Gran Turismo 7 Custom Race thread
- Thread starter 05XR8
- 2,472 comments
- 363,012 views
- 50,889
- Australia
- SpacedustDaddy
The throttle percentage does factor in the programming, but it's still less than 100% with the AI still boosting to keep ahead and stay on my tail. I haven't checked throttle percentage in the SOPHY races or any other race event, for that matter.Me too, but unfortunately I still don't have all the cars needed
I had this bug yesterday in a BTCC custom race at Brands Hatch: my car (an Alfa Giulia with a Rob Austin replica livery and a BTCC-style tune), and three others, without their proper custom livery, regularly applied in the garage. Tried to restart the race, same issue. Then I exit the custom race mode, go back to the main menu, then back again to custom race mode and all was fine this time.
I wonder if this issue could depend by the use of the new "Change car" button in the circuit menu, because before going to the custom race I was actually in a completely different car, and this issue had never arisen before.
I need DTM tunes too... But I'm not in a hurry, as said before I still don't have all cars needed
I still have the AI that uses 95% of the throttle even with Boost Weak (tested yesterday in that BTCC race) and honestly I can't figure why this happens.
- 661
- France
Hi everyone !
So, Because I'm tired of having to repeat that rubber banding exists and isn't just a feeling, here is a short test I did today to show it better, supporting screenshoots. It may be quoted if useful.
3 identical races at Trial Mountain, the only difference between them being my pace. The grid is filled with my 1988 Group C cars for Le Mans. Official AG Porsche and Silk Cut Jaguar are meant to be a little bit quicker on the long run, but it doesn't really matters here. The rest of the field is meant to be on par.
CUSTOM RACE SETUP
Every penalty and flag are removed so it won't interfere for the last race. No tyre or fuel usage so it won't affect car performances. Time / Weather are fixed too for the same reasons, all of the 3 races must have the same track conditions. Boost is set off (At least that's what they're trying to make us believe).
RACE 1 :
RACE 2 :
RACE 3 :
So, to summarize the results :
Race 1 / racing into the pack
Race 2 / racing for the win
Race 3 / dead last
I tryied a 6h endurance race on LM no chicane with that very grid. If you make the choice to save fuel and overcut, you'll be P1 after AI pitted which means they will push as hell to close the gap, but you.. you still need to pit.... So you box but will be back on track dead last after everyone made its pit stop... Then, AI slows down to let you pass again..
Set the boost off / weak / high doesn't disable rubber banding, it may increase or decrease its effects in weird ways depending on your cars / track combo, but it stays. That issue has to be addressed.
So, Because I'm tired of having to repeat that rubber banding exists and isn't just a feeling, here is a short test I did today to show it better, supporting screenshoots. It may be quoted if useful.
3 identical races at Trial Mountain, the only difference between them being my pace. The grid is filled with my 1988 Group C cars for Le Mans. Official AG Porsche and Silk Cut Jaguar are meant to be a little bit quicker on the long run, but it doesn't really matters here. The rest of the field is meant to be on par.
CUSTOM RACE SETUP
Every penalty and flag are removed so it won't interfere for the last race. No tyre or fuel usage so it won't affect car performances. Time / Weather are fixed too for the same reasons, all of the 3 races must have the same track conditions. Boost is set off (At least that's what they're trying to make us believe).
RACE 1 :
The goal here is just to chill and stay out of the podium range.
My best lap was a 1'46'771, AI Jaguar set its fastest lap with a 1'45'522 on lap 2. I had to slow down from the 9th lap so I wouldn't disturb the 3 leading cars. I spun on lap 3, though...
Half of the field crossed the line after the 20 minutes mark, this is why they are so far, they have already finished their race.
My best lap was a 1'46'771, AI Jaguar set its fastest lap with a 1'45'522 on lap 2. I had to slow down from the 9th lap so I wouldn't disturb the 3 leading cars. I spun on lap 3, though...
Half of the field crossed the line after the 20 minutes mark, this is why they are so far, they have already finished their race.
RACE 2 :
Here we go again, but now it is time to push so we can see if rubber banding is happening. I took the lead on lap 7 :
So let the **** begin :
From 1'45'522 earlier during the first race to 1'41'985 now as I am in front of the pack...
Results :
First but with a short margin, #23 Nissan took the best time on lap 9. That second race is 55 seconds shorter than the first one.
Now let's take a look on that #23 Nissan's top speed before my pass for the lead, and then, as it was setting the fastest lap of the race :
No slipstream involved here. Luckily, this isn't La Sarthe without chicane where I like to take these cars out.. I chose Trial Mountain just to save a little bit of my time, but it's really scary to see at LM.
So let the **** begin :
From 1'45'522 earlier during the first race to 1'41'985 now as I am in front of the pack...
Results :
First but with a short margin, #23 Nissan took the best time on lap 9. That second race is 55 seconds shorter than the first one.
Now let's take a look on that #23 Nissan's top speed before my pass for the lead, and then, as it was setting the fastest lap of the race :
No slipstream involved here. Luckily, this isn't La Sarthe without chicane where I like to take these cars out.. I chose Trial Mountain just to save a little bit of my time, but it's really scary to see at LM.
RACE 3 :
The last test is the easiest one, stay in a corner where it won't slow down the AI :
And here are the results :
1'53'643 and one lap less? Well done AI, I'm really impressed and afraid !
And here are the results :
1'53'643 and one lap less? Well done AI, I'm really impressed and afraid !
So, to summarize the results :
Race 1 / racing into the pack
- 12 laps
- time needed : 21'38'633
- best AI lap : 1'45'522
Race 2 / racing for the win
- 12 laps
- time needed : 20'43'409
- best AI lap : 1'41'093
Race 3 / dead last
- 11 laps
- time needed : 20'54'276
- best AI lap : 1'53'643
I tryied a 6h endurance race on LM no chicane with that very grid. If you make the choice to save fuel and overcut, you'll be P1 after AI pitted which means they will push as hell to close the gap, but you.. you still need to pit.... So you box but will be back on track dead last after everyone made its pit stop... Then, AI slows down to let you pass again..
Set the boost off / weak / high doesn't disable rubber banding, it may increase or decrease its effects in weird ways depending on your cars / track combo, but it stays. That issue has to be addressed.
MonoTonouS
(Banned)
- 656
- United States
- 661
- France
We talk about a 12 seconds gap per lap (not 7), not for a duration of 20 minutes. 12 laps x 12s = a 2m24s gap between the fastest and the slowest pace. This is massive.
Last edited:
- 75
- Boston, USA
- pritchyTim
MonoTonouS
(Banned)
- 656
- United States
- 75
- Boston, USA
- pritchyTim
Thanks for all the research effort and receipts! You made this very easy to follow. Can I trouble you to run a fourth test where you have tuned your car to run at the Race 3 AI pace legitimately?
I can’t say I am optimistic but perhaps if we tuned everything PERFECTLY it could work? Race 3 seems to suggest that, without reacting to the player, the AI simply drives at THAT pace. Knowing GT, I am sure it is more convoluted and frustrating, but I think you may be very close to helping us confirm.
- 661
- France
Actually, endurance races are kind of possible with multiclass, only if you run into the lower category. This is what I did with my WEC championship, damages set on hard increased a lot the randomness though.
I'm going to try this tomorrow then.
I could even launch other races for each level of difficulty and every level of boost, but that would be a bunch of them.
- 661
- France
So here we go with a detuned car (probably not enough but still very slow)..
RACE 4 :
Losing ground until lap 5
I start to climb on lap 6 despite not being faster
I think having reached P10 is important here
Results :
AI P1 lap times :
I start to climb on lap 6 despite not being faster
I think having reached P10 is important here
Results :
AI P1 lap times :
Player VS leading AI Lap times :
I was able to overtake P19 AI into the first lap but it immediately took back its position, then AIs started to pull over from me until lap 5. Lap 6 is a turning point as I started to gain ground and then overtook them until I reached P10.
Once I was P10, the leading top 3 AI started to increase its pace. Even if I kept overtaking the others, these 3 were able to maintain the gap. I think rubber banding would have let me take 4th place but no more given the performance handicap that I inflicted on my car.
- 815
- Amsterdam
- HennyGingerale
Great vid!
I also finished my DTM grid, so much fun.
Sometimes GT7 feels like a big playbox, it's really the best racing game ever. But it feels weird that we have to go such great lengths to create these kind of races. We have all the tools to recreate some amazing historical races, with complete liveries and most accurate cars, like a DTM season. Oh and some amazing Sophy AI!
But the setup is lot of work (and ingame money) and still not possible to create custom championships etc. Or to share these amazing custom race settings.
Imagine the possibilities of that.
Edit: I currently own 33 WRX Gr. 4, 22 BMW M3 Evolution, 9 Sierra Cossies, 5 190E, 4 R33, 6 R32... All just to have some JTC/DTM/BTCC racing.
Last edited:
- 75
- Boston, USA
- pritchyTim
Thank you - the pictures really help to show what’s going on! Not that I fully understand it…
It seems lap 6 could be relevant as a midpoint in the race, as I have felt that ‘things’ start to happen around the 50% mark that give the player a fighting chance at winning/the podium. Similarly, 10th place stands out (though to a lesser degree, to me) as the 50% mark of the pack of cars.
The lap times are not completely clear, when it comes to the specifics/reasons. Aside from that second lap, it seems the lead AI car drove at a ‘natural’ pace (but not exactly the 1:53ish ‘race 3 FL’ you balanced at) until it switched into granny mode for those few laps as you climbed through the field. Once you reached 10th, it went back to natural mode. Why the second (fastest) lap, though? And why not show similar times to Race 3?
Is it fair for me to assume the top 3 were the only ones that increased pace when you reached 10th? I can see you still passed a few more, and you commented you would’ve been eligible for 4th if given enough time. I’d think we’d see more of the granny times for the rest of the AI pack. Still curious about that second lap…
I guess I am as properly confused (and frustrated) as before? haha. “Weak boost” comes to mind as another tool in our toolbox, but it’s evident the rubberband is very in play and seemingly more difficult to ‘match’ than I had hoped! Mixed feelings about continuing the investigation, I’m not sure it can have a perfectly happy ending.
- 661
- France
First two laps look more like a "sample standard" pace for AI (yes, I guess AI laps at a similar pace on lap 1, it's just that the timer isn't kept). Nor aggressive, nor very slow, just the basic rhythm AI will follow in every race until it starts to react to player's position.
- 50,889
- Australia
- SpacedustDaddy
Revisiting my 1973 Daytona 24H top 19 finishers grid. Waiting for the “washing machine” to return. It’s the closest we have to the Porsche 908/02(finished 12th overall) from that race.
- 3,567
- London, ON
- A_Mindful_Place
- 338
- England
Now that money is a bit easy to come by I’m thinking of recreating the 2016 wec season. Just need to buy duplicates of the gr1 cars
- 118
- United States
The tires issue is pretty annoying. I race against all of my tuned time attack cars, only to see them get nerfed by RH tires.
- 50,889
- Australia
- SpacedustDaddy
Trans Am Street
Been waiting for a Challenger to join the game again.
Boost Weak. AI we’re loving these. Got my behind whooped. Of course, computer cars have all aids on. With 801BHP(Dodge), 761BHP(Ford) and 718BHP(Chevy), all limited to about 300kph, makes it very lively in tight groups. Couldn’t move up the grid much and mad a mistake at the end of the penultimate lap. Didn’t get to complete the last lap as time ran out.
I‘m happy with this first test.
Been waiting for a Challenger to join the game again.
Boost Weak. AI we’re loving these. Got my behind whooped. Of course, computer cars have all aids on. With 801BHP(Dodge), 761BHP(Ford) and 718BHP(Chevy), all limited to about 300kph, makes it very lively in tight groups. Couldn’t move up the grid much and mad a mistake at the end of the penultimate lap. Didn’t get to complete the last lap as time ran out.
I‘m happy with this first test.
- 1,133
- Murray Bridge
- ItzRacing
Hey everyone, long time reader first time poster in this thread!
I can definitely say I’ve been a big fan of the inclusion of the custom races since GT7 was released, and I knew it was something that would help me live out all of those racing fantasies from years gone by. Having used the custom races a little bit on GT Sport, I did think the idea of being able to add a full field of cars with your own liveries and setups wouldI transform the mode into something special which turned out to be true. It has certainly created a number of ideas for races and custom championships. It’s a shame that the tyre bug is still here and sometimes the rubberbanding can be frustrating in longer races, but it’s still a great experience nonetheless. I have my fingers crossed with that thing that’s become popular recently, that maybe PD will give some attention to some other features of custom races that need fixing.
I thought I would share a project I’ve been working on this year. I’ve trialled a hybrid Gen 2/Gen 3 Supercar Series with the Camaro SS, Mustang GT, Lexus RC F and BMW M4. Instead of doing an accurate grid, I instead created some fantasy teams with friends and family as drivers/managers to make a 20 car grid. I’ve had a few teething issues with balancing the performance of the cars early on (much like the real thing this year) and then once I got that sorted I did some endurance races where fuel consumption across the 4 cars became an issue as well. Thankfully I had a 500km race that helped make me aware of the problems and allowed me to make some fixes for running a Bathurst 1000 race. The fixes worked a lot better and all cars seemed to be on a more even playing field and created an interesting race watching the teams all move up and down the order!
It’s been a nice little trial run in 2023 and I’ll be making some changes for a 2024 series which will see the M4 dropped and maybe even the RC F going as well in favour of a Mustang v Camaro series. I’ve also started building a few different car sets like a Muscle Car Series, Truck Series, Mini Challenge, Porsche Cup and even a couple of Formula Series using the SF23 and Red Bull Junior.
Here are some photos I’ve grabbed from the opening 2 rounds of my Touring Car Series at Dragon Trail and Deep Forest!
I can definitely say I’ve been a big fan of the inclusion of the custom races since GT7 was released, and I knew it was something that would help me live out all of those racing fantasies from years gone by. Having used the custom races a little bit on GT Sport, I did think the idea of being able to add a full field of cars with your own liveries and setups wouldI transform the mode into something special which turned out to be true. It has certainly created a number of ideas for races and custom championships. It’s a shame that the tyre bug is still here and sometimes the rubberbanding can be frustrating in longer races, but it’s still a great experience nonetheless. I have my fingers crossed with that thing that’s become popular recently, that maybe PD will give some attention to some other features of custom races that need fixing.
I thought I would share a project I’ve been working on this year. I’ve trialled a hybrid Gen 2/Gen 3 Supercar Series with the Camaro SS, Mustang GT, Lexus RC F and BMW M4. Instead of doing an accurate grid, I instead created some fantasy teams with friends and family as drivers/managers to make a 20 car grid. I’ve had a few teething issues with balancing the performance of the cars early on (much like the real thing this year) and then once I got that sorted I did some endurance races where fuel consumption across the 4 cars became an issue as well. Thankfully I had a 500km race that helped make me aware of the problems and allowed me to make some fixes for running a Bathurst 1000 race. The fixes worked a lot better and all cars seemed to be on a more even playing field and created an interesting race watching the teams all move up and down the order!
It’s been a nice little trial run in 2023 and I’ll be making some changes for a 2024 series which will see the M4 dropped and maybe even the RC F going as well in favour of a Mustang v Camaro series. I’ve also started building a few different car sets like a Muscle Car Series, Truck Series, Mini Challenge, Porsche Cup and even a couple of Formula Series using the SF23 and Red Bull Junior.
Here are some photos I’ve grabbed from the opening 2 rounds of my Touring Car Series at Dragon Trail and Deep Forest!
Attachments
Last edited:
- 3,567
- London, ON
- A_Mindful_Place
Photo mode during custom race replays can be mesmerizing sometimes.
Similar threads
