Released July 2022 Release Candidate Now Available

Just an FYI, there is something wrong with the TEST Porsche Cup car. I just tried it at Portland, and with the default setup, the rear of the car is very low and the rear wheels poke through the wheel arches. The tires also produce a continuous screech even when driving at low speed in the paddock and pits. Needs some more work...
 
burgesjl said:
Just an FYI, there is something wrong with the TEST Porsche Cup car. I just tried it at Portland, and with the default setup, the rear of the car is very low and the rear wheels poke through the wheel arches. The tires also produce a continuous screech even when driving at low speed in the paddock and pits. Needs some more work...
Click to expand...
None of the Porsche Cup or GT3 cars have the full sound update yet, so reduce your legacy sound effects slider dramatically to balance that out til more sound updates arrive.
 
Some observations with the AI on a wet track in the RC.

I did several short tests at Silverstone GT, both in the v1130 RC and the current v1128, all with these starting conditions:
tYN2AbT.jpg


AI settings:
Strength 100%, Aggression 33%, Limiter 1%

I tested both the Porsche 911 RSR GTE and the Audi R8 GT3 2019, so one car with the latest tyre tech, one without. Results as follows:

Porsche 911 RSR GTE
v1130 RC
c9ioCH7.jpg


v1128
hugZTTj.jpg


Audi R8 GT3 2019:
v1130 RC
L5PFePq.jpg


v1128
Iaemqee.jpg


As you can see, my own lap times are comparable. whereas the AI is either much slower (v1128), or much quicker. I'm guessing this is something to do with the RealRoad changes.

It would be nice for the AI to be quicker in the wet, as they've always been slow in such conditions, but this does seem a bit much! By way of comparison, I also did a quick dry-weather test in v1130 RC with the Porsche GTE. The fastest AI time with the car was 2:03.973, so a time of 2:10.019 in the wet is just 4.87% slower. That seems... optimistic in conditions that the sim describes as "Extremely Wet."

@Paul Jeffrey Maybe this is due to an intentional change? If not, perhaps it can be looked into before the RC becomes the main build.
 
jayarrbee36 said:
Some observations with the AI on a wet track in the RC.

I did several short tests at Silverstone GT, both in the v1130 RC and the current v1128, all with these starting conditions:
tYN2AbT.jpg


AI settings:
Strength 100%, Aggression 33%, Limiter 1%

I tested both the Porsche 911 RSR GTE and the Audi R8 GT3 2019, so one car with the latest tyre tech, one without. Results as follows:

Porsche 911 RSR GTE
v1130 RC
c9ioCH7.jpg


v1128
hugZTTj.jpg


Audi R8 GT3 2019:
v1130 RC
L5PFePq.jpg


v1128
Iaemqee.jpg


As you can see, my own lap times are comparable. whereas the AI is either much slower (v1128), or much quicker. I'm guessing this is something to do with the RealRoad changes.

It would be nice for the AI to be quicker in the wet, as they've always been slow in such conditions, but this does seem a bit much! By way of comparison, I also did a quick dry-weather test in v1130 RC with the Porsche GTE. The fastest AI time with the car was 2:03.973, so a time of 2:10.019 in the wet is just 4.87% slower. That seems... optimistic in conditions that the sim describes as "Extremely Wet."

@Paul Jeffrey Maybe this is due to an intentional change? If not, perhaps it can be looked into before the RC becomes the main build.
Click to expand...
Hi! This post really caught my attention as I've just yesterday had an article published on Racedepartment concerning this very thing!

Basically its all about sims that feature dynamic weather and the big problems it can cause with A.I. speed abilities in the wet being very different to their ability in the dry, to the point where it renders all the hard work done by the devs on these fantastic weather features completely useless for offline racing in changing conditions.

I'd love you're thoughts on this (especially if you are knowledgable about A.I. programing), weather you think its feasible or not, or if you have a better idea!

I may make a seperate thread for this, but in the meantime, here's the article with a supporting video at the end, there is also an example illustrating how big this problem can be at the end using AMS2 (as RF2 is not the only sim to have this problem, I generaly find the A.I. in RF2 with most cars become too fast in the rain.). Btw, this is not just a problem with A.I. its also to do with different players abilities in the wet, which of course the devs have no control over.

Anyway here it is: https://www.racedepartment.com/news...ing-weather-conditions-possible-solution.689/
 
Paul McC said:
https://www.racedepartment.com/news...ing-weather-conditions-possible-solution.689/
Click to expand...
Hi Paul, read your article and enjoyed your video :)

Conceptually, I think your idea has merit, although I have no idea about the programming involved so couldn't say whether it's feasible or not. There's a lot that would need to be worked out, and I'm thinking especially in races with changing conditions, from wet to dry and/or the opposite. For example, how does the sim define when the track has become "wet" (or "dry") and incorporate the increased/decreased skill level? Would that extra/lesser skill be added/taken away all at once, or gradually somehow? Could that be defined by linking to a percentage of track "wetness"?

I think this issue is well worth considering since, as you and I have both identified, sim racing AI rarely works as well in the wet as in the dry, and it's a particular problem in changing conditions. And that's a damned shame when the transitions are executed so well in sims such as rF2. I've always felt, ever since Studio 397 first incorporated some good wet-weather visuals (raindrops on the windscreen, track surface reflections) that those, plus the FFB, give a really good sense of how conditions and grip levels are changing. It'd be nice to make more use of that.
 
Has really nobody checked the new BTCC-cars with the current "RC"-candidate? Both undrivable 'thanks' to the new brake-sound:eek:
How can this crap passes any QC?
 
davehenrie said:
It has been noted, and they expect a hotfix before Q3 release.
Click to expand...
Interesting, but Q3 is already there and still: How could this happen and why isn't it fixed already?
I'm more involved into the AMS2-beta process and many past as well coming improvements are based on my feedback, like the upcoming far more reasonable aero-efficency of the F3-cars. Such a huge bug wouldn't take 48 hours to fix for Reiza. They also don't need feedback for such big mistakes, because they play the game...
 
Leynad said:
Interesting, but Q3 is already there and still: How could this happen and why isn't it fixed already?
I'm more involved into the AMS2-beta process and many past as well coming improvements are based on my feedback, like the upcoming far more reasonable aero-efficency of the F3-cars. Such a huge bug wouldn't take 48 hours to fix for Reiza. They also don't need feedback for such big mistakes, because they play the game...
Click to expand...
It's coming tomorrow, don't worry.
 
Leynad said:
Interesting, but Q3 is already there and still: How could this happen and why isn't it fixed already?
I'm more involved into the AMS2-beta process and many past as well coming improvements are based on my feedback, like the upcoming far more reasonable aero-efficency of the F3-cars. Such a huge bug wouldn't take 48 hours to fix for Reiza. They also don't need feedback for such big mistakes, because they play the game...
Click to expand...
the Release Candidate is here, Q3 is in August
 
The simulation was just stopped for a couple of times in every session I tried. At first I thought is was a complete freeze, and then I realized that everything else worked fine, but the car. It stopped. Marshals were talking to each other, shadows were moving, but the car just stopped for 3-4 seconds each time. The SSD was working hard, and then everything went back to normal. And then it started again half a minute later. 4-5 times in a lap.
 
You must log in or register to reply here.
Back
Top