, in the real world used cars form a considerable part of the market.
GearCity models this to some extent, with used cars playing a factor in sales. The modeling only goes back a few years, which makes sense when I think of which cars dealers that also sell new cars are likely to sell, especially the "certified pre-owned" or off-lease markets.
Still, I'd long thought, "what would the entire market look like?" Could we model the number of Model T's on the road? How rare might my former colleague, who drove his Model A to work on good-weather days in the early 2010s, be? How long are all those phaetons relevant, anyway?
Thus, this early-version report. It looks at the dependability score, now that it's not constantly-decreasing thanks to the new bounty, to see how likely a car is to survive; the more dependable it is, the more likely, but there's always a chance it will be destroyed in a wreck. Then it's just a bit of math to approximate how many might still be on the road, and from there, to figure out how much fuel is being burned.
In this image, we see that while sales peaked in 1928 before the Depression, and 1928 model-year cars are more numerous than '29 through '33 model years. But due to their aging and the gradual revival of the market, 1934 model years are more common than '28, with 1941 being the most common before the war tanks the market. Meanwhile, '38s use less fuel than '37s despite being more numerous, thanks to their improved efficiency. Not yet in the table is that the overall-average miles-per-gallon is also 24.
So far this is all a high-level approximation based on yearly average dependability/fuel economy, but I have the groundwork to make it per-model - so if the most reliable cars are highly-inefficient trucks, they'll drive the fuel economy down for a long time, for instance.
From there, there are lots of directions it can go. Making fuel consumption more accurate, via historical data on mileage driven per year, which I've found dating back to 1936. CO2 emissions, based on the miles driven, fuel economy, and type of fuel. More general greenhouse gas emissions via approximation of N2O and CH4 (nitrous oxide and methane) emissions, thanks to the new emissions system in GearCity. Preliminary testing shows CO2 will almost always be responsible for 95 - 98% of greenhouse gas emissions, but exceptionally dirty cars may push other contributors above 10%, such as a vehicle one of the AIs built with an emissions score of 68,000. Another possibility is showing an evolution of the figures over time, rather than the current snapshot of the current year picture. We see 724 million gallons of fuel were burned in 1944, but don't yet have a good view of how that compared to 1940 (although, for context, 724 million gallons represents about 2 days' worth of consumption in the U.S. in 2022, and actual fuel consumption in the U.S. in 1944 was 11.1 billion gallons).
I doubt I'll get as detailed as making award-winning cars survive longer as collectors' items, or modeling how many cases of lung disease were caused by highly-polluting vehicles. Or how many lives were saved by the future safety regulations. As usual, these reports are driven by what questions I find interesting, and which I find answerable with the data available. More likely, a "contracts history" page to show your prowess with raking in the contract dough, or perhaps a variant of the "superlatives" screen with more filters ("highest emissions" superlative, anyone?) will come first. I'm also tempted to add a "displacement" trend chart, showing the rise and eventual fall of the 6.2L V8 Hemi engine.
A hypothetical picture of what the worldwide auto fleet might look like in a GearCity game. The game, of course, focuses on selling new cars. But as anyone who's been behind a late-'80s Mercedes 420 SEL at a traffic light recently knows, in the real world used cars form a considerable part of the market.
GearCity models this to some extent, with used cars playing a factor in sales. The modeling only goes back a few years, which makes sense when I think of which cars dealers that also sell new cars are likely to sell, especially the "certified pre-owned" or off-lease markets.
Still, I'd long thought, "what would the entire market look like?" Could we model the number of Model T's on the road? How rare might my former colleague, who drove his Model A to work on good-weather days in the early 2010s, be? How long are all those phaetons relevant, anyway?
Thus, this early-version report. It looks at the dependability score, now that it's not constantly-decreasing thanks to the new bounty, to see how likely a car is to survive; the more dependable it is, the more likely, but there's always a chance it will be destroyed in a wreck. Then it's just a bit of math to approximate how many might still be on the road, and from there, to figure out how much fuel is being burned.
In this image, we see that while sales peaked in 1928 before the Depression, and 1928 model-year cars are more numerous than '29 through '33 model years. But due to their aging and the gradual revival of the market, 1934 model years are more common than '28, with 1941 being the most common before the war tanks the market. Meanwhile, '38s use less fuel than '37s despite being more numerous, thanks to their improved efficiency. Not yet in the table is that the overall-average miles-per-gallon is also 24.
So far this is all a high-level approximation based on yearly average dependability/fuel economy, but I have the groundwork to make it per-model - so if the most reliable cars are highly-inefficient trucks, they'll drive the fuel economy down for a long time, for instance.
From there, there are lots of directions it can go. Making fuel consumption more accurate, via historical data on mileage driven per year, which I've found dating back to 1936. CO2 emissions, based on the miles driven, fuel economy, and type of fuel. More general greenhouse gas emissions via approximation of N2O and CH4 (nitrous oxide and methane) emissions, thanks to the new emissions system in GearCity. Preliminary testing shows CO2 will almost always be responsible for 95 - 98% of greenhouse gas emissions, but exceptionally dirty cars may push other contributors above 10%, such as a vehicle one of the AIs built with an emissions score of 68,000. Another possibility is showing an evolution of the figures over time, rather than the current snapshot of the current year picture. We see 724 million gallons of fuel were burned in 1944, but don't yet have a good view of how that compared to 1940 (although, for context, 724 million gallons represents about 2 days' worth of consumption in the U.S. in 2022, and actual fuel consumption in the U.S. in 1944 was 11.1 billion gallons).
I doubt I'll get as detailed as making award-winning cars survive longer as collectors' items, or modeling how many cases of lung disease were caused by highly-polluting vehicles. Or how many lives were saved by the future safety regulations. As usual, these reports are driven by what questions I find interesting, and which I find answerable with the data available. More likely, a "contracts history" page to show your prowess with raking in the contract dough, or perhaps a variant of the "superlatives" screen with more filters ("highest emissions" superlative, anyone?) will come first. I'm also tempted to add a "displacement" trend chart, showing the rise and eventual fall of the 6.2L V8 Hemi engine.