Then we may have different standards I know its never going to be perfect but there are 4 collisions in that 20 second clip and several significant solo slides. Thats immersion breaking situations every few seconds (although as a driver youre luckily only there to witness them once a lap ) and overall 8 spins in one race. With how the ai handle yellow flags thats a potentially race breaking situation every 4 minutes. (And i know programming wise thats probably wishful thinking but theoretically a higher skilled ai driver should be less prone to making mistakes)
Wrong thread and while its not up to me to decide i wouldnt expect your request to be granted, afterall you literally just said you deliberately put people into the wall... (and if im not mistaken this isnt your first ranked ban either)
That l Looks way better than when i reported real spins at lv 105 like a year ago. Cars were actually spinning out there on lap 1. Bis this looks actually fine. and car contact is more than realistic especially at that level. I do understand the problem with yellow flags though but this small slide actually triggers a yellow already? that is a bit too sensitive
Car contact is realistic just because of going a bit faster? But no, nothing in this clip triggered a yellow. This clip was just to show them generally struggling through there and that because of that its not surprising they end up actually spinning sooner or later. In that race there were around 8 actual spins through turn 1, each obviously triggering a yellow flag. A couple spinners got going again very quickly, a few took quite a while with other ai crawling past in the meantime and one spinner kept getting rammed by another ai going back and forth for a fair bit
And yet when watching drivers in real life theyre not constantly having those moments and youd imagine the pros are as close to the limit as physically possible. Plus its not like the ai are actually close to the limit, most of them had best race laps that put them 1.5+ seconds behind what these cars are capable of
I did some AI tests on different levels and i saw that level 80 (and probably some other lower levels) are a bit problematic. I mainly tested levels 80, 100 and 120. See, cars on higher levels (100 and above) usually are 1 second faster in qualy than in practice and about 1 second slower in race than in practice. On level 80 however the cars are about 10(!) seconds slower in race than compared to their practice time. That seems to be very wrong. Combos i tested were Mazda Dpi on Road America, DTM2020 on Assen GP and Mercedes AMG GT3 2020 on Hockenheim GP. I will probably test lv 90 soon to see how many levels are problematic and will report back EDIT: Tested DTM20 lv 90 on Assen GP and the fastest laps in race session are 2 seconds slower than practice which I'd say is fine. So Lv 90 downwards seems to need fixes.
Dtm 2020 at brands hatch gp is a bit slow. Had 16 ai run a 15 minute quali session at 120 strength and pole position was still a second off the leaderboard p1
Adding onto the turn 1 problems for the ai at chang, now that i found some time in the setup the dtm 2020 ai are overall too slow at chang (in a race anyways). Started in p10 and was p1 after 5 laps at 120 strength. They were able to keep up and even repass after that for a bit but once we hit the gt3 traffic they fell behind and never really got closer again
Do the AI actually follow the racing line? sometimes they seem to be all over the place. In the good old days of rFactor, you could keep them a bit under control using corridors etc, I suppose it's a bit different these days.
So further to my previous post, AI at Oschersleben use way too much curb to the point of 4 x wheels well over the white lines. Also, they seem to be able to hang out wide in what would be the marbles IRL and go 2 or 3 abreast for multiple corners, again unrealistic. Love the track, predominantly the MC version as the AI can't seem to handle the tight T1 of the main course, but bit over being punted off by AI charging down the inside over the curb or darting all over the track unpredictably... ...just MHO