Automobilista 2 V1.5.5.0, Le Mans & Endurance Pack Pt1 RELEASED - Now Updated to V1.5.5.6

Is possible to update daytona road circuit ? the chicane have new design. I hope there will be the mustang GT3 in GT3 gen2

 

Without being pedantic then I guess you/we cant use a compare between AI and player car to say "it doesn't seem right".
Because when its 2 complete different physics (AI cars and Player car) then of course AI bots will be able to brake shorter than the player car.
And will sometimes also be able to accelerate faster than poor player car

 

But AI drivers should have braking zones as close to the player (depending on difficulty setting) as possible and not brake much later in ways that are absolutely impossible for a player. Same with acceleration, there is nothing that would justify it being any other way. I play and also played other racing games where that doesn't happen.

 

Principally you are right.
And the only thing Im more or less implying is that because the AI car physics is completely different (compared to player) then there will allways be small or huge differences in the behaviour of AI/player car. And even if Reiza worsened the AIs braking ability then there would probably be other situations where the AIs was braking too bad now.
The problem is the 2 different physics. Not as example braking to good/bad.

 

Which physics the AI are using is not the deciding factor. The AI could be using identical physics and be programmed to brake much earlier or later than the player and make a mess of a corner. The "reduced physics" AI that we have can be programmed to brake earlier or later and have their braking performance adjusted so it most closely matches the player car.

This latter step is what has not yet occurred fully since the focus has been on the updating the physics of the player cars to the maximum extent possible. So, two things will occur next:

1) Some player cars need some brake performance tweaking because they cannot brake as well as they should. I'm not sure if any brake too well, but I have only encountered cars with brakes that are too weak or too prone to early lock-up. These are a minority.

2) Reiza will systematically review and adjust the AI so their various performances, including braking, mirror the latest physics just delivered in the massive recent update.

Once these two are complete, the AI will no longer be able to out-brake the player so dramatically as reported by @MarcG. It is a real issue, as is any time the AI can accelerate better/put the power down better, corner with higher g-forces, etc. There is sadly no shortcut to tweaking all these values and due to the innumerable major (and welcome) physics updates to the player cars, Reiza hasn't caught-up on this work yet. Renato just recently confirmed in the beta thread that it will be occurring, as will some changes to track paths that the AI use to guide their decision-making.

Which physics the AI use is not a major determinant in how good they are to race against. We have had these discussions many times in the forum (unfortunately spread over multiple threads and months/years past). Although AI using the identical physics to the player would provide certain advantages, it is impossible to implement using today's available hardware, so arguing the minutae of it is pointless. We can just focus on what can work and I know Reiza appreciates the specific examples presented by users. Those outliers help can identify what to address first in the very long list of adjustments required.

 

Another pretty big issue (in my opinion) is the class of cars whereby AI at 120% are too slow, for example I took the 1999 Stock Cars round Watkins Glen Short this evening which was mega fun throwing that thing round sideways, but at 120% strength I was still 3-4 tenths faster per lap and I'd not touched the setup.

I/we either wait for every AI car & track combo to receive specific individual adjustment (a monumental task indeed as proving), or in the meantime we get a "boost" in the UI to go beyond 120% perhaps, food for thought.

 

No, no...
I've heard iRacing does it the right way or at least the I would implement it.
100% is almost impossible to beat, alien territory. Perhaps a game could go to 105% because sometimes a Verstappen steps in and we need to welcome him .
But I guess marketing strategies lead studios to make 100% an accessible goal so we can say we are great drivers .
On a more serious note, it will never happen, but I would like sims to have 100% as the ultimate goal of a sim racer. It's just a cognitive bias thing I guess, but a very persistent one.

 

Yeah I said "meantime" to mean something to get us by whist we wait for specific changes, I totally agree 100% should be on the cusp of Alien but it's never that easy as proven in a fair few racing Sims over the years.

Dynamic Strength changes like what RaceRoom had/has would be decent, especially for championships where it gets annoying to have to drastically alter the strength for each track.

Respectfully it is a fine art to get all this right, I don't envy Renato and the team with this particular task

 

Here’s a multi-class combo that worked really well at Bathurst 1983: Group A & Omega 99.

These cars could serve as Australian Class A competitors during the early 1990s.

 

Yeah and its pretty logical that if both the AI use a complete different physics AND then have to use scripts that tries to compensate for this then there is even more strange variables involved.
And the more variables the more difficult it is to get the AI bots to drive/behave like the player car.

 

For all:

It's very important to remember it is a two-tier system. General variables and vehicle/track specific changes. So, in the example above, if possible, before we all report, try to clarify if the '99 Stock cars are generally too slow relative to other series, or, if they only misbehave at WG Short, or if WG Short is consistently slow in all series compared to other layouts (at WG or otherwise). Or, if you are a testing freak and have done enough multiple combos to have some statistical validity, you could say it was all three or two of the three.

Depending on which is the answer, a global change might address the majority of the issue (good for everyone) versus very deep, specific changes required.

 

Has somebody else noticed LMDh tyres being maybe too durable?

I ran an online 60 minute race last night at Spa, driving the Porsche 963 on Soft tyres. Had to make a stop for fuel, but there was no need to take new tyres. The track was quite cold due to in game date being at March, and running most of the race through the night, so I was struggling to keep the fronts up to temp. The tyres wore down around 30% according to Second Monitor, and had no pace loss at any point.

 

Yes:

and the answer:

 

It was just the invoice number, not a card number or a bank number, ok

 

With the content manager everyone can use mods

 

It would be nice if there was a way to choose the cars for the trip and not just class the car. Independently.

 

Not entirely convinced.

The practice of doing 4 or 5 stints in a row on a single set of tyres is something you see only at Le Mans, not at the rest of the tracks. Le Mans also favors the use of softer tyres due to both being a track that does not eat them, and also having plenty of time at straightlines to cool them down.

While I would be perfectly fine with cuadruple-stinting Softs at LM, doing the same at any other track does not seem right.

 

unrelated funny stuff, when will online ppl learn what a yellow flag is?

 

Regarding the AI super brakes, I don't think it has anything to do with the AI not being on the same physics, it's more like a superficial "boost" they get when within 1 to 1.5 seconds behind the player or each other. I think it's actually one of the main culprits of the AI pack racing, because often times the faster car gets by on one corner but cannot get away from the slower car because it closes on the next corner due to his super brakes, rinse and repeat.

It also causes an infinite overtaking loop between 2 evenly matched cars. You can see this issue easily in a track like Barcelona with the F-Ultimates, the car with DRS passes the other car on the middle of the straight but the passed car gains its super brakes and closes on corner entry, then passes the car on the next lap and they can stay like that the whole race.

I really enjoy the AI in AMS2, but this issue in particular is pretty annoying because when it happens against you, you don't get a real challenge because once the AI passes you, they are extremely easy to overtake again and just pester you, because you can easily overtake them again. And when it's between AI is annoying watching 2 AI repeat the same moves and overtaking each other for 30 laps.

 

I had the same though about rebranding the Chevy Omega to an Opel Omega 3000 24V to perfectly fit into the 1991 DTM season.