Expect a much longer than usual gap between n18 and n19. I want to bring Endurance back and I'm finding myself bogged down a bit fixing and adjusting random things. Therefore I will write out a plan for what I'm attacking and the order in which I will do it.
The general outline is: - Raft of improvements to singleseater AI to finally clear out the roughness. They are very good where they are right now, but they are still missing a few things here and there that would allow them to be closer to humans with regard to general strategy. Lots of things aren't consistent through the chain of methods that let them pick tyres, fuel, and desired pit laps. This should hopefully be the last time this AI goes through significant changes, unless a feature turns up some way down the line that requires it
- First iteration of the new Endurance AI, which will borrow a lot from singleseater AI but will need its own set of strategies. The aim is to make Endurance *playable but not perfect*, then it will follow the same thing of me making new iterations every so often to clean them up and make them a force who can compete with, and beat, the player
- There will be *no new features* in n19, only these two items
- I will make updates to n18 if there are critical issues.
Quick update on what I have been doing since I posted the above:
- Created various utility methods to allow every part of the AI strategy choices to use the same values. No more random offsets here and there to maintain consistency, it's all one source of truth now.
- Rewrote/refactored/reworked everything the AI use for their strategy - tyre selection, fuel load, strategy selection, and strategy alteration. So far I have got 40% tanks and 80% tanks finished and tested, I am happy with where they are. I also got fired from my team for poor results, read into that what you will. I will be starting no refuelling this weekend.
- Rewrote the cornering speed method, both to refactor it, but also clear out roughness and make the cars act a little more naturally.
- Rewrote the avoidance logic so that they are more able to avoid cars that they view as an obstacle. I also changed how that system works - it looks for any car with a critical problem and prioritises avoiding them before anything else, then looks to the closest car to them physically only.
- Fixed a couple of issues with AI qualifying run logic that made it flawed, much improved from where it was before. Also improved the yielding in qualifying for cars who are on a flying lap. I will still leave collisions disabled as I still spotted a couple of cars getting stuck on others even after improving things, see it as a necessary evil.
Took some time away from the mod over the past few weeks, wanted to allow any issues to filter out from the test build I put on Discord, and I was sick of looking at MM after finding some balancing issues with overtake rolls, and after having spent every hour I was not at work staring at a screen and playtesting MM for a month and a half solid.
Yesterday I made some major progress in expanding the grid with regards to the number of cars on track at one time, specifically for Endurance so those 12 and 24 hour races will not feel so empty. I made a post on the Motorsport Manager subreddit if anyone cares to look with a screenshot of ERS with a 40 car starting grid. I played the whole race and everything ran fine, but it does become more intense for your PC, so I will have expanded grids as an optional set of assets once Endurance returns. I also wanted larger grids for the moment in the future that I also revived GT, so it will also be used there.
Endurance qualifying logic is proving to be a bit of a headache where two drivers need to be considered, mainly around changeable weather because you may only end up with time for one dry run, for example. Race strategy will be easy to do, just a few hour's work. There is also time cost and car performance balancing to be done, which will take a little while to do.
Thanks for the mod! Do I have to have all the DLCs to play this mod? Because the requirement "Rebirth: Logo Pack" on Steam is said to need all the DLCs, and I don't have all of them?
Firstly, I chose the start that gives you a car with 3 stars in every field. With that car, and some luck, I managed to somehow eke out 7th of 12 in the first season on Make Your Team.
Second season came around and it turns out I'm only allowed to buy 0-star parts for my chassis. Now I'm struggling to even get 10th of 12, but the chairman thinks I should be getting at least 8th since I got 7th last seaspon, and it's clear that I'm going to get fired and lose the save.
Secondly, my drivers are constantly bleeding morale because my car sucks. I don't seem to be able to do anything about this except keep making new parts in an attempt to catch up, but they're losing 20+ morale every race, and it's not looking like that's going to stop for another season or two if I can even get that far.
Maybe I'm doing something wrong? I feel like my biggest mistake was getting 7th in the first season, that a more optimal play would have been to sabotage myself so that the chairman doesn't have unreasonable expectations next season.
The chassis selection is currently governed by your player Financial stat, which means that for the first several seasons you will have access to poor options. It is a rather unpopular feature that was carried over from the original Rebirth mod, there is a new system that is being created that I will merge into n19 whenever I finish it.
Good drivers for your series will care more about the quality, or lack thereof, of the car they drive. For example, in ERS, three star drivers and up will care deeply about the quality of the car they drive, while those at two stars and below will largely be happy to be there and will not care so much. It is a balance of getting decent drivers in when you have the car to support them, or accept that you will have a revolving door of drivers who simply use you as a stepping stone in their career.
I'm quite enjoying this mod so far. Doing a CYOT from scratch (in ERS). First season was not so awful as I thought. I am now in the second season, which is not better than the previous (mainly because I could not choose any supplier, since they weren't deeming doing business with me). I still manage to get some points.
But today I encountered something unexpected. I was having a race in Milan, with 40% rain probability, from what my team told me. Since it's a 33 laps race (with 40% fuel), I deem a good strategy to make 2 stops with 11 laps stints.
What struck me as odd after 14 laps is that I'm the only one who has made stops. And suddenly, rain starts to drop. Then, they all go in the pits to change their tyres, as if every team knew exactly when the rain would start.
Then it rains for quite a while. I have to stop for refuelling (seeing how much it rained, I guessed it would be a wet race until the end), but rain unexpectedly stopped 5 laps before the end of the race, and again, everyone seemed to know they had to stop 3 laps before the end to change tyres and refuel.
EVERY… SINGLE… TEAM… KNEW… (sorry for the caps, but that strikes me as odd 🤣).
Is there a bug? I pretty much doubt that every one of my competing teams have their forecast center fully upgraded. So maybe there's something wrong in the IA strategy calculation?
Long answer - I have been having a long-term internal battle over how I want to handle wet weather going forward. As it stands, right now, I take the view that if the player can see the track wetness, then it is only fair that the AI can as well. My intention/desire is to explore removing the track wetness rollover (so you cannot see the future wetness) but leave the weather forecast in. Then build in mechanics to make the AI both able to see the track wetness for the purposes of the wider strategy, but force offsets to create the illusion of other teams misjudging the weather, as the player will also have to make judgements based on the forecast and having the AI 'get it wrong' will give the player some leeway.
In the meantime, I was also toying with the idea of only giving the AI the lookahead that matched their Forecast Centre, to match the player, but I quickly realised that players would simply work out who could see the furthest and simply copy them, thus putting other teams at a disadvantage.
The wall I have hit several times is that human intelligence will simply outsmart AI because all they can do is follow the set of instructions I give them (you're essentially playing against me at this point), and finding the right way to implement AI options can take some time. I know AI 'cheating' can be a turn off for some people, myself included, but I want to find a fair way to implement the wet weather experience without needlessly penalising one party. As the AI are likely to drop the ball far more often than the player, I am willing to let them have this one for the time being.
If I can give you my advice : Try to be as fair as possible.
Sure, the player can try to copy the AI strategy. But you don't know how much fuel they've loaded. But the player has a much more unfair advantage: He can run the race, reload before the race and adapt his strategy. There is no way the AI can win this game.
But you may consider that if the AI knows something that the player cannot know, it's an unfair advantage (I mean, if he cannot know while having the same buildings, etc…).
Well, it's my vision of things. You are the lead director of this project 😉
Yes, the player can save scum - but a player who wishes to save scum should uninstall my mod and play something easier. Vanilla has the guaranteed gratification they are looking for, not Redux. If they are not prepared to take the losses then this is not the mod for them.
As I said, the AI knowing everything is the middle ground of the available options when factoring in people intelligence against the AI where they are right now. Making them more fuzzy on it requires other work to be done first, and as it stands I am a solo developer working almost every hour of his free time and I still have a large list of things to complete, and it all needs to be completed in the right order.
And you can still see how much fuel the AI have through the data centre during races, so if you really wanted to you could simply check their fuel after they pass the start/finish line and copy that.
Do not ever think I am being unfair, I am simply working with the tools I have and have set things up in a certain way for now because that is how they will function best until I can focus on changing it. The way I am releasing updates means the AI is evolving every time and becoming more capable. There will be some help for them until they have the proper logic in place because they need to be functional across all areas of race strategy.
If I make them fuzzy on strategy now I am handing success to the player, and I refuse to do so.
Hello again, I went from F3 to F2 in n17. But I had the worst car, I didn't get a single point for the season (more teams didn't get a point) and I was relegated to F3 again. I was losing so much that there was nothing I could do about it, plus only the first eight vots are scored. Now I've been promoted to F2 again and the situation is repeating itself. Is this solved in the n18 version? Do you plan to address this or do I need to improve my gameplay? Thanks a lot!
Trying my best not to sound rude, but you need to improve. The AI is relatively unforgiving and will make good strategy choices most of the time, but the place where the player will be superior for some time to come is in part development (as long as the AI difficulty slider is at a low value). I am making the AI even more robust for n19, so they will be even more difficult to score easy results against. Strong AI is one of the main parts of this mod and I will not be turning them down. I playtest on a relatively high difficulty (currently playing on an unfair difficulty) and when I succeed in a fashion I feel wasn't challenging enough I look to correct it. I like my games difficult, and the struggle is just as critical to my enjoyment as the successes are.
Player demotion was almost impossible in vanilla unless you actually tried to be bad, and it might take some getting used to for some players that they are now vulnerable to demotion.
Some tips for now - reverse grids in APS can help, especially with the new catchup system, and getting tracks with Extreme rated overtaking difficulty would then work in the favour of the slower cars, mostly at the beginning of each performance cycle. As of n18 the AI are weaker on 40% fuel tanks due to gaps in their strategy logic, so you can use that to your advantage.
Hi, thanks a lot for the mod! I'm sure it was very challenging. I'd like to point out a few bugs I found, hopefully that will help you: Attached images - Driver loses abilities even though he is very young - the newly developed parts do not match the predicted performance (in the picture the new part has even less performance) I don't know what to do, after 5 seasons I'm still 11/12 team with the car and I'm developing it to the maximum. I do have progressively more suppliers to choose from before the new season, but they are still terrible. Will that change? What can I do to influence this? Thanks to anyone for any suggestions! Translated with DeepL.com (free version)
1) Redux includes traits that will reduce a driver's Improvability stat, and this can make it negative. If the trait is temporary they will eventually return to their normal rate of improvement, but if it's permanent then that driver will never recover.
2) Player parts that are built during the season now have their performance reduced if the AI Difficulty slider in "Preferences" is greater than 0. I have recently swapped to this system because the previous method of keeping the player from overperforming was causing the field to spread out too much. Try reducing the slider to 0 and see how it plays, then increase it if you want more difficulty keeping up. CYOT has no facilities, so you are rightly suffering because you cannot create the same quality parts as the better teams. There is still some work to be done to properly convey this feature to the player in-game.
3) Supplier availability starts to open up mainly with your player's Financial stat, as long as you keep a positive cash balance that stat will improve and open up more suppliers. This system is in the process of being changed to something more realistic, so I am just waiting on that to be finished before I throw out this old system and replace it with a better one.
Endurance is in my plans to bring back early this year, it was the whole reason the day night cycle was implemented and I hate not having it around. Most of the core AI work is done for race day now, so (finger's crossed) it's a case of adapting that code over to Endurance.
GT is still in no man's land - it needs to not be single seaters with a roof, but something that stands out on its own. The amount of work to get it to that point means it is at the bottom of the list of things to be done. As a solo developer I have to pick my battles carefully, and GT is not one I'm ready to fight.
There is a bug where the AI teams constantly stops at the pitstop. The AI teams are constantly being repaired during race and don't repair the car. I can overall win the race by just not crashing with a class B car, because same thing happens in all class A and B cars.
As a trade off I did add the feature where you start off a session with some practice knowledge already if your driver is experienced enough, with an extra bonus if they have been with their current mechanic for over a year.
With experienced drivers and long term mechanics, both of my drivers have very small green bars after one run which make it easy to get the setup right.
The whole ethos of Redux is to get the player to earn what they have.
64 comments
The general outline is:
- Raft of improvements to singleseater AI to finally clear out the roughness. They are very good where they are right now, but they are still missing a few things here and there that would allow them to be closer to humans with regard to general strategy. Lots of things aren't consistent through the chain of methods that let them pick tyres, fuel, and desired pit laps. This should hopefully be the last time this AI goes through significant changes, unless a feature turns up some way down the line that requires it
- First iteration of the new Endurance AI, which will borrow a lot from singleseater AI but will need its own set of strategies. The aim is to make Endurance *playable but not perfect*, then it will follow the same thing of me making new iterations every so often to clean them up and make them a force who can compete with, and beat, the player
- There will be *no new features* in n19, only these two items
- I will make updates to n18 if there are critical issues.
- Created various utility methods to allow every part of the AI strategy choices to use the same values. No more random offsets here and there to maintain consistency, it's all one source of truth now.
- Rewrote/refactored/reworked everything the AI use for their strategy - tyre selection, fuel load, strategy selection, and strategy alteration. So far I have got 40% tanks and 80% tanks finished and tested, I am happy with where they are. I also got fired from my team for poor results, read into that what you will. I will be starting no refuelling this weekend.
- Rewrote the cornering speed method, both to refactor it, but also clear out roughness and make the cars act a little more naturally.
- Rewrote the avoidance logic so that they are more able to avoid cars that they view as an obstacle. I also changed how that system works - it looks for any car with a critical problem and prioritises avoiding them before anything else, then looks to the closest car to them physically only.
- Fixed a couple of issues with AI qualifying run logic that made it flawed, much improved from where it was before. Also improved the yielding in qualifying for cars who are on a flying lap. I will still leave collisions disabled as I still spotted a couple of cars getting stuck on others even after improving things, see it as a necessary evil.
I have released a version of n19 on the Discord for testing so we can clear out any bugs or oversights in the core AI strategy.
In the meantime, I will be making some adjustments to AI reactions to SC and VSC calls, and I have begun the planning stage of Endurance.
As I am a solo developer and an amateur programmer, everything takes a bit of time, so please bear with me.
Took some time away from the mod over the past few weeks, wanted to allow any issues to filter out from the test build I put on Discord, and I was sick of looking at MM after finding some balancing issues with overtake rolls, and after having spent every hour I was not at work staring at a screen and playtesting MM for a month and a half solid.
Yesterday I made some major progress in expanding the grid with regards to the number of cars on track at one time, specifically for Endurance so those 12 and 24 hour races will not feel so empty. I made a post on the Motorsport Manager subreddit if anyone cares to look with a screenshot of ERS with a 40 car starting grid. I played the whole race and everything ran fine, but it does become more intense for your PC, so I will have expanded grids as an optional set of assets once Endurance returns. I also wanted larger grids for the moment in the future that I also revived GT, so it will also be used there.
Endurance qualifying logic is proving to be a bit of a headache where two drivers need to be considered, mainly around changeable weather because you may only end up with time for one dry run, for example. Race strategy will be easy to do, just a few hour's work. There is also time cost and car performance balancing to be done, which will take a little while to do.
Simple answer is I am not sure, I have never not had all the DLCs, all you can do is try and see what happens.
Firstly, I chose the start that gives you a car with 3 stars in every field. With that car, and some luck, I managed to somehow eke out 7th of 12 in the first season on Make Your Team.
Second season came around and it turns out I'm only allowed to buy 0-star parts for my chassis. Now I'm struggling to even get 10th of 12, but the chairman thinks I should be getting at least 8th since I got 7th last seaspon, and it's clear that I'm going to get fired and lose the save.
Secondly, my drivers are constantly bleeding morale because my car sucks. I don't seem to be able to do anything about this except keep making new parts in an attempt to catch up, but they're losing 20+ morale every race, and it's not looking like that's going to stop for another season or two if I can even get that far.
Maybe I'm doing something wrong? I feel like my biggest mistake was getting 7th in the first season, that a more optimal play would have been to sabotage myself so that the chairman doesn't have unreasonable expectations next season.
Good drivers for your series will care more about the quality, or lack thereof, of the car they drive. For example, in ERS, three star drivers and up will care deeply about the quality of the car they drive, while those at two stars and below will largely be happy to be there and will not care so much. It is a balance of getting decent drivers in when you have the car to support them, or accept that you will have a revolving door of drivers who simply use you as a stepping stone in their career.
I am now in the second season, which is not better than the previous (mainly because I could not choose any supplier, since they weren't deeming doing business with me). I still manage to get some points.
But today I encountered something unexpected. I was having a race in Milan, with 40% rain probability, from what my team told me.
Since it's a 33 laps race (with 40% fuel), I deem a good strategy to make 2 stops with 11 laps stints.
What struck me as odd after 14 laps is that I'm the only one who has made stops. And suddenly, rain starts to drop. Then, they all go in the pits to change their tyres, as if every team knew exactly when the rain would start.
Then it rains for quite a while. I have to stop for refuelling (seeing how much it rained, I guessed it would be a wet race until the end), but rain unexpectedly stopped 5 laps before the end of the race, and again, everyone seemed to know they had to stop 3 laps before the end to change tyres and refuel.
EVERY… SINGLE… TEAM… KNEW… (sorry for the caps, but that strikes me as odd 🤣).
Is there a bug? I pretty much doubt that every one of my competing teams have their forecast center fully upgraded. So maybe there's something wrong in the IA strategy calculation?
Long answer - I have been having a long-term internal battle over how I want to handle wet weather going forward. As it stands, right now, I take the view that if the player can see the track wetness, then it is only fair that the AI can as well. My intention/desire is to explore removing the track wetness rollover (so you cannot see the future wetness) but leave the weather forecast in. Then build in mechanics to make the AI both able to see the track wetness for the purposes of the wider strategy, but force offsets to create the illusion of other teams misjudging the weather, as the player will also have to make judgements based on the forecast and having the AI 'get it wrong' will give the player some leeway.
In the meantime, I was also toying with the idea of only giving the AI the lookahead that matched their Forecast Centre, to match the player, but I quickly realised that players would simply work out who could see the furthest and simply copy them, thus putting other teams at a disadvantage.
The wall I have hit several times is that human intelligence will simply outsmart AI because all they can do is follow the set of instructions I give them (you're essentially playing against me at this point), and finding the right way to implement AI options can take some time. I know AI 'cheating' can be a turn off for some people, myself included, but I want to find a fair way to implement the wet weather experience without needlessly penalising one party. As the AI are likely to drop the ball far more often than the player, I am willing to let them have this one for the time being.
If I can give you my advice : Try to be as fair as possible.
Sure, the player can try to copy the AI strategy. But you don't know how much fuel they've loaded.
But the player has a much more unfair advantage: He can run the race, reload before the race and adapt his strategy. There is no way the AI can win this game.
But you may consider that if the AI knows something that the player cannot know, it's an unfair advantage (I mean, if he cannot know while having the same buildings, etc…).
Well, it's my vision of things. You are the lead director of this project 😉
As I said, the AI knowing everything is the middle ground of the available options when factoring in people intelligence against the AI where they are right now. Making them more fuzzy on it requires other work to be done first, and as it stands I am a solo developer working almost every hour of his free time and I still have a large list of things to complete, and it all needs to be completed in the right order.
And you can still see how much fuel the AI have through the data centre during races, so if you really wanted to you could simply check their fuel after they pass the start/finish line and copy that.
Do not ever think I am being unfair, I am simply working with the tools I have and have set things up in a certain way for now because that is how they will function best until I can focus on changing it. The way I am releasing updates means the AI is evolving every time and becoming more capable. There will be some help for them until they have the proper logic in place because they need to be functional across all areas of race strategy.
If I make them fuzzy on strategy now I am handing success to the player, and I refuse to do so.
thanks for your dedicated work on this game.
My question might have been answered before, yet i didnt find it:
Is the mod compatible with "real drivers/logos" mods?
I went from F3 to F2 in n17. But I had the worst car, I didn't get a single point for the season (more teams didn't get a point) and I was relegated to F3 again. I was losing so much that there was nothing I could do about it, plus only the first eight vots are scored. Now I've been promoted to F2 again and the situation is repeating itself. Is this solved in the n18 version? Do you plan to address this or do I need to improve my gameplay?
Thanks a lot!
Player demotion was almost impossible in vanilla unless you actually tried to be bad, and it might take some getting used to for some players that they are now vulnerable to demotion.
Some tips for now - reverse grids in APS can help, especially with the new catchup system, and getting tracks with Extreme rated overtaking difficulty would then work in the favour of the slower cars, mostly at the beginning of each performance cycle. As of n18 the AI are weaker on 40% fuel tanks due to gaps in their strategy logic, so you can use that to your advantage.
Attached images
- Driver loses abilities even though he is very young
- the newly developed parts do not match the predicted performance (in the picture the new part has even less performance)
I don't know what to do, after 5 seasons I'm still 11/12 team with the car and I'm developing it to the maximum.
I do have progressively more suppliers to choose from before the new season, but they are still terrible. Will that change? What can I do to influence this?
Thanks to anyone for any suggestions!
Translated with DeepL.com (free version)
2) Player parts that are built during the season now have their performance reduced if the AI Difficulty slider in "Preferences" is greater than 0. I have recently swapped to this system because the previous method of keeping the player from overperforming was causing the field to spread out too much. Try reducing the slider to 0 and see how it plays, then increase it if you want more difficulty keeping up. CYOT has no facilities, so you are rightly suffering because you cannot create the same quality parts as the better teams. There is still some work to be done to properly convey this feature to the player in-game.
3) Supplier availability starts to open up mainly with your player's Financial stat, as long as you keep a positive cash balance that stat will improve and open up more suppliers. This system is in the process of being changed to something more realistic, so I am just waiting on that to be finished before I throw out this old system and replace it with a better one.
I have 2k hours on this game, played a lot of mods, but this is sure the best I have played.
Any plans to get it also working again with GT and Endurance? I would love to be able to play them too.
GT is still in no man's land - it needs to not be single seaters with a roof, but something that stands out on its own. The amount of work to get it to that point means it is at the bottom of the list of things to be done. As a solo developer I have to pick my battles carefully, and GT is not one I'm ready to fight.
Endurance is 100% dead as it stands right now because I removed a lot of AI logic which is waiting for me to rewrite it.
1) Assembly-CSharp.dll & Assembly-CSharp-firstpass need to be pasted into ...\steamapps\common\Motorsport Manager\MM_Data\Managed
2) resources.assets needs to be pasted into ...\steamapps\common\Motorsport Manager\MM_Data
Honestly, this is the only feature I don't like.
With experienced drivers and long term mechanics, both of my drivers have very small green bars after one run which make it easy to get the setup right.
The whole ethos of Redux is to get the player to earn what they have.