GT4 AI: OMG, it's even worse than I thought.

Discussion in 'Gran Turismo 4' started by RaidoGT, Oct 19, 2009.

  1. Austin343

    Austin343

    Messages:
    1,903
    Location:
    United States
    In general, you're wrong. You only have to look at the exceptionally competitive world of racing competitions in GT4, such as the inter-forum Board Challenge or the Weekly Race Series, to find that top drivers essentially never use aids. Believe me when I say that competitors in those series are doing whatever they can to go faster, and are not keeping aids off out of principle. In certain rare circumstances, such as when a car has bad wheelspin, a TCS setting of 1 (but never more) does give an advantage. Also, extremely rarely, an ASM Understeer setting of 1 allows certain cars to be driven with more precision and speed. However, in Arcade Mode, having no driving aids is always fastest, because the levels of TCS are non-adjustable and are set too high to be useful. In general, avoiding wheelspin is largely achieved by selecting a much higher gear than is "recommended," rather than by activating the traction control. It's true, though, that for more casual players, having aids on can be faster in powerful and/or ill-handling cars.

    I apologize for the off-topic post, but I couldn't help but refute that statement.
     
  2. niky

    niky Moderator

    Messages:
    23,800
    Location:
    Philippines
    I can never go faster with any ASM than with it completely off.. no matter how hard I try.

    See... I did a series of road tests a few years ago (filed under "Road & Trek", in GT4 Race Reports), where I did driving tests of simulated cars for a fictional magazine.

    For one test, I decided that since it was almost utterly impossible to turn off the electronic stability control of Mercs in real-life, I would simulate their effect. And try to simulate the effects of a BMW M5 stability system... one that would allow you to powerslide at a slight angle without kicking in, yet save you when you drove incorrectly.

    I could simulate the Merc's almost shutting down if you tried to throw it sideways. I could not simulate a BMW system at all. Even at its weakest, ASM was way too aggressive to help you drive a fast race.

    TCS, set at 1-3, helped with the stupidly powerful cars, but even at 1, if you drove with a pedal, you would still be faster with it off given the proper technique.

    GT4's driving aids are actually woefully ineffective compared to some of the better real-life driving aids available in the market right now...

    -

    As for rubberbanding AI... nope. I drove tons of endurance races... and when I played arcade, I did very long races... 10-15 laps... every so often. Whether playing in A-Spec or B-Spec, the results were always the same. There were slow cars and the occassional jackrabbit. Whether you were right behind him or not, the jackrabbit would be X seconds ahead of the next AI car by lap 5. If you pass him and don't hold him up, he will still be X seconds ahead... give or take a few tenths depending on whether you nudged or punted him when you passed... :D

    From your tale... it sounds like you're holding up the lead AI car... plain and simple. He's faster, you've got track position... Sounds like. The validation on your part would be... how far ahead of him are you when you took the lead? What are his laptimes? Did you accidentally hit him when you passed?

    Start a 20-30 lap race. Identify your "rabbit". The first time, do NOT even get near the rabbit till he gets to the font... then pass him, build up a ten second lead so he cannot draft you, then observe his laptimes. Restart... with the same line-up right away. Same thing... do NOT interfere with the rabbit's jockeying through the field. This time, when he gets to the front, stick to his tail. Observe his laptimes.

    I guarantee you his laptimes through the 15th lap will be nearly identical in both cases.
     
  3. Seismica

    Seismica Premium

    Messages:
    4,346
    Location:
    United Kingdom
    The driving aids when set up correctly, make the driving easier, with the right ASM setting, you get more grip in the corners, this allows for faster and more consistent driving without the need for great skill. I am sure around 99% of all the hardcore players play with the aids off, because 1: its more fun, 2: Its more realistic, 3: Online comps usually don't allow aids, or they get called noobs for doing so.

    Playing with aids off requires more throttle control, so it would be suited to a wheel user, because pedals > buttons. Like i said, it can't be compared unless someone does a bit of research on it, but they would have to be one of the fastest GT4 drivers otherwise the experiment would be pointless.
     
  4. niky

    niky Moderator

    Messages:
    23,800
    Location:
    Philippines
    The only aid that makes sense for any user, pad or otherwise, is a little TCS.

    ASM will always... always slow you down... even if you set it to catch understeer only.

    Understeer is very easy to catch without ASM. With ASM on, it slows you down too much... even on the lowest setting... delaying that moment when you can get back on the power. It may not be faster to understeer a little bit past the apex and hammer the throttle compared to slowing down properly and clipping the apex... but the ASM never does slow you down properly... and is a distant third in effectiveness to the other two strategies.

    ASM on oversteer is even worse. On the lowest setting, it cuts in way before you start sliding... making for massive understeer.

    Fastest laptimes are always without aids at all... and top ranks are filled by guys using no aids or only using a little TCS. This is still true in GT5P... people at the top don't use aids not because it's more challenging... but simply because it is faster.

    I'm a pad user, and I've never been able to come close to my best times with TCS only or with no aids when I've turned ASM on, in GT4. I can sometimes even get better laptimes with TCS off, if the car isn't stupidly powerful (like a Speed12 on S-tires), because I use the right analog to accelerate and brake.
     
  5. Parnelli Bone

    Parnelli Bone Premium

    Messages:
    10,506
    Location:
    United States
    I've driven quite a few cars with all aids on "1" (to try and simulate modern cars which would have these on while they're stock). I'm always significantly faster once I've turned them off.


    Off-topic now, but this is important. Is anybody else remotely alarmed or concerned about these quips from Radio GT? Or am I the only one?


    Isn't this a major violation of the AUP?

    Some of you are gonna say "well, the AUP rule concerning insults and threats only applies to actual members of GTP". To which I'll say in advance...

    1). It's completely possible (I'd say probable, actually) some of PD's staff has joined GTP at some point, hence there could be PD workers here.

    2). I'd actually consider any of PD's staff an honorary member of GTP. These are the creators of one of the greatest racing games of all time, which gives us so much. How dare anybody threaten them with violence.


    Grrr....:mad:
     
    Last edited: Oct 23, 2009
  6. Austin343

    Austin343

    Messages:
    1,903
    Location:
    United States
    Yes, they do make it easier, but, critically, you're mistaking "easier" for "faster." And as a participant myself in online competitions, you have to believe me when I say that the fastest drivers always consider using aids. And, precisely because of their skill, activating any one of them is almost always slower. But occasionally, a particular car and track combination will see the fastest drivers all on the top of the results board, and all using TCS. And pedals may be superior to buttons - although in reality, with a skilled driver, only slightly - but they are about equivalent to the analog sticks of the DS2 with regard to throttle control. I use the sticks, and objectively, I am one of the fastest drivers still playing GT4 (see my signature). I dislike using aids, but I do consider them when there is bad wheelspin, and only in really extreme cases such as the Speed 12 on Economy tires is turning on the TCS faster.

    Extremely rarely, there are exceptions to that rule. In particular, there was a WRS race a long time ago that demonstrated the value of aids in certain cases: Week 55. Pay particular attention to posts 29 and 33. holl01 is probably the single fastest GT4 driver ever, and he was totally demolished by a very skilled but nevertheless lesser driver who was using ASM Understeer to improve the smoothness and speed of the 22B's cornering. I'm sure you all know how there are some cars that turn better under braking than they do while coasting, and I think that's what happened here: the ASM applies the brakes while cornering under full throttle, thereby not only maintaining acceleration but also improving the weight balance of the car.
     
  7. niky

    niky Moderator

    Messages:
    23,800
    Location:
    Philippines
    I guess it would depend. What tire were you guys using in that challenge? Hmmm... super-softs...

    If I ever get my PS2 back, I've got to try that.

    I suppose this only works for some of the worse AWD cars? I've never ever experimented with this and found it worth it, but of course, with the number of cars in the game, I'd never gotten around to trying it on more than a few dozen.... and many of those were cars that were really tricky or twitchy to drive... hence, cars that should benefit more from the tuning.

    It's interesting to see that the most useful application of ASM would be on a car that technically doesn't need it! :lol:
     
    Last edited: Oct 23, 2009
  8. Austin343

    Austin343

    Messages:
    1,903
    Location:
    United States
    I wasn't actually participating; in fact, I didn't even own GT4 when that particular race took place! And although I know the "SS" you read was vague, it meant Sports Soft tires, not Racing Super Soft.
     
  9. Lain

    Lain Premium

    Messages:
    6,908
    Location:
    United States
    Plus isn't assault a crime? Doesn't the AUP say you're not allowed to discuss illegal activities, or something to that effect?

    Anyway... after playing GT without driving aids for so long, they actually mess me up pretty badly if I turn them on.
     
  10. Seismica

    Seismica Premium

    Messages:
    4,346
    Location:
    United Kingdom
    I did neglect to mention the car and track play a big part too. And putting 1 on TCS and ASM understeer is not necessarily a good setup. Cars are often tuned with aids off, so i doubt there are many good setups that use driving aids on these forums; someone may set up a suspension that prevents too much understeer, if they put ASM on aswell, it will ruin the balance. The aids have to be set in conjunction with everything else.

    As a general rule though like you said, you don't want to aids too high otherwise they interfere too much with the driver behind the wheel. It requires a different driving style (less coasting more throttle? :dopey:). That is probably why most people are noticeably slower when they turn aids on.

    Just to add, left foot braking with pedals, you can't really do that with the right analogue stick. Left foot braking is sometimes necessary. But again, it depends on the car, track and driving style. But at the same time throttle control is more difficult using the standard control setup, which is why pedals are better than controller :D
     
    Last edited: Oct 24, 2009
  11. Parnelli Bone

    Parnelli Bone Premium

    Messages:
    10,506
    Location:
    United States
    Yup. Well said. I happen to suspect one past member who used to post frequently here at GTP of definately being part of PD's staff, tho I can't prove this.


    Not to mention the original topic of this thread (the Ai supposedly cheating) is a bunch of horse-hockey. I say if you can't keep up with 'em, and are getting that frustrated...move on to another game, or keep practicing GT4 till you get bettter. Don't drive on crappy tires expecting the sims to honor you with crappy driving...of course they're gonna murder you! :sick:
     
  12. RaidoGT

    RaidoGT

    Messages:
    137
    Oh dear oh dear, someone seems to have lost their sense of humour or be a smidgen too sensitive for the Internet perhaps.

    (I haven't slapped anyone *yet*, of course. The last time I did must've been somewhere around primary school as well, but in this case I'll happily make an exception.)


    I *do* hope that PD is listening in, btw, and invites mr. Mountainhouse (Yamauchi) and his aides to the pitfight. Entrance fees will be moderate! But I'll be happy to smack 'em silly on behalf of everyone who ever felt cheated because of the bloody AI bumping or pushing you off the track again. (Hands up everyone who feels that the four years development time for the AI from GT3 to GT4 was really really well spent! :tup: )

    Like I said, some people really need to pick up their reading skills. It had nothing to do with tyres. I was following them around on that particular race on N2's, and could keep up or win. Had I used S2's, it just would've been easier to follow, but the result - the AI running away if you're chasing but not if you're in front - would've been just the same. And no, I wasn't holding them up either. (Just won my 4th race in the Japanese Championship on N2's as well btw (car used: 392 BHP Evo), just one to go... so despite being sucky, they're not completely useless I guess. And that's without aids - which I never use since it's me who needs to drive the car, not the bleedin' computer -, or downforce, or TC. The latter I'd only use on supercars and only if they got it in real life, but that's about it.)

    And no, not gonna repeat that AI test on the 80's race 'till I'm blue in the face. Not interested to keep doing the same over and over, and putting up photo's probably wouldn't really convince anyone either.
    By repeating it several times it was painfully obvious what was going on, the AI *did* cheat to keep you behind in certain circumstances. Whether you believe me or not, just try it for yourself if you happen to drop into the 80's series. Sooner or later someone is bound to bump into the same phenomenon. It was just too repeatable to be attributed to random variations.
     
  13. Parnelli Bone

    Parnelli Bone Premium

    Messages:
    10,506
    Location:
    United States
    You specifically threatened someone else with violence. Twice. Now you've done it a third time.

    Are you directing this third threat towards me? That's definately AUP violation there, buddy. :mischievous:






    What's "painfully obvious" to me is that you're not nearly as good a driver as you think you are.

    I've done plenty of "catch up" races (literally hundreds) where the Ai was vast seconds ahead of me. Got the replays saved so I can show them to you & everyone here someday (got to get a DVR first). In some cases, I was able to catch up & pass...in others, I was merely able to catch up. In ALL cases, at some point it looked as there was no way I'd be able to win.
     
  14. Smallhorses

    Smallhorses Staff Emeritus

    Messages:
    7,260
    *edit* reopened after cooling-off period.
     
    Last edited: Oct 30, 2009
  15. PS2 GT4

    PS2 GT4

    Messages:
    104
    After playing this game for over a dozen years I seem to always discover, through researching, something that I never heard of before... In this case it's "RUBBER BANDING" or what I use to refer to as the slingshot effect...
     
    Last edited by a moderator: Dec 24, 2016
  16. BobK

    BobK Premium

    Messages:
    6,970
    Rubber banding is a well-known phenomenon in the GT series. It was very pronounced in original Gran Turismo, and quite evident in GT2 and 3. GT4 actually didn't have any pronounced rubber banding. It was back, though, in GT5.

    Hard to believe someone can play the GT series for twelve years and never have heard of rubber banding before.
     
    Last edited: Aug 29, 2016
    TeamCZRRacing and PS2 GT4 like this.
  17. PS2 GT4

    PS2 GT4

    Messages:
    104
    Well BobK... believe or not, I rarely researched anything on the net before about GT4 until picking the game up again this year after some years passed and my grandson got older. We started playing the sim together after getting the bug again and finally completed it earlier this year... We obviously noticed it racing over the years and we use to call it slingshoting, but had never heard of the rubber banding term...
     
    Last edited by a moderator: Dec 24, 2016
  18. C0rd0syl

    C0rd0syl

    Messages:
    61
    I see here that nobody provides any evidence for their claims...so I went to check it out myself.
    I conducted a test on one race, with the same starting grid each time.
    Test was ran on a real PS2, on the PAL version of Gran Turismo 4.

    Player car: TVR Cerbera Speed 12 '00, completely stock.
    Race: TVR Blackpool Racers race 5: Fuji Speedway.
    I chose this track because the long straights are excellent for rubberbanding.

    The rabbit is the other TVR Speed 12, starting in 4th place.

    1st test: parking my car off the track.
    Rabbit time: 5'49.064

    2nd test: player driving at racing speeds. I managed to pull 8 seconds per lap on average.
    Rabbit time: 5'49.330
    My time: around 5'25 (I had to wait for the rabbit to cross the finish line for its timing)

    3rd test: riding the rabbit's rear bumper for the entire race without interfering.
    Rabbit time: 5'48.117

    Conclusions:
    According to OP, the rabbit could speed up indefinitely as long as I was right behind him. Yet I cannot make the AI drive at my racing speeds by doing so. My teasing only made him 1 second faster, when it should be a good 20 seconds.

    In addition, I kept an eye on when the rabbit overtakes the Tuscan Speed 6, starting in 1st place. No matter what I did, the rabbit would always pass the Tuscan on the straightway at the end of the 1st lap. He should overtake the Tuscan earlier if my driving made a difference to his speed.

    All this said, I'd like to see someone find evidence that their driving has a meaningful effect on the AI. My findings state otherwise.
     
    BobK likes this.
  19. Obelisk

    Obelisk Premium

    Messages:
    8,470
    I'll need to look into this as well, since I still have my GT4 and PS2.

    I'm thinking I'll test DTM cars to see if there's rubberbanding in GT4. Or better yet, a mono-spec car.
     
  20. PS2 GT4

    PS2 GT4

    Messages:
    104
    Rubber Banding becomes very evident at the Nurburgring 24 Hour in the Endurance Hall if ya run the first lap in "A spec" as follows... Early on get past the first three cars as soon as the opportunity comes up... Then catch up to the next two cars and tail them for awhile until you decide to pass the second place car... Then it clearly starts until you get past the first place car and the second place car eases up. But then the first place car reacts just like the second place car rubber banding you until you get past both of them with a lot of distance... Try it like I have many many many times...
     
    Last edited by a moderator: Dec 24, 2016
  21. C0rd0syl

    C0rd0syl

    Messages:
    61
    You do know that slipstreaming is strong in Gran Turismo 4, especially with race cars? AI will always "rubberband" if you're close ahead of them.
     
    PS2 GT4 likes this.
  22. submaniac93

    submaniac93

    Messages:
    1,332
    Location:
    France
    I personally believe that the AI in endurance races is set whether you're in A-spec or B-spec as B-spec. And usually in B-spec, once a car starts lagging behind, it'll seem as if he's going into "1" speed when you're in B-spec, while the AI that's closest to 1st will still be at 4 or even 5. I've seen multiple instances of me lapping a car that's been alone for a while (like on the Nurburgring) and it suddenly picking up the pace, then becoming slow again after I passed it for a while. It's prety easy to see that when you first start as A-spec, and then switch to B-spec, and then compare each AI's fastest lap vs they're latest.
    However, in Arcade and non endurance races (except sunday cup and the easiest races) I've never seen any kind of catch up, except the difference in power between cars.
     
    PS2 GT4 likes this.
  23. PS2 GT4

    PS2 GT4

    Messages:
    104
    Today, after the real IMSA races were over in Virginia, I decided to run a quick lap around the Nurburgring Circuit with the Option Stream Z and delete my old video at the Nurburgring circuit previously posted. I did this video just now and posted it up on YouTube. Toward the middle to 3/4 through the video I pulled away with a very good lead. Toward the end of the video the Audi A4 Red Bull racing team seems to get a boost of super power catching up on my lead in short order and I really had to drive my butt off without crashing or bumping to stay ahead. If it's not rubber banding then I don't know what to call it... Being straight up, frankly it just seems like I'm the only one that posts any video proof of stuff to back up my posts...

     
    Last edited by a moderator: Dec 24, 2016
  24. submaniac93

    submaniac93

    Messages:
    1,332
    Location:
    France
    Hmm... I'd like to try myself. Were you playing in Arcade mode? or something else? Also, I'd like your time for this lap if you still have it, as well as the tires you were using.

    'Cause, not to sound disrespectful, especially because I can't tell that much without a HUD, but you seemed awfully slow from what I saw (full last half of the video). Of course, you might have been slowing down on purpose just to prove your point, which then invalidates all that I said, but I wanna try :)

    I'll be posting a video when it's done.

    EDIT: Okay I watched the whole video to make sure I didn't miss anything. Here's what I think:
    0. Opponents:
    JGTC Impreza - 300hp
    JGTC MR-S - 300hp
    JGTC Celica - 310hp
    DTM Calibra - 420hp/450hp (Vauxhall version) Slowest DTM
    DTM A4 - 450hp (clearly fastest apart from the 350Z)

    1. The A4 was being slowed down by the GT300 Celica for the first half of the track, which is why it suddenly got faster. I mean, a 450hp DTM car versus a 300hp JGTC would be way faster of course, once free of the opposition.
    2. I really don't think you were going all out. You were paying safe, yes. There was zero aggression, and you only passed when it was safe to do so, BUT, you were overall very slow from my perspective, which is why the A4 caught up to you.

    I don't want to sound like a know-it-all, negative person, but I don't think your video truly show a rubberband effect.

    I would've agreed to a rubberband effect if the opponents all had similar specs such as all GT500 cars, or all DTM, as they're very close in overall performance, but this car roster cannot show anything considering the big difference in overall performance.
     
    Last edited: Aug 29, 2016
    C0rd0syl and JakeMR2 like this.
  25. JakeMR2

    JakeMR2

    Messages:
    1,106
    I can say one thing though, at least it isn't GT1 boost, if you have used the "Use Any Car In Arcade" the car boosts to no end, and the A.I does too. GT4 I hardly even noticed, everytime I passed them they were slow as pigs could get.
     
  26. submaniac93

    submaniac93

    Messages:
    1,332
    Location:
    France
    True. GT1 and 2 had very noticeable opponent boost, especially early in the game. GT3.. I don't know.
    In GT4, mostly in Rally races, I've always found the AI to have either more grip (I tried driving like they do, but always end up sliding around like crazy) or tires with longer lifespans. But that only adds to the challenge.
     
  27. PS2 GT4

    PS2 GT4

    Messages:
    104
    First, I didn't slow down for anything... In fact I had to drive almost to a point of crashing. I race like real life with no crashes and or game bumping crap. The Nurburgring videos seem to make things look slow, but I assure you that's not the case. I do look forward to your video posting soon and I think you'll see what I'm saying. ALL MY VIDEOS use campaign mode with cars in my garage... Not and never arcade. I used the 24 hour race in the "Endurance Hall" for the lap, so I had no choice of other competitor cars in selection either. You can work out a rough time by looking at the video length time, but you must measure it from finish line to finish line, not beginning to end... IE....VIDEO TIME TOTALS 8.43...SUBTRACT 0.26 GETTING TO THE START LINE... SUBTRACT ANOTHER 0.41 FOR THE ENDING OVER THE FINISH LINE....TIME IS 7.76. My Option stream is and was fully moded and loaded as most know and I ran medium race tires... Again looking forward to your video soon...

    Also, please keep in mind I wasn't looking for a 5 minute record lap with my Minolta or Formula... So... to compare properly and fairly, please use the Gran Turismo mode, the Endurance Hall 24 hour Nurburgring, and the same car "without" smashing it up like in a child's arcade at a pizza parlor...:D

    It's also clearly me in A Spec, as I doubt a B Spec driver pulls over at the end of a lap to look at the cars and let them pass by...

    And one more thing to ponder... please know also that it's a "start out lap" and not a "pass through lap" thereafter, which are obviously faster cause of the speed being carried over the finish line from beginning to end line...
     
    Last edited by a moderator: Dec 24, 2016
  28. submaniac93

    submaniac93

    Messages:
    1,332
    Location:
    France
    Alright! Here it is :3



    Stock (808hp with oil change) Option Stream Z. As similar a grid as I could (Weds Celica GT300 replaced with C-West S15 GT300)
    Played on controller.
     
    Last edited: Aug 29, 2016
  29. Obelisk

    Obelisk Premium

    Messages:
    8,470
    Currently setting up my experiment. I'm going to be using my DFGT wheel and the Option Stream 350Z. Now to get the right field... (My next post in this thread will be my results).

    Edit: Had an idea. Decided to change the car I'm running my experiment with to something with a lot of closely-spec'd cars. I'm running a 2005 Mazdaspeed 6 in my test. Will be doing the Nurb as well.
     
  30. PS2 GT4

    PS2 GT4

    Messages:
    104
    OK... Hmmmmm... So..... Ya may wanna share the secret to get a "stock" Option Stream to move along at 342 miles per hour on the final straight per your vid ??? Mine fully loaded doesn't even come close to that on any straight ??? Surely that might account for the lack of competition from the other cars...
     
    Last edited by a moderator: Dec 24, 2016