20 Pages<<1415161718>>PrevNext
Rank: D-Class Racing License
 1 user liked this post.
#376 Posted : Monday, February 24, 2020 5:28:40 AM(UTC)
Originally Posted by: BriskArtist7600 Go to Quoted Post
1.5 year after launch - buggies class still broken on higher diff. levels. ZERO response from the developers so far. I mean come on? Everyone knows the RZR Polaris is completely BROKEN in it's current state and can't be beaten on certain tracks (mostly track that have few props).
At least stop adding Offroad Buggies to seasonal championships.


Every time there are offroad buggies races in the playlist, I bring the maverick or any cvt buggy. It evens the field.
Rank: B-Class Racing License
#377 Posted : Monday, February 24, 2020 9:01:42 AM(UTC)
I think I figured out what is going on with unbeatables. They're all AWD swapped. I just tuned this fairly stock ( awd swapped and suspension maxed ) 650S, and I can beat unbeatables with no issues. If I leave it RWD I get stomped.

They also play almost exactly like people do online so if you have a problem with drivatar driving, you're at fault.

Something else I've noticed is that if you beat unbeatables enough, the game retaliates. They start getting more aggressive. It's annoying.

Edited by user Monday, February 24, 2020 9:53:13 AM(UTC)  | Reason: Not specified

Rank: S-Class Racing License
#378 Posted : Monday, February 24, 2020 9:07:59 AM(UTC)
Originally Posted by: RashestQuasar99 Go to Quoted Post
Originally Posted by: BriskArtist7600 Go to Quoted Post
1.5 year after launch - buggies class still broken on higher diff. levels. ZERO response from the developers so far. I mean come on? Everyone knows the RZR Polaris is completely BROKEN in it's current state and can't be beaten on certain tracks (mostly track that have few props).
At least stop adding Offroad Buggies to seasonal championships.


Every time there are offroad buggies races in the playlist, I bring the maverick or any cvt buggy. It evens the field.


I've never lost offroad buggies to Ai, because the Ai take wide turns, and I can undercut them every time. If you time your slide right, and hit the inside of the checkpoints you get a massive lead over them. Also the Ai don't jump very well even though somebody said that they do. I usually jump over them.
Rank: A-Class Racing License
#379 Posted : Monday, February 24, 2020 1:18:11 PM(UTC)
Originally Posted by: teksenic Go to Quoted Post
I think I figured out what is going on with unbeatables. They're all AWD swapped. I just tuned this fairly stock ( awd swapped and suspension maxed ) 650S, and I can beat unbeatables with no issues. If I leave it RWD I get stomped.

They also play almost exactly like people do online so if you have a problem with drivatar driving, you're at fault.

Something else I've noticed is that if you beat unbeatables enough, the game retaliates. They start getting more aggressive. It's annoying.

AWD is easier if you can't drive. Sure, you can beat them with AWD, but with RWD too. They don't need AWD either.

I don't think they play like anybody, maybe a bit, but mostly not.

Some mood swings are there but hard to say what causes it.
Rank: C-Class Racing License
#380 Posted : Monday, February 24, 2020 2:57:46 PM(UTC)
Originally Posted by: turtleCZ Go to Quoted Post
Originally Posted by: teksenic Go to Quoted Post
I think I figured out what is going on with unbeatables. They're all AWD swapped. I just tuned this fairly stock ( awd swapped and suspension maxed ) 650S, and I can beat unbeatables with no issues. If I leave it RWD I get stomped.

They also play almost exactly like people do online so if you have a problem with drivatar driving, you're at fault.

Something else I've noticed is that if you beat unbeatables enough, the game retaliates. They start getting more aggressive. It's annoying.

AWD is easier if you can't drive. Sure, you can beat them with AWD, but with RWD too. They don't need AWD either.

I don't think they play like anybody, maybe a bit, but mostly not.

Some mood swings are there but hard to say what causes it.


Teksenik, you posted on Feature Request topic about your experience with '08 BMW M3. I run test runs with it in both Road Racing and Street Scene events. I didn't make a new post but edited old my old reply. I think your experience with BMW could be debated further here.

So I run that BMW bone stock through every Road and Street event in game. I was able to win every event with it but struggled really bad at the Monument Wynds, Scenario was against AWD Audi starting from front of starting grid. I used only default blueprints for season. I screen capped results and they can be seen here, first for Street Scene and further back Road events. Note that there are multiple captures from Monument Wynds as it looked like I might be able to win the race but not necessarily beat the best time by fastest Drivatar. That turned out to be the case. My wining time captured here and best Drivatar time here. I think I needed about 10 races to win it but I only captured 5 shots from there total, which should present fastest, slowest and thus give points to rubber band / force field bracket speed dynamics relative to player performance.

What I would need is some players to take that bone stock BMW and run it against Drivatar lineup front lining that Audi and see how many races it takes to win and screen cap results. That is so we could decide if there's something worth submitting a ticket. I run manual, no TSC, ABS, etc, Sim-steering, default controller dead zones with controller that came with my Xbox One X.

My motivation is improving the product if there's something broken on that route and create opportunities for better peer support. Who is fastest or most hardcore player, I don't want anything to do with that. This is peer support topic and here we could have a chance to perhaps discover something that would matter, not just for us but every one enjoying Horizon.

You may notice that there are over 1000 screencaps at my TrueAchievements accounts. While there are some captures from Playlist events, most are data from my Street Scene test. All versus Unbeatable with same settings described above. I originally planned to make nice summary, send it to Retro with link to screenshots as proof, but now we don't have a community liaison anymore.

So when I write about Unbeatable being actually beatable, I don't talk about based to my opinion or some sort feeling but based on data, hard evidence. I have been sitting on that for couple of months actually. Past couple of weeks, while posting here all sort of stuff, I have been trying to figure out what do I do with all that data now.

Edited by user Monday, February 24, 2020 2:59:12 PM(UTC)  | Reason: Fixed link

Rank: S-Class Racing License
#381 Posted : Monday, February 24, 2020 4:03:49 PM(UTC)
Originally Posted by: ClawedScroll554 Go to Quoted Post
Originally Posted by: turtleCZ Go to Quoted Post
Originally Posted by: teksenic Go to Quoted Post
I think I figured out what is going on with unbeatables. They're all AWD swapped. I just tuned this fairly stock ( awd swapped and suspension maxed ) 650S, and I can beat unbeatables with no issues. If I leave it RWD I get stomped.

They also play almost exactly like people do online so if you have a problem with drivatar driving, you're at fault.

Something else I've noticed is that if you beat unbeatables enough, the game retaliates. They start getting more aggressive. It's annoying.

AWD is easier if you can't drive. Sure, you can beat them with AWD, but with RWD too. They don't need AWD either.

I don't think they play like anybody, maybe a bit, but mostly not.

Some mood swings are there but hard to say what causes it.


Teksenik, you posted on Feature Request topic about your experience with '08 BMW M3. I run test runs with it in both Road Racing and Street Scene events. I didn't make a new post but edited old my old reply. I think your experience with BMW could be debated further here.

So I run that BMW bone stock through every Road and Street event in game. I was able to win every event with it but struggled really bad at the Monument Wynds, Scenario was against AWD Audi starting from front of starting grid. I used only default blueprints for season. I screen capped results and they can be seen here, first for Street Scene and further back Road events. Note that there are multiple captures from Monument Wynds as it looked like I might be able to win the race but not necessarily beat the best time by fastest Drivatar. That turned out to be the case. My wining time captured here and best Drivatar time here. I think I needed about 10 races to win it but I only captured 5 shots from there total, which should present fastest, slowest and thus give points to rubber band / force field bracket speed dynamics relative to player performance.

What I would need is some players to take that bone stock BMW and run it against Drivatar lineup front lining that Audi and see how many races it takes to win and screen cap results. That is so we could decide if there's something worth submitting a ticket. I run manual, no TSC, ABS, etc, Sim-steering, default controller dead zones with controller that came with my Xbox One X.

My motivation is improving the product if there's something broken on that route and create opportunities for better peer support. Who is fastest or most hardcore player, I don't want anything to do with that. This is peer support topic and here we could have a chance to perhaps discover something that would matter, not just for us but every one enjoying Horizon.

You may notice that there are over 1000 screencaps at my TrueAchievements accounts. While there are some captures from Playlist events, most are data from my Street Scene test. All versus Unbeatable with same settings described above. I originally planned to make nice summary, send it to Retro with link to screenshots as proof, but now we don't have a community liaison anymore.

So when I write about Unbeatable being actually beatable, I don't talk about based to my opinion or some sort feeling but based on data, hard evidence. I have been sitting on that for couple of months actually. Past couple of weeks, while posting here all sort of stuff, I have been trying to figure out what do I do with all that data now.


You can't say something is broken if Unbeatable is unbeatable. It is all the other tracks that are broken... lol. To be more precise it is not a valid complaint to a game designer that their unbeatable Ai are unbeatable. If you want something improving in the game it would be that the Ai behave more like real people, and can drift more realistically. That can happen when the Xbox speeds up in the future. At the moment the Ai have restricted physics to keep the frame rate high.
Rank: A-Class Racing License
#382 Posted : Tuesday, February 25, 2020 3:30:44 AM(UTC)
This thread is a pure gold! I started to read it from start and so much info there. I came to the party a bit late but at least I really started to play seriously with Horizon 4. Racing n00b :D Where are those researchers now? They moved to other games?
Rank: C-Class Racing License
 1 user liked this post.
#383 Posted : Tuesday, February 25, 2020 12:11:52 PM(UTC)
Originally Posted by: AquaPainter168 Go to Quoted Post


You can't say something is broken if Unbeatable is unbeatable. It is all the other tracks that are broken... lol. To be more precise it is not a valid complaint to a game designer that their unbeatable Ai are unbeatable. If you want something improving in the game it would be that the Ai behave more like real people, and can drift more realistically. That can happen when the Xbox speeds up in the future. At the moment the Ai have restricted physics to keep the frame rate high.


I don't believe this to be so. This could be valid point, if it weren't but one route and in very narrow conditions (AWD starting from first row when player car is RWD in 5th. row).

For Road and Street events Purpose of "Unbeatable" AI is two folded.
1. Developers choice to use word "Unbeatable" for highest difficulty level, is clearly for psychological effect and not to be interpreted literally.
2. Highest difficulty works like all other difficulties, presenting challenge for player that pushes his/her playing skills and so expands player skill envelope.

Game is very consistent in this. Unbeatable is beatable in its all 26 Road Racing events and 18 Street Scene events out of 21, Monument Wynds being inconsistency and only in very specific circumstances, described above. I exclude Ambleside Ascent and Otleydale Dash at least for now, because I need to revisit my notes to see if win percentages are conclusive or not.

This one event, although only in specific scenario presents 2,27% of 44. So it's inconstant with 97,73 % tarmac events. I think it's very clear that we are not talking about something that is intended design, but an anomaly and appealing to choice of word "Unbeatable" is just semantic excuse, I can see programmer could use though. This is why we need managers, but I digress.

In 97,73% of tarmac events winning the race is a matter of player skill and in 2,27% it's up to random factor, luck, depending of what Drivatar lineup game draws from possible combinations, it's IMO very clear that feature is not working as intended and thus confusing players, it's clearly bugged.




What comes to AI behaving more like real people, certain Drivatar behaviour is curiously similar to tactics used in real racing with factory teams. I don't know what term was used globally, if there even was one, but in the US one pilot in team was driving as "Hare". Purpose of Hare was fast start, get lead, push as fast possible to wear competing teams drivers and technology down. This what Ford did in the '60s at least.
Rank: A-Class Racing License
#384 Posted : Saturday, February 29, 2020 4:11:38 PM(UTC)
I am half way through the reading this thread. Many ppl think it's cheating but for some perspective it's not that bad. I like road racing on unbeatable even if it's not racing, it's about passing everybody and racing good enough to have a chance. It's not much wrong with the AI and maybe that's the reason why it stayed as it is. I rather like AI how it is even if it's not real racing. New consoles has more power so maybe new AI will be there too.
Rank: A-Class Racing License
#385 Posted : Wednesday, March 4, 2020 6:45:40 AM(UTC)
Finally I read everything. It's a good thread but many people don't know how the AI works. I think it could be a good place here to research more how AI works.

I am happy with current design because it works for my needs. Are there some strange situations? Yeah, for sure. The most strange is asking you if you want a harder challenge. You don't want it?! Come on, you will get it anyway! It's so funny...
Rank: A-Class Racing License
#386 Posted : Wednesday, March 4, 2020 1:27:51 PM(UTC)
I maybe saw the force fields :D I went straight and bang I was out of road immediately :D Hard to say what it was but it must the field!
Rank: C-Class Racing License
#387 Posted : Friday, March 6, 2020 4:15:51 AM(UTC)
I seriously cannot believe this thread is still going.

Perfect Dark had it's hardest AI doing things humans could never achieve, and I heard no complaints. They would run twice as fast, shoot faster, easily clap you across the map. Working as intended.

If anything you should be flattered to have the ability to beat an AI that cheats.

Edited by user Friday, March 6, 2020 4:18:16 AM(UTC)  | Reason: Not specified

Rank: S-Class Racing License
#388 Posted : Friday, March 6, 2020 5:53:51 AM(UTC)
Originally Posted by: SuperHornetA51 Go to Quoted Post
I seriously cannot believe this thread is still going.

Perfect Dark had it's hardest AI doing things humans could never achieve, and I heard no complaints. They would run twice as fast, shoot faster, easily clap you across the map. Working as intended.

If anything you should be flattered to have the ability to beat an AI that cheats.


Well you just mentioned Perfect Dark on a site not even connected to it so it's failures are spreading everywhere.
Rank: C-Class Racing License
 1 user liked this post.
#389 Posted : Friday, March 6, 2020 1:58:10 PM(UTC)
Had No traffic cars glitch happen today at Derwent Valley Dash Street Race. No idea what caused it but it resulted similar situation that I have encountered before, Drivatars got a speed boost. I couldn't even beat the fastest one, which was finishing at x.53's vs my best that was in x.54's. Did some other race with same car (JB Lotus Esprit stock) and went back, traffic was there again and no issues winning, Drivatar taking second position finishing in x.59's! So glitch impact in this case was ~6 seconds and that's in C-class. Was too lazy to screencap, so nothing more exact than that though.

In glitched race visually it looked like Drivatar top speed, acceleration and probably even grip was boosted, grip being the biggest issue as I couldn't catch it with sort of power glides you can do on tarmac. It didn't really need to slow down that much on any part of route.

Edited by user Friday, March 6, 2020 1:58:56 PM(UTC)  | Reason: Not specified

Rank: A-Class Racing License
#390 Posted : Friday, March 6, 2020 2:40:55 PM(UTC)
I don't think grip is important because of rubber banding. If they should be fast they will no matter what. They don't care about grip, speed, obstacles, etc. It's just a show for you. It's like different AI cars, it's not real.
Rank: S-Class Racing License
 1 user liked this post.
#391 Posted : Friday, March 6, 2020 4:54:40 PM(UTC)
Originally Posted by: turtleCZ Go to Quoted Post
I don't think grip is important because of rubber banding. If they should be fast they will no matter what. They don't care about grip, speed, obstacles, etc. It's just a show for you. It's like different AI cars, it's not real.


"The grip isn't real" gives the Ai an advantage when you slide. That's what people mean.
Rank: C-Class Racing License
#392 Posted : Saturday, March 7, 2020 11:52:29 AM(UTC)
Originally Posted by: turtleCZ Go to Quoted Post
I don't think grip is important because of rubber banding. If they should be fast they will no matter what. They don't care about grip, speed, obstacles, etc. It's just a show for you. It's like different AI cars, it's not real.


Remember we are talking about a glitch here, not ordinary race. When I rerun with without no traffic glitch everything worked like usually. I actually posted about same glitch earlier in this topic and you just wrote that you read entire topic. Were you a bit sleepy perhaps?

Scenarios where Drivatars practically ignore most of game worlds physics if they get far enough distance relative to player has been proved to consistently happen in Cross Country and only in Cross Country. What is interesting here is that whatever boost "no traffic" glitch is causing, it appears to be at least somewhat similar to boost as in CC but happens regardless to relative distance to player.

Rubber band / force field dynamic effect sets relative performance bracket for Drivatars. In normal Street/Road race and mostly what I have seen in Dirt Racing, Drivatars can boost a lot to catch up to player and sometimes overtake by exceeding Drivatar car PI value a bit, which gives bad impression, but in the end with good enough racing player wins. There are two factors for this: 1. General time bracket that is dynamic depending of player car PI and perhaps other factors and 2. player performance in race. So perceived rubber banding where Drivatars slow down and it looks like they do it to allow player to catch up, happens because Drivatars has to stay within certain time bracket first, and secondly to allow time window for player to win. So not only Drivatars get faster with each difficulty level up, but also time window for player opportunity win gets smaller and smaller with increased difficulty.

That was actually discussed about a year ago but unfortunately I don't recall much of that conversation being in this topic. What we didn't had was enough data to prove that. What is different now, is that now I have that data, about thousand screencaps.

What comes to Drivatars ignoring physics and cars being purely visual presentation and Drivatars always finishing in same time. Nope, doesn't work like that. This is very easy to verify at Ambleside Ascent. If you sideswipe a Drivatar in S-bends section and it gets on grass, it slows down and it doesn't recover. Actually it can finish with far slower time that its intended time bracket. Why does that happen, is that S-bends section is relatively close to end of race and Drivatar doesn't fall behind enough to catch up boost mechanic to activate, or even if it does it's too late, for player race is finished. I actually have screencaps of relative times from there and posted examples in other data thread, which I recall PM'n you a link.

Second good spot is at Derwent Reservoir Trail. After the hill and left turn over the bridge if player is fast enough it can be often seen how Drivatars start messing things up. Hitting walls, hitting traffic and that doesn't only affect their momentary speed, if leading car has bad enough accident, they negotiate new leading Drivatar from ones that is further behind, which becomes new fastest Drivatar, sometimes with quite insane time considering how much further back that one can then get to second place. That is example of catchup boost mechanic when it has time to react. Drivatars cannot however ignore physics, take shortcuts through fences without it affecting their speed, take curves at 100 mph. etc. once they get close enough to player.

For most part what is discussed in this topic, it doesn't really matter if Forza AI is called Rubber Band or Force Field it's semantics. But to understand why we have a good dynamic AI instead of bad dynamic AI it begins to matter. Aqua Painter's Force Field theory would explain why we see environmental effects to have influence to Drivatars, not only visually but their times, yet they are able to catch up quite fast once far enough behind. In bubble close to player all these hitting traffic, hitting wall, losing grip, have effect to Drivatar speeds, though not as much as for player. Outside player bubble force field (when applied properly) only limits their performance to stay within preset time targets for route and player car PI (possible other factors). Preset theory (that might be dynamically generated just pre-race) is supported by that even unbeatable AI happens to have upper limit. It's IMO more noticeable in Road Racing, but better you get with the car, after a while winning margin gets larger and larger.

So I don't know. It's good to read stuff I suppose. Sometimes I think it would perhaps contribute to peer support a lot if good racing lines, alternative racing lines, benefits of leaving behind some racing aids would be discussed here. Stuff that actually works you know. If there we were to tell that instead of trying to beat Drivatars in first few corners, learning to take all corners more or less well would help a lot, trying to beat your best time and stuff like that. Might make players who have questions, whom have puzzling experience, make them feel a bit better about the game and themselves. Yeah... if only everybody understood what they read, maybe this topic wouldn't be going on in circles. Don't you think?
Rank: A-Class Racing License
#393 Posted : Sunday, March 8, 2020 4:10:30 PM(UTC)
Originally Posted by: AquaPainter168 Go to Quoted Post
Originally Posted by: turtleCZ Go to Quoted Post
I don't think grip is important because of rubber banding. If they should be fast they will no matter what. They don't care about grip, speed, obstacles, etc. It's just a show for you. It's like different AI cars, it's not real.


"The grip isn't real" gives the Ai an advantage when you slide. That's what people mean.

Yeah, but it's something like time based. You need to meet the time and rubberbanding. They can pass you with unreal speed and then slow down to meet the criteria. Who cares they drive 450 km/h S1 class car when they slow down later.

Edited by user Sunday, March 8, 2020 4:12:47 PM(UTC)  | Reason: Not specified

Rank: A-Class Racing License
#394 Posted : Sunday, March 8, 2020 4:12:24 PM(UTC)
Originally Posted by: ClawedScroll554 Go to Quoted Post
Originally Posted by: turtleCZ Go to Quoted Post
I don't think grip is important because of rubber banding. If they should be fast they will no matter what. They don't care about grip, speed, obstacles, etc. It's just a show for you. It's like different AI cars, it's not real.


Remember we are talking about a glitch here, not ordinary race. When I rerun with without no traffic glitch everything worked like usually. I actually posted about same glitch earlier in this topic and you just wrote that you read entire topic. Were you a bit sleepy perhaps?

Sure, I skipped some things. You have too long texts for me because I am not even sure if I understand everything there :D Will read it first...
Rank: R-Class Racing License
#395 Posted : Sunday, March 8, 2020 6:44:26 PM(UTC)
Originally Posted by: SuperHornetA51 Go to Quoted Post
I seriously cannot believe this thread is still going.

Perfect Dark had it's hardest AI doing things humans could never achieve, and I heard no complaints. They would run twice as fast, shoot faster, easily clap you across the map. Working as intended.

If anything you should be flattered to have the ability to beat an AI that cheats.


So you finally accept the AI cheats? lol

Every arcade game is the same. Numbers 2 to 10 are scrubs and #1 is glued to your tail like a magnet. NFS has the exact same system as Forza Horizon. It's lazy difficulty.

The comparison with Perfect Dark is useless since a DarkSim could be exploited, even against a PerfectSim it was suicidal to engage them in the open, so you resort to cheap shots with OP guns like the charged Mauler. I beat all 30 challenges solo back in the day and camping the OP weapons/items was the key to victory against harder simulants. In Perfect Dark save for Challenge #16 you even have the radar so you can easily anticipate any simulant, from MeatSim to DarkSim, unless cloaked of course.

A racing game AI bolstered by rubberbanding on the other hand cannot be exploited since it'll be on your tail no matter how fast you are. The exploits you can use to defeat such AI are due to poor pathing, not flaws in its rubberbanding. I prefer to call it "magnetic AI" instead since when you put on a big enough lead the fastest car will slow down to the pace of the others... both in Forza and NFS.

Edited by user Sunday, March 8, 2020 6:47:29 PM(UTC)  | Reason: Not specified

May the forced induction be with you.

Alice >>>>>>>>>> Keira
Rank: A-Class Racing License
#396 Posted : Monday, March 9, 2020 4:54:35 AM(UTC)
It's true but I like it. It's just a show for me. I wouldn't call it cheating because if you have a good time and race cleanly you can win which is important for me. Sure, there are some broken situations but it's uncommon.
Rank: C-Class Racing License
#397 Posted : Monday, March 9, 2020 10:19:50 AM(UTC)
Originally Posted by: NightDriver7800 Go to Quoted Post

Every arcade game is the same. Numbers 2 to 10 are scrubs and #1 is glued to your tail like a magnet. NFS has the exact same system as Forza Horizon. It's lazy difficulty.

The comparison with Perfect Dark is useless since a DarkSim could be exploited, even against a PerfectSim it was suicidal to engage them in the open, so you resort to cheap shots with OP guns like the charged Mauler. I beat all 30 challenges solo back in the day and camping the OP weapons/items was the key to victory against harder simulants. In Perfect Dark save for Challenge #16 you even have the radar so you can easily anticipate any simulant, from MeatSim to DarkSim, unless cloaked of course.

A racing game AI bolstered by rubberbanding on the other hand cannot be exploited since it'll be on your tail no matter how fast you are. The exploits you can use to defeat such AI are due to poor pathing, not flaws in its rubberbanding. I prefer to call it "magnetic AI" instead since when you put on a big enough lead the fastest car will slow down to the pace of the others... both in Forza and NFS.


You saying you can't exploit them just tells me you don't know how to exploit them. You are allowed to ride the wall, use them as walls, drive next to them so they're forced to slow down to let you go ahead. Easy first place, never struggled with this AI.
Rank: C-Class Racing License
#398 Posted : Monday, March 9, 2020 10:39:50 AM(UTC)
Originally Posted by: turtleCZ Go to Quoted Post

Sure, I skipped some things. You have too long texts for me because I am not even sure if I understand everything there :D Will read it first...


I haven't really used my English that much for a decade or so. There's a lot I have forgotten from my vocabulary and my posts tend to get wordy. That's on me I admit.

Even then though, research requires certain things. Numbers without contexts doesn't necessarily mean anything and enable others interested to participate. Transparency is also a must, it's not about me, it's not about you, it's about anybody being able to verify results. In context of this discussion Car, PI (Setup), Location (also season, especially for Street Scene), Time Target, Difficulty. All must be there for others to be able to repeat it or in short form: BMW M3 '03 stock at The Colossus, Summer, 6:20.230, Unbeatable.. Then there are other things, possible driving aids, transmission setting auto/manual/manual with clutch, etc.

And it may look like a bit of work, but without baselines and transparency enabling others to verify results I wouldn't be posting data. I would be posting about my subjective opinion and that's not how research works.

Even when we enter into more speculative space, about how dynamic AI works and where we really don't have much real knowledge, we can distill bit and pieces of information by sticking in same principle as above, do repeats. Notice that "No traffic" glitch also have influence to Drivatar speeds. Gather a mass of data and notice that in general Drivatar AI isn't unfair at least on most Street and Road events and cases where projected goal for player performance is perhaps beyond human ability, are such a small percentage, that we could even ignore the whole matter if it's intended to be that way or not, and just claim that such inconsistency is not building towards good Horizon experience for anyone on basis that such inconsistency is confusing for players and makes peer support inconvenient (For example: it's not the car, it's you, except if it happened at route X with Drivatar lineup Y [this sort of talk belongs to real tech support, not peer support]).

So in the process we can make multiple discoveries and reach multiple goals. One goal is met when we can separate game mechanical aspect (Drivatars in general aren't unfair in sense that they are unwinnable) from psychological aspects (Racing, even winning against rubber band AI isn't always satisfying experience for everybody). Let's see what we can do with that.

Originally Posted by: NightDriver7800 Go to Quoted Post
Originally Posted by: SuperHornetA51 Go to Quoted Post
I seriously cannot believe this thread is still going.

Perfect Dark had it's hardest AI doing things humans could never achieve, and I heard no complaints. They would run twice as fast, shoot faster, easily clap you across the map. Working as intended.

If anything you should be flattered to have the ability to beat an AI that cheats.


So you finally accept the AI cheats? lol

Every arcade game is the same. Numbers 2 to 10 are scrubs and #1 is glued to your tail like a magnet. NFS has the exact same system as Forza Horizon. It's lazy difficulty.

[..]

A racing game AI bolstered by rubberbanding on the other hand cannot be exploited since it'll be on your tail no matter how fast you are. The exploits you can use to defeat such AI are due to poor pathing, not flaws in its rubberbanding. I prefer to call it "magnetic AI" instead since when you put on a big enough lead the fastest car will slow down to the pace of the others... both in Forza and NFS.


Okay here we have sample SuperHornet and Sample NightDriver. Let's call them A and B.

Sample A appears to have a meta approach to topic based on some sort of moral which is based on other game from different genre. Sample A appears not to be interested nor understand what is actually discussed here but just decided to post something totally irrelevant because for some reason sample A is irritated that there is this discussion in this topic. So useless meta.

Sample B answers to sample A but sample B has something relevant to add to discussion. At least from psychological aspect I mentioned earlier there are valid points IMO.

Note: I don't mean to belittle anyone. Just did what I did for demonstration purposes.

@NightDriver7800

By Forza do you mean FM7 or FH4? as for my experience there are major differences between the two in Drivatar behaviour.

In FM7 it's possible to lap Drivatars, very easy to repeat if you do oval races. It's also possible to build a meta car that you can "drive" oval races by setting rubber band (real one) around controller in a way that keeps the throttle trigger down and then just steer with stick a bit around those ovals and win with so many laps over fastest Drivatar even on Unbeatable, that it's ridiculous. Used to be popular among credit grinders.

What was gathered via telemetry FM7 Drivatars can boost their cars above PI limit in certain situations. I don't recall how much boost was but that it was some percentage of base PI. I haven't seen any evidence that FM7 Drivatars ignored game physics, but I don't know that much about that game.

That leaves us with FH4.

Only place where I have ever lapped a Drivatar was Castle Cross Country Circuit and it was not under normal circumstances. It was Trial event and there's a jump from sand dune to Bamburgh castle courtyard. One Drivatar must have messed up that jump, perhaps repeatedly. I rerun that Trial few times and was able to repeat it, but that time there was one team mate who had set up camp at dune jump and I think actively ramming Drivatars there so they would miss that jump. It was very effective tactic.

AFAIK one effect of FH4 "rubber band" is that under normal circumstances Drivatars can't be lapped, not in single player at least (correct me if I'm wrong). I haven't really tried to do longer 7+ laps circuit races in FH4 but used default blueprints. But I could imagine that on longer races lapping Drivatatars, even you are always racing against the clock, could be rewarding experience and help keeping focus. Doing many nearly identical laps to stay in first position can be a bit exhausting, demanding to stay focused and patient. Lapping opponents can break monotony of such race and give some feedback, I'm doing well, I don't need to worry about these Drivatars anymore, etc. that can be rewarding.

Other aspect with magnetic effect is that it tends to keep player on toes. For example no matter what how well you drive in Street Scene and Road Racing Sprints, there's no such thing as safe lead (though the Marathon and the Goliath can be forgiving, but it's situational). So more attention and focus, more on the edge players is to perform perfect race, more punishing it actually gets if something goes wrong. With stock cars at least, it's more or less trying to always be at best performance and moments of relaxation don't come from room to not flat out, but about knowing limits of car and route. It can be also immersion breaking, having huge lead on Goliath and then see from minimap how fastest Drivatars catches up no matter how fast player is going.

It has its pros to. Mechanic forces player to adapt to making less and less mistakes as difficulty goes up, if forces player to understand more about racing line, strengths and disadvantages of whatever car is used. It removes time consuming meta, such as what car should I use? As almost any car can be dream car that can win. Dynamic AI also adds longevity of game as there's no point where player enters in every race in practical God Mode.

So there are pros and cons. The big question is, what would make a better Drivatars? And just "make them more human" doesn't cut it. There were awfully lot of rammers in FM7, not Drivatars but actual humans. What kind of Drivatars would make Horizon better experience?

Originally Posted by: turtleCZ Go to Quoted Post
It's true but I like it. It's just a show for me. I wouldn't call it cheating because if you have a good time and race cleanly you can win which is important for me. Sure, there are some broken situations but it's uncommon.


Very uncommon in Street and Road events at least. I have encountered situations with bone stock RWD's on some Dirt Racing events where I haven't been able to win, but then I don't do Dirt Racing that much. A lot of my troubles I think comes down how independently left and right side wheels of car is allowed to work in curves. Cross Country is something else entirely IMO.
Rank: A-Class Racing License
#399 Posted : Monday, March 9, 2020 10:50:40 AM(UTC)
Originally Posted by: ClawedScroll554 Go to Quoted Post
Originally Posted by: turtleCZ Go to Quoted Post
[quote=turtleCZ;1151519]It's true but I like it. It's just a show for me. I wouldn't call it cheating because if you have a good time and race cleanly you can win which is important for me. Sure, there are some broken situations but it's uncommon.

Very uncommon in Street and Road events at least. I have encountered situations with bone stock RWD's on some Dirt Racing events where I haven't been able to win, but then I don't do Dirt Racing that much. A lot of my troubles I think comes down how independently left and right side wheels of car is allowed to work in curves. Cross Country is something else entirely IMO.

Yeah, I do road races only but wanted to start dirt more so we will see. I had some troubles on ubeatable with dirt races year ago. I thought it's the car but probably not.
Rank: C-Class Racing License
#400 Posted : Monday, March 9, 2020 11:08:42 AM(UTC)
Originally Posted by: turtleCZ Go to Quoted Post
Originally Posted by: ClawedScroll554 Go to Quoted Post
Originally Posted by: turtleCZ Go to Quoted Post
[quote=turtleCZ;1151519]It's true but I like it. It's just a show for me. I wouldn't call it cheating because if you have a good time and race cleanly you can win which is important for me. Sure, there are some broken situations but it's uncommon.

Very uncommon in Street and Road events at least. I have encountered situations with bone stock RWD's on some Dirt Racing events where I haven't been able to win, but then I don't do Dirt Racing that much. A lot of my troubles I think comes down how independently left and right side wheels of car is allowed to work in curves. Cross Country is something else entirely IMO.

Yeah, I do road races only but wanted to start dirt more so we will see. I had some troubles on ubeatable with dirt races year ago. I thought it's the car but probably not.


I should expand a bit. The whole thing Dirt is like tarmac, problem comes that I somehow come to make things and expect things in how real life and rally games work. It could be said that it's a matter of self control instead of car because that's what there is. But still, I haven't seen anyone doing power glide or flick like IRL, so no it's just not me.

Human factor I have been wondering the most is if using manual with clutch, staying on power band while coasting through curve would give me needed edge instead of doing the trick with shifting on lower gears, which is, should be at least, less effective method.
20 Pages<<1415161718>>PrevNext

Notification

Icon
Error