1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

To Reiza: What physics do you actually want?

Discussion in 'Automobilista 2 - General Discussion' started by Richard Wilks, Jan 3, 2023.

Tags:
  1. Mhad

    Mhad Active Member AMS2 Club Member

    Joined:
    Apr 13, 2020
    Messages:
    327
    Likes Received:
    215
    Back to the 98T. I've done another lap at Cascais trying to recreate Senna's.

    I am by no means a great driver, but I ended up 2tenths off (1:16.864). From the external view the lap looks very similar.


    I did modify the set up, this lap is currently number 1 on the TT boards for this combo. (Probably not for long). SlickD tyres, 3laps fuel.

    Also, my SecondMonitor telemetry is attached.
     

    Attached Files:

    • Like Like x 4
    • Informative Informative x 4
  2. Kuku

    Kuku Flying Kiwi AMS2 Club Member

    Joined:
    Aug 30, 2020
    Messages:
    1,494
    Likes Received:
    952
    Hats of to the Replay Camera
    Car feel pretty good around there .
     
    • Like Like x 1
  3. Maser V6

    Maser V6 Assume nothing._ Verify everything._Have fun AMS2 Club Member

    Joined:
    Mar 27, 2020
    Messages:
    2,502
    Likes Received:
    1,422
    Thanks for efforts and time.
    Different from op then.
    Does not confirm for some an issue does not exist of course, though illustrates at least in vid from 1 side that not everyone has same experience with reported handling. How is the car for you on default? Curious!
    Car looks more settled than Rl car in some corners.@ 0.40
    sidenote. Video reminds me of @GregzVR comparisons..
    very good.
     
    Last edited: Jan 7, 2023
    • Like Like x 2
    • Friendly Friendly x 1
  4. Renato Simioni

    Renato Simioni Administrator Staff Member

    Joined:
    Feb 14, 2016
    Messages:
    4,956
    Likes Received:
    47,136
    Without reading through the whole thread but hopefully answering what the premise of the OP is in essence - generally speaking there is very little "want" in physics development if you are serious about it; you model the car with the data and understanding you have of said car, of vehicle dynamics and the physics engine representing them for a model that is as accurate as possible to the knowledge you have of all fronts at the time. With that said new things are always being learned, and there is always some latitude in what´s reasonable for every variable that allows some "design" work so it´s less exact science and finite process than one would hope and most sim racers seem to understand it to be - but if the developer approaches it as essentially a work of design he/she is unlikely to get it right in objective terms.

    Thus some cars may err for being too grippy, others too little; some from being too forgiving, others too little; for being too fast, or too slow - there may on occasion be a common flaw to most of them (for ex. issues we had earlier on due to general lack of tire carcass stiffness or bugged LSD modelling) but generally speaking each car is its own model. Users often try to draw patterns as physics engine "issues" as humans usually will do, but most of the time that´s down to a fundamental misunderstanding of how physics development actually works.

    Taking the case in point, the Lotus 98T is the one car in the pack I´m less happy with in terms of handling, and I´m sure as we continue to pour more time into its development we´ll find some errors in it and rough edges to trim as one does with every physics model to one degree or another - but to put things bluntly, if you consider the reference point to be the same car in another game in which cars often pull 30-50% higher lateral Gs and are seconds per lap faster than real counterparts, you´re unlikely to ever be satisfied with our representation.
     
    Last edited: Jan 7, 2023
    • Informative x 17
    • Winner x 6
    • Like x 4
    • Useful x 2
    • Agree x 1
  5. Racinglegend1234

    Racinglegend1234 AMS2 wiki founder AMS2 Club Member

    Joined:
    Jul 28, 2022
    Messages:
    5,749
    Likes Received:
    1,994
    Speaking of physics, any news on V1.5?
     
  6. sgsfabiano

    sgsfabiano Well-Known Member AMS2 Club Member

    Joined:
    Oct 28, 2019
    Messages:
    2,058
    Likes Received:
    1,232
    Did I miss something about this v1.5 you keep asking for?

    This is a legit question, as this specific version of the game has been asked for around for a bit now.
     
    • Agree Agree x 1
  7. Racinglegend1234

    Racinglegend1234 AMS2 wiki founder AMS2 Club Member

    Joined:
    Jul 28, 2022
    Messages:
    5,749
    Likes Received:
    1,994
    It was lasted spoken about in the August development update. In it it say it is expected either late 2022 or early 2023. I’m just wondering if it is still expected early 2023 or not.
     
    • Like Like x 1
  8. Roar McRipHelmet

    Roar McRipHelmet Well-Known Member AMS2 Club Member

    Joined:
    Feb 25, 2020
    Messages:
    1,577
    Likes Received:
    1,266
  9. Renato Simioni

    Renato Simioni Administrator Staff Member

    Joined:
    Feb 14, 2016
    Messages:
    4,956
    Likes Received:
    47,136
    He´s probably referring to the fact that unlike our approach up until v1.4, we have opted to refrain from introducing major incremental physics developments with every update, but rather concentrate them to bigger milestone updates (as v1.3 and v1.4 were, and v1.5 within the next few months will be) - this allows us to work through our findings more methodically and apply them wholesale to all cars the game, at the same time preserving handling / performance of cars for longer periods (instead of forcing users through the old routine of having cars change almost on a monthly basis).

    And in answer to the original question: yes there are substantial physics developments already being worked on for v1.5, the newer cars inevitably already benefit from them to one degree or another as we can´t unlearn what we have learned to model flawed cars purely for the sake of consistency :)
     
    Last edited: Jan 8, 2023
    • Like Like x 21
    • Informative Informative x 3
    • Winner Winner x 2
    • Friendly Friendly x 1
  10. Dady Cairo

    Dady Cairo "Son of Spartakus" and "Leisure Nostradamus" AMS2 Club Member

    Joined:
    Jun 12, 2017
    Messages:
    2,261
    Likes Received:
    782
     
    • Funny Funny x 2
    • Like Like x 1
  11. sgsfabiano

    sgsfabiano Well-Known Member AMS2 Club Member

    Joined:
    Oct 28, 2019
    Messages:
    2,058
    Likes Received:
    1,232
    Cheers! It is undeniable the game is evolving, one way or another, at one pace or another.

    Even if Richard is 100% right about something wrong with physics (which in my ignorance I'd agree to some degree), this "flawed" BT52 still is the most fun F1 car I've raced so far in any sim.

    Finally, like some have already mentioned, reading the technical exchanges between Richard and steelserv was enlightning. Too bad it ended in attacks... if both worked together I think something good could happen for the sim.
     
    • Agree Agree x 7
    • Like Like x 2
  12. Renato Simioni

    Renato Simioni Administrator Staff Member

    Joined:
    Feb 14, 2016
    Messages:
    4,956
    Likes Received:
    47,136
    All racing sims are “flawed” to a degree or another as their physics engines are inevitably composed of interpretations & aproximated / simplified models of very complex dynamics; any developer that says otherwise is selling you something, and they all are :p

    We do have the most sophisticated physics model of all commercial sims IMO, and I have insight into most of the main ones to say that with authority - but that also means more room to get things wrong. I’m very happy with the physics in AMS2 as they are now, but in a year or so it’s probable I’ll feel very differently as we move along further - it’s the equally frustrating and fascinating nature of the process.
     
    Last edited: Jan 7, 2023
    • Like Like x 26
    • Agree Agree x 3
    • Informative Informative x 1
  13. oez

    oez Mayor of Long Beach AMS2 Club Member

    Joined:
    Aug 27, 2020
    Messages:
    697
    Likes Received:
    680
    This is really what's so difficult about working on this stuff. You can't do everything alone so you need a team. But then you can't just do whatever you want. It takes patience and careful communication a lot of the time to work as a group towards something. I wish our lizard brains were better at filtering out unnecessary words or nuance from feedback, but unfortunately in the heat of the moment and under pressure/stress it does matter how things are expressed. So within a group of people you have to build chemistry, learn to communicate in an effective way that minimises friction and get the job done. Only then can anyone's knowledge, experience and added value do more good than harm. It really isn't for everyone. Working alone can easily be a nicer experience to some people and there's no shame in that.

    By the way since it was asked early on, but I'm not sure was ever addressed: Renato is still the lead physics developer. Whatever steelreserv or anyone in the team might be working on is ultimately accepted by him. If something slips in that shouldn't be there, he's going to let the team know and correct it.
     
    Last edited: Jan 7, 2023
    • Informative Informative x 7
    • Like Like x 3
  14. Richard Wilks

    Richard Wilks Active Member

    Joined:
    Feb 12, 2018
    Messages:
    183
    Likes Received:
    236
    Thank you Oez, for the straight answer, i am glad that Renato is still the main physics man, he is a man of experience, who knows his way around this, so that at least lets me know that there is some sound authority in the team in terms of physics lead.
     
    • Dislike Dislike x 9
    • Like Like x 1
  15. Richard Wilks

    Richard Wilks Active Member

    Joined:
    Feb 12, 2018
    Messages:
    183
    Likes Received:
    236
    I do agree that the madness engine is an impressive piece of kit, and people who know me for sure saw me already defending it, either be on discord or forums. It has more than enough potential to be the best thing out there at the moment, but of couse, i do understand that complexity also means new and more complex challenges.

    Like i stated, it was not my purppose to slag the team, or the car, or much less the game. I always agreed with your premise that these cars are usually much more controllable than many virtual representations of them lead us to believe.

    I am glad to hear that this is still an open book, and there there are cars (like the 98T), which are still wip and that the team wants to improve upon them.

    My thread was basically to try to shed some light, and bring some of my observations to the table in terms of what one can feel in the end user role, and to know that team is on it. I am glad it is. Thank you for the response.
     
    • Like Like x 4
    • Friendly Friendly x 1
    • Useful Useful x 1
  16. Richard Wilks

    Richard Wilks Active Member

    Joined:
    Feb 12, 2018
    Messages:
    183
    Likes Received:
    236

    Thank you for your video, because it perfectly ilustrates what i was saying in terms of lack of lateral integrity in the rear whenever there is weight transfered to it.

    Now again, that might be the diff overreacting (if indeed its locking the rears everytime you go on the power) Or it might be what i said initially, the rear tires are basically "overwhelmed" in terms of slip angles by the weight transfer, so the car wanders a bit sideways with a "self steering" effect everytime you press on them. Thats very visible in the stadium section of your lap. The front tires indeed seem a lot "strong" than the rears in that regard, so the car never "plows" forward in acceleeration, because the rears are overwhelmed first.
     
  17. Richard Wilks

    Richard Wilks Active Member

    Joined:
    Feb 12, 2018
    Messages:
    183
    Likes Received:
    236

    Thanks for the response.

    Just to make something clear, i never said that the 98T in Pcars2 with vintage/historic tires (again important) was perfect. I just commented how the tires seemed to work properly in those situations i described, and how they communicate things to the driver. The car might indeed be too quick, altough from my testing at least the historic tires werent so far off. I am NOT talking about the default tires of said game.
     
    • Like Like x 2
  18. BrunoB

    BrunoB TT mode tifosi BANNED

    Joined:
    Jun 13, 2020
    Messages:
    2,582
    Likes Received:
    651
    Oh no..
    Focus on some of the other cars before - please.
    I have been looking forward to get this car into AMS2 because the L98T was the only car I REALLY liked in AC.
    And then to my positive surprise the AMS2 version was similiar but slightly more brutal.
    Yeah man I thought - now we are really doing businezz :D
    But now - mainly because a lot of casual users cannot control this car - you are allready talking about changing the behaviour of this marvel.

    Repeat: Focus on some of the other cars before - please.:(
     
    • Funny Funny x 3
  19. Richard Wilks

    Richard Wilks Active Member

    Joined:
    Feb 12, 2018
    Messages:
    183
    Likes Received:
    236

    Picking again this example, notice how even these massive inputs that the real driver is giving the car steering wise, seem to have (at first) no bearing on the attitude of the car. That is because it actually is not easy to change the attitude of a car when its in motion, unless there are massive weight transfer shifts. The cars in the game seem to respond well in this, and they do maintain some lateral integrity at all times just to the steering alone, but whenever the accelerator is involved, this lateral stability disappears instantly in the rear.

    Some people classified this as the "shopping trolley" effect, and i guess its a blunt way of putting, because the car seems to self steer.

    Now, i know that SOME self steering due to slip angle ramping is important. In fact, its CRUCIAL, specially on ovals. Many games and mods dont aknowledge this, and we end up with far too understeery cars that are unwilling to turn "around" a corner properly.

    So this is a GOOD decision from the physics team to explore this. But i guess the effect is too pronounced whenever there is some weight transfer ocurring, at least for now. So it might be a matter of fine tunning this between the rears and the fronts.
     
    • Like Like x 6
  20. Racinglegend1234

    Racinglegend1234 AMS2 wiki founder AMS2 Club Member

    Joined:
    Jul 28, 2022
    Messages:
    5,749
    Likes Received:
    1,994
    Can’t you change the setup to make it even MORE BRUTAL?
     
    • Like Like x 2
    • Funny Funny x 1

Share This Page