Worldwide Rush | Blogpost 2
Hi everyone!
Most of this week was about AI. Adding features and testing. It is very time-consuming making sure that AI behaves the way you want. Sometimes small changes can break their brains, but at least it makes me chuckle looking at them. It is like teaching your own child how to play, but it takes all instructions literally, and you are the only one to blame when you contradict yourself. But as I understood, it will be the most iterated system in the game before and after release.
Now let's talk about what we added.
Note, that this is not yet in the demo build.
[h2]AI[/h2]
AI is built from the grounds up to be deterministic and not demanding. The demo will feature up to 3 competitors, but the full game will have up to 100. This number can be bigger, but the vehicle's amount grows silly with more. Already with 10 vehicles per company, you will have 1000 vehicles total. After a decade or so, a single AI can grow to 5-10k vehicles, which really tests performance. Granted, then many others will be bankrupt or bought out.
AI thinks in steps per frame. This way it mimics a human player and does not act like a god that does 1000 decisions per frame. AI will search for available paths and vehicles that fit them. It will do so indefinitely, stopping to reevaluate existing routes or to search for a new hub. AI will invest in good routes that it found if there are enough finances to do so. Otherwise, it will plan ahead discarding, improving and finding new potential routes. Easy AI will have less capacity to think per frame, while the hard one will have more.
Hopefully, AI will provide difficulty some of you are searching for. They definitely make you play slower and more methodical. Last time I played in China it was a very interesting experience trying to compete against the local companies, losing country's trust and making some small plans to take it back. But then you zoom out and see that the whole world has skirmishes and plans. For those who never wanted competition, then good news! They will be optional. But there is still much to do.
[h2]Difficulty[/h2]
AI will have several difficulty options that will allow you to experiment with them more. Easy AI will not buy other hubs or companies, they will simply try their best to thrive. Medium AI will buy other companies and hubs, but it will not try to buyout players company. But with Hard AI, there will be no friendliness. It will use all tricks available to try and be the dominant company in the world. AI will not cheat as there is no code for them to do that, they will have the same rules and information as you do.
[h2]Passengers[/h2]
This topic requires a blog post on its own. Today, I just wanted to mention how passengers choose which vehicles to board. This should help to compete against AI. Here is the list of current factors:
- Travelers will ignore vehicles that are taking longer paths than the shortest available. By providing short direct routes, you can always win passengers if competing against indirect routes.
- Then there is trust. Passengers will prioritize vehicles that are from trusted companies. Low trust means that it will be very difficult to compete for the same destination. But you can always pick the neglected destination and earn trust that way.
- Vehicle tier can bypass trust if you are offering higher tier vehicles. Comfort is more important than company loyalty.
- And passengers will prioritize vehicles that are in the city longer. This will only influence your own company vehicles and will not affect competition.
With that being said, if there are enough passengers to board, they will board all vehicles. Passengers do not wait in line, they just board whoever accepts them, starting from the most factor favored vehicle.
[h2]Cities Shrink[/h2]
Cities will be able to shrink now. For this to happen, the overall fulfillment needs to be below 30%. Cities will stop shrinking if they reach the minimum level, which is 2 at the moment. With this, you can redirect your traffic through AI city and look how it fails to handle the increased demand.
[h2]Last Destination[/h2]
Another addition for neglected cities is that the last destination can change at the end of the month if its fulfillment is below 30%.
[h2]Demand Price[/h2]
Not all cities and destinations need to be the same. The next update will include destination demand price, which basically means how much passengers are willing to pay. This value will change based on supply. Neglected destinations will decrease its price, while fully satisfied will increase it.
[h2]Notifications[/h2]
Finally, I wanted to mention new notifications in the game. With added features, it is easy to get lost in what is happening and changing. Notifications should help by notifying you about the most important events that you should know about.
Until the next Friday, where we will talk more about Worldwide Rush development. Also, we will share the demo beta access to those who will want to test out all the new systems that will come with the next update.
Most of this week was about AI. Adding features and testing. It is very time-consuming making sure that AI behaves the way you want. Sometimes small changes can break their brains, but at least it makes me chuckle looking at them. It is like teaching your own child how to play, but it takes all instructions literally, and you are the only one to blame when you contradict yourself. But as I understood, it will be the most iterated system in the game before and after release.
Now let's talk about what we added.
Note, that this is not yet in the demo build.
[h2]AI[/h2]
AI is built from the grounds up to be deterministic and not demanding. The demo will feature up to 3 competitors, but the full game will have up to 100. This number can be bigger, but the vehicle's amount grows silly with more. Already with 10 vehicles per company, you will have 1000 vehicles total. After a decade or so, a single AI can grow to 5-10k vehicles, which really tests performance. Granted, then many others will be bankrupt or bought out.
AI thinks in steps per frame. This way it mimics a human player and does not act like a god that does 1000 decisions per frame. AI will search for available paths and vehicles that fit them. It will do so indefinitely, stopping to reevaluate existing routes or to search for a new hub. AI will invest in good routes that it found if there are enough finances to do so. Otherwise, it will plan ahead discarding, improving and finding new potential routes. Easy AI will have less capacity to think per frame, while the hard one will have more.
Hopefully, AI will provide difficulty some of you are searching for. They definitely make you play slower and more methodical. Last time I played in China it was a very interesting experience trying to compete against the local companies, losing country's trust and making some small plans to take it back. But then you zoom out and see that the whole world has skirmishes and plans. For those who never wanted competition, then good news! They will be optional. But there is still much to do.
[h2]Difficulty[/h2]
AI will have several difficulty options that will allow you to experiment with them more. Easy AI will not buy other hubs or companies, they will simply try their best to thrive. Medium AI will buy other companies and hubs, but it will not try to buyout players company. But with Hard AI, there will be no friendliness. It will use all tricks available to try and be the dominant company in the world. AI will not cheat as there is no code for them to do that, they will have the same rules and information as you do.
[h2]Passengers[/h2]
This topic requires a blog post on its own. Today, I just wanted to mention how passengers choose which vehicles to board. This should help to compete against AI. Here is the list of current factors:
- Travelers will ignore vehicles that are taking longer paths than the shortest available. By providing short direct routes, you can always win passengers if competing against indirect routes.
- Then there is trust. Passengers will prioritize vehicles that are from trusted companies. Low trust means that it will be very difficult to compete for the same destination. But you can always pick the neglected destination and earn trust that way.
- Vehicle tier can bypass trust if you are offering higher tier vehicles. Comfort is more important than company loyalty.
- And passengers will prioritize vehicles that are in the city longer. This will only influence your own company vehicles and will not affect competition.
With that being said, if there are enough passengers to board, they will board all vehicles. Passengers do not wait in line, they just board whoever accepts them, starting from the most factor favored vehicle.
[h2]Cities Shrink[/h2]
Cities will be able to shrink now. For this to happen, the overall fulfillment needs to be below 30%. Cities will stop shrinking if they reach the minimum level, which is 2 at the moment. With this, you can redirect your traffic through AI city and look how it fails to handle the increased demand.
[h2]Last Destination[/h2]
Another addition for neglected cities is that the last destination can change at the end of the month if its fulfillment is below 30%.
[h2]Demand Price[/h2]
Not all cities and destinations need to be the same. The next update will include destination demand price, which basically means how much passengers are willing to pay. This value will change based on supply. Neglected destinations will decrease its price, while fully satisfied will increase it.
[h2]Notifications[/h2]
Finally, I wanted to mention new notifications in the game. With added features, it is easy to get lost in what is happening and changing. Notifications should help by notifying you about the most important events that you should know about.
Until the next Friday, where we will talk more about Worldwide Rush development. Also, we will share the demo beta access to those who will want to test out all the new systems that will come with the next update.