Gran Turismo 7 Custom Race thread

[Meythia]

I have made an observation that helps to explain why Sophy can suffer with the GT3 cars and some street set ups. It appears that Sophy is very abrupt with it's application of brakes and throttle, which means, it is likely engaging the game's stability systems. And, many of us know that those systems slow you down.

That is likely down to Sophy excepting Gr.3 cars to be used with any racing tyre by default.

 

[Voodoovaj]

That is likely down to Sophy excepting Gr.3 cars to be used with any racing tyre by default.

But also, many on this forum drive the Gr.3 cars with TC turned on, because they find it challenging to drive without it. Sophy is likely triggering the TC as well, which will slow it down.

 

[suomi1]

So I let Sophy a last chance at Le Mans, trying to reverse the roles (2015 LM field). The idea was to slow down AI cars until they drive more or less on the same pace as me when rubberbanding hits its maximal effect. Here I could set 3'28 laps, so I tuned down AI cars so they would hit this target.

It took me a lot of time and 10 races to set up the BoP correctly, 6 races of 10 laps for the LMP1, 4 races of 6 laps for the LMGTE Am. Still the results were very convincing.

I did a 4 lap qualifying race to determine the starting order that gave this :

Spoiler

1 - Audi #7 - 3:28.240
2 - Mazda #55 - 3:28.390 (me, obviously)
3 - Toyota #2 - 3:28.418
4 - Audi #8 - 3:28:506
5 - Porsche #19 - 3:28:508
6 - Audi #9 - 3:28.546
7 - Toyota #1 - 3:28.580
8 - Porsche #17 - 3:28.607
9 - Porsche #18 - 3:28.964
10 - Bykolles #4 - 3:30.442
11 - Rebellion #13 - 3:31.077
12 - Nissan #23 - 3:31.660
13 - Rebellion #12 - 3:31.680
14 - Nissan #22 - 3:31.726
15 - Nissan #21 - 3:31.834
16 - Ferrari #66 - 3:50.606
17 - Dodge #53 - 3:50.784
18 - Porsche #77 - 3:50.794
19 - Aston Martin #96 - 3:50.894
20 - Chevrolet #50 - 3:50.950

Pretty close actually, even if some cars (#12 / #13) were trapped behind a Nissan. Privateers and Nissan are slow on purpose.

I went for a 6 hours race but didn't go all the way because, yeah, it was trash despite being good on the paper... So 2 hours + were more than enough to see where all of this was going. From the beginning actually. I immediately took the lead and Sophy refused to make a pass, even after putting the wheels in the gravel several times.

I went off at "Virage du Pont" on the 11th lap and had to pit to repair the damages. Now look at this, my lap times and AI's before / after I crashed and fell last in class :

Spoiler

And here is AI's best lap so we can compare :

Spoiler

It slowed down by around 17 seconds. 17 seconds... I knew that this would happen, but not to this degree considering Sophy is far more steady. 5 seconds I wouldn't have said anything, totally tolerable. Not 17, lol. You can't race properly at Le Mans, whether with basic AI or Sophy.

It's sad because setting BoP for race cars is far more reliable with Sophy, plus the whole field tries to set the best possible laps even when they are last. With PD's AI, the further down the leaderboard a car is, the slower it will run regardless of its actual potential, not Sophy.
That being said, I remain delighted with my BoP, the five GT cars were still packed together after 2 hours, same goes for top cars in the LMP1 category.

I stopped the race after this figure, I was completely disillusioned and unfocused :

We'll meet up again the day rubberbanding disappears. See ya!

 

[05XR8]

So I let Sophy a last chance at Le Mans, trying to reverse the roles (2015 LM field). The idea was to slow down AI cars until they drive more or less on the same pace as me when rubberbanding hits its maximal effect. Here I could set 3'28 laps, so I tuned down AI cars so they would hit this target.

It took me a lot of time and 10 races to set up the BoP correctly, 6 races of 10 laps for the LMP1, 4 races of 6 laps for the LMGTE Am. Still the results were very convincing.

I did a 4 lap qualifying race to determine the starting order that gave this :

Spoiler

1 - Audi #7 - 3:28.240
2 - Mazda #55 - 3:28.390 (me, obviously)
3 - Toyota #2 - 3:28.418
4 - Audi #8 - 3:28:506
5 - Porsche #19 - 3:28:508
6 - Audi #9 - 3:28.546
7 - Toyota #1 - 3:28.580
8 - Porsche #17 - 3:28.607
9 - Porsche #18 - 3:28.964
10 - Bykolles #4 - 3:30.442
11 - Rebellion #13 - 3:31.077
12 - Nissan #23 - 3:31.660
13 - Rebellion #12 - 3:31.680
14 - Nissan #22 - 3:31.726
15 - Nissan #21 - 3:31.834
16 - Ferrari #66 - 3:50.606
17 - Dodge #53 - 3:50.784
18 - Porsche #77 - 3:50.794
19 - Aston Martin #96 - 3:50.894
20 - Chevrolet #50 - 3:50.950

Pretty close actually, even if some cars (#12 / #13) were trapped behind a Nissan. Privateers and Nissan are slow on purpose.

I went for a 6 hours race but didn't go all the way because, yeah, it was trash despite being good on the paper... So 2 hours + were more than enough to see where all of this was going. From the beginning actually. I immediately took the lead and Sophy refused to make a pass, even after putting the wheels in the gravel several times.

I went off at "Virage du Pont" on the 11th lap and had to pit to repair the damages. Now look at this, my lap times and AI's before / after I crashed and fell last in class :

Spoiler

View attachment 1463481

View attachment 1463482

View attachment 1463483

View attachment 1463484

And here is AI's best lap so we can compare :

Spoiler

View attachment 1463485

It slowed down by around 17 seconds. 17 seconds... I knew that this would happen, but not to this degree considering Sophy is far more steady. 5 seconds I wouldn't have said anything, totally tolerable. Not 17, lol. You can't race properly at Le Mans, whether with basic AI or Sophy.

It's sad because setting BoP for race cars is far more reliable with Sophy, plus the whole field tries to set the best possible laps even when they are last. With PD's AI, the further down the leaderboard a car is, the slower it will run regardless of its actual potential, not Sophy.
That being said, I remain delighted with my BoP, the five GT cars were still packed together after 2 hours, same goes for top cars in the LMP1 category.

I stopped the race after this figure, I was completely disillusioned and unfocused :

View attachment 1463490


We'll meet up again the day rubberbanding disappears. See ya!

I think you had posted the video of the R92CP at Lemans slowing down. In your testing and results, do you feel it's the programming for SOPHY or the programming to slow the car's performance?

 

[suomi1]

I think you had posted the video of the R92CP at Lemans slowing down. In your testing and results, do you feel it's the programming for SOPHY or the programming to slow the car's performance?

That's the rubberband programming, not Sophy, the same applied to PD's AI which is known to be competitive when not affected by the same rubberband effect.

Edit : about the R92CP, it was at Trial Mountain.

 

[Marty_KAR]

We'll meet up again the day rubberbanding disappears. See ya!

Like a lot of you peeps here, I too have been researching SOPHY and Regular AI races, and SOPHY is definitely a better and more challenging opponent (Boost set to Weak, ...always).

Last weekend I did a few races (at Le Mans) where, after 2-hours, SOPHY wins quite convincingly, or after I limit the (SOPHY) field a bit, and I'm in a faster (more powerful) car, I can only just get the win, ...under extreme pressure from SOPHY, ....and almost unrealistic pressure, ...like SOPHY has nitrous on the straights, or their worn tires suddenly have renewed grip through corners and corner exit 😒

When I'm racing SOPHY (or Regular AI) I don't want to have "wins" handed to me on a plate, ...but when I do work hard to create a winning gap to SOPHY, I also don't want to see, what appears to be, "sudden & magical" (unrealistic) performance from SOPHY either.

Just for ****s & giggles, ...today I'm going to Le Mans and climb aboard a Group 1 car, and I'll give SOPHY Group 2 cars. After 2-hours, I should win by a few laps, ...Right?

I'll post my findings later.

 

[Nebuc72]

Just for ****s & giggles, ...today I'm going to Le Mans and climb aboard a Group 1 car, and I'll give SOPHY Group 2 cars. After 2-hours, I should win by a few laps, ...Right?

I'll post my findings later.

I do believe with the new BoPs some of the Gr.2 cars are faster than some Gr.1 cars so, depends what you pick

 

[Marty_KAR]

I do believe with the new BoPs some of the Gr.2 cars are faster than some Gr.1 cars so, depends what you pick

Thanks Nebuc72.

My aim is to see if I can win against SOPHY by a large margin, so I'll definitely consider the cars. Thanks for the heads-up 👍

 

[Theufcveteran]

My aim is to see if I can win against SOPHY by a large margin

May I recommend a nuclear powered jet engine?

Jokes aside, best of luck! I’d be interested in reading your findings. Do a Lewis Hamilton 2020 on them!

 

[suomi1]

So I let Sophy a last chance at Le Mans, trying to reverse the roles (2015 LM field). The idea was to slow down AI cars until they drive more or less on the same pace as me when rubberbanding hits its maximal effect. Here I could set 3'28 laps, so I tuned down AI cars so they would hit this target.

It took me a lot of time and 10 races to set up the BoP correctly, 6 races of 10 laps for the LMP1, 4 races of 6 laps for the LMGTE Am. Still the results were very convincing.

I did a 4 lap qualifying race to determine the starting order that gave this :

Spoiler

1 - Audi #7 - 3:28.240
2 - Mazda #55 - 3:28.390 (me, obviously)
3 - Toyota #2 - 3:28.418
4 - Audi #8 - 3:28:506
5 - Porsche #19 - 3:28:508
6 - Audi #9 - 3:28.546
7 - Toyota #1 - 3:28.580
8 - Porsche #17 - 3:28.607
9 - Porsche #18 - 3:28.964
10 - Bykolles #4 - 3:30.442
11 - Rebellion #13 - 3:31.077
12 - Nissan #23 - 3:31.660
13 - Rebellion #12 - 3:31.680
14 - Nissan #22 - 3:31.726
15 - Nissan #21 - 3:31.834
16 - Ferrari #66 - 3:50.606
17 - Dodge #53 - 3:50.784
18 - Porsche #77 - 3:50.794
19 - Aston Martin #96 - 3:50.894
20 - Chevrolet #50 - 3:50.950

Pretty close actually, even if some cars (#12 / #13) were trapped behind a Nissan. Privateers and Nissan are slow on purpose.

I went for a 6 hours race but didn't go all the way because, yeah, it was trash despite being good on the paper... So 2 hours + were more than enough to see where all of this was going. From the beginning actually. I immediately took the lead and Sophy refused to make a pass, even after putting the wheels in the gravel several times.

I went off at "Virage du Pont" on the 11th lap and had to pit to repair the damages. Now look at this, my lap times and AI's before / after I crashed and fell last in class :

Spoiler

View attachment 1463481

View attachment 1463482

View attachment 1463483

View attachment 1463484

And here is AI's best lap so we can compare :

Spoiler

View attachment 1463485

It slowed down by around 17 seconds. 17 seconds... I knew that this would happen, but not to this degree considering Sophy is far more steady. 5 seconds I wouldn't have said anything, totally tolerable. Not 17, lol. You can't race properly at Le Mans, whether with basic AI or Sophy.

It's sad because setting BoP for race cars is far more reliable with Sophy, plus the whole field tries to set the best possible laps even when they are last. With PD's AI, the further down the leaderboard a car is, the slower it will run regardless of its actual potential, not Sophy.
That being said, I remain delighted with my BoP, the five GT cars were still packed together after 2 hours, same goes for top cars in the LMP1 category.

I stopped the race after this figure, I was completely disillusioned and unfocused :

View attachment 1463490


We'll meet up again the day rubberbanding disappears. See ya!

So just for the record, I recreated the exact same race but with Polyphony's AI (and my personnal BoP for Sophy). I simulated a crash at the end of lap 6, entered the pit lane and went back on track P11 and 45 seconds behind the leading car. I quickly took back the lead at lap 9 between Tertre Rouge and the first chicane.

Here are the lap times (don't mind the "S" indicating a pit stop, it's always bugged) :

Spoiler

That's roughly a 23 seconds slow-down against 17 seconds for Sophy. Worst indeed, but still very bad in both scenarios.

 

[bliprunner]

So just for the record, I recreated the exact same race but with Polyphony's AI (and my personnal BoP for Sophy). I simulated a crash at the end of lap 6, entered the pit lane and went back on track P11 and 45 seconds behind the leading car. I quickly took back the lead at lap 9 between Tertre Rouge and the first chicane.

Here are the lap times (don't mind the "S" indicating a pit stop, it's always bugged) :

Spoiler

View attachment 1463659

View attachment 1463660

View attachment 1463661

View attachment 1463662

That's roughly a 23 seconds slow-down against 17 seconds for Sophy. Worst indeed, but still very bad in both scenarios.

Here's a suggestion to possibly make this work without rubber banding ruining it:

• Use Sophy with Boost Weak.
• If you want to drive in LMP1 add a "Special Prototype" entry that is so fast you'll never overtake it.
• Or drive in LMGT so you'll never overtake the LMP1 leader.

With Boost Weak all car lap times should remain consistent and Sophy won't slow or speed up - as long as you don't take the lead - this is the key.

 

[suomi1]

Here's a suggestion to possibly make this work without rubber banding ruining it:

• Use Sophy with Boost Weak.
• If you want to drive in LMP1 add a "Special Prototype" entry that is so fast you'll never overtake it.
• Or drive in LMGT so you'll never overtake the LMP1 leader.

With Boost Weak all car lap times should remain consistent and Sophy won't slow or speed up - as long as you don't take the lead - this is the key.

I know all of this and I already kind of tried. That was my best race so far at LM under those conditions.
That's a good idea I have considered, but in this case it comes with the sacrifice of immersion in many ways.

First I will have to tune down my car until I do 3:42 or 3:43 laps at best which looks lurid for a top class car. Then I will have to remove an interesting car to put an UFO instead, with heavy mechanical damages it might end ugly for no reason. I don't want to drive LMGTE cars, this I already did many times and a full multiclass championship too. There is an other issue with fuel economy for the AI able to do far longer stints, this will stay whatever I try.

The best answer stays to set a 3 category race and select a car from the intermediate class. Of course, this will only work if you are into one of the cars running in the middle of the pack. That's not the goal here.

In any case, we shouldn't be in a situation where we need such stratagems, hours and hours of tests and discussions (not to mention grinding, lol...) in order to obtain a race that is barely acceptable.

 

[Revengel]

I know all of this and I already kind of tried. That was my best race so far at LM under those conditions.
That's a good idea I have considered, but in this case it comes with the sacrifice of immersion in many ways.

PSVR2 & GT Race Engineer FTW! (sorry, shameless plug for immersion)

In any case, we shouldn't be in a situation where we need such stratagems, hours and hours of tests and discussions (not to mention grinding, lol...) in order to obtain a race that is barely acceptable.

I have a different perspective on this.

First, my frame of (relevant) reference: In single player I mostly use road cars on street tires. I used to race against Reggie and had to increase the power of the field to get a solid race; I now race against Sophy in equal machinery. That's been detailed on a couple of previous pages.

With that understood I suspect - while understanding everyone can have different results - you would likely have a better time racing against road cars on these tracks in terms of how competitive Sophy is.

In the past I used Sophy as a warm-up for GTWS events (where applicable; some races were not on Sophy tracks) just to get some warmup time in traffic and to measure tire wear. I would certainly do the occasional full race, but I wasn't out for a victory. Those races were (almost) purely to measure tire wear and get some door-to-door practice. The few full races I did were enjoyable for me; I'm not going to pretend I've put as much effort into Sophy using race cars as you have.

I believe I have put in at least as much time regarding road cars, however.

When I want an enjoyable race, I go to road cars. Part of that is I have an H-pattern shifter in my setup. Part of that is pretending I'm racing against some characters/people for fun. Part of it is I enjoy the threat of heavy damage and the skills that come when you need to save fuel. In this regard, Sophy is unmatched in my personal experience. Based on my 17-track series I've found that there are certainly cars quicker than mine, but fuel plays a factor. For even longer races tire wear also rears it's ugly head . . . but I digress. You already stated you use both fuel and damage; I just wanted to disclose I use them as well.

Feel free to try road cars a few times; I'd be happy to share what I use if that helps.


As always, just an opinion. Thank you for reading.

 

[bliprunner]

I know all of this and I already kind of tried. That was my best race so far at LM under those conditions.
That's a good idea I have considered, but in this case it comes with the sacrifice of immersion in many ways.

First I will have to tune down my car until I do 3:42 or 3:43 laps at best which looks lurid for a top class car. Then I will have to remove an interesting car to put an UFO instead, with heavy mechanical damages it might end ugly for no reason. I don't want to drive LMGTE cars, this I already did many times and a full multiclass championship too. There is an other issue with fuel economy for the AI able to do far longer stints, this will stay whatever I try.

The best answer stays to set a 3 category race and select a car from the intermediate class. Of course, this will only work if you are into one of the cars running in the middle of the pack. That's not the goal here.

In any case, we shouldn't be in a situation where we need such stratagems, hours and hours of tests and discussions (not to mention grinding, lol...) in order to obtain a race that is barely acceptable.

Yes, very much agree, all the messing about shouldn’t be necessary.

 

[suomi1]

I have a different perspective on this.

First, my frame of (relevant) reference: In single player I mostly use road cars on street tires. I used to race against Reggie and had to increase the power of the field to get a solid race; I now race against Sophy in equal machinery. That's been detailed on a couple of previous pages.

With that understood I suspect - while understanding everyone can have different results - you would likely have a better time racing against road cars on these tracks in terms of how competitive Sophy is.

In the past I used Sophy as a warm-up for GTWS events (where applicable; some races were not on Sophy tracks) just to get some warmup time in traffic and to measure tire wear. I would certainly do the occasional full race, but I wasn't out for a victory. Those races were (almost) purely to measure tire wear and get some door-to-door practice. The few full races I did were enjoyable for me; I'm not going to pretend I've put as much effort into Sophy using race cars as you have.

I believe I have put in at least as much time regarding road cars, however.

When I want an enjoyable race, I go to road cars. Part of that is I have an H-pattern shifter in my setup. Part of that is pretending I'm racing against some characters/people for fun. Part of it is I enjoy the threat of heavy damage and the skills that come when you need to save fuel. In this regard, Sophy is unmatched in my personal experience. Based on my 17-track series I've found that there are certainly cars quicker than mine, but fuel plays a factor. For even longer races tire wear also rears it's ugly head . . . but I digress. You already stated you use both fuel and damage; I just wanted to disclose I use them as well.

Feel free to try road cars a few times; I'd be happy to share what I use if that helps.


As always, just an opinion. Thank you for reading.

This I have tried too, it looks better at first but as soon as a gap opens behind or in front of you it's the same thing all over. You can even mix race cars with road cars and it "works", you can put AIs into GR010 against Valkyries and see them having fair battles.

I spend a lot of time with road cars too, it's a lot of fun. Yet the way I enjoy the most road cars is into "stupid" grids where I mix max tuned road cars and stock ones just for the sake of it. The aim isn't to race, but to take a stroll and cross paths with little / old cars while being overtaken by crazy rockets lol.

It goes from a Nissan powered Alpine A110 to a stock Renault 4L. Sure rubberbanding is the last of my problems in this situation :

 

[2cam16]

Did a small-sided all Ford race today with the cars I had in my garage. One at RBR, the other at Dragon Trail.
All on RH. I tried to even the field of 9 cars out and it seems pretty close. The Dragon Trail race was fun! 8 laps and we were all bunched up for the most part. At times we were 3 and 4 wide! Sophy was giving it a good go. I barely won it by less than half a second.
Red Bull Ring_ (23) by 2cam16, on Flickr
Dragon Trail - Seaside__1 (22) by 2cam16, on Flickr
Dragon Trail - Seaside_ (29) by 2cam16, on Flickr

 

[Yard_Sale]

This might be the best Sophy race I’ve done yet. 18 laps. Started at 20th. Fuel x7. Tires x6.



My 81’ Porsche is tuned to 653pp I think. All other cars are tuned anywhere from 679pp to 660pp, depending on the car. My grid is starting to get close to ready for my series 😎

 

[05XR8]

Not much else I can say about GT SOPHY.

 

[suomi1]

• If you want to drive in LMP1 add a "Special Prototype" entry that is so fast you'll never overtake it.

I tried this even if that was at the cost of immersion, and it worked to the perfection! Thank you, you probably saved the rest of my journey on GT7.

So here it is, I placed a Red Bull at the top of the field instead of the drunk "ByKolles Audi" I removed. I raced normally until the end of lap 5. The 6th loop was a deceleration lap, I was 1:05 in front of the LMP1 pack at that point. I entered the pit lane just as I was about to be caught up, to not interfere with their lap times. Lap 7 was another slow one so AI could escape and continue to set their laps without me overtaking them too quickly.
I went back to a normal rhythm from the 8th lap, being 30 seconds behind the slow privateers and Nissan. Gaps weren't shown in the replay but here is were I was when I started to race again :

And here are the lap times from Audi, Porsche and Toyota :

Spoiler

The method works perfectly. APPROVED!
AIs do not give a **** about where I am as long as the Red Bull stays far in front.
Now I just need to remove some of their ballast and lower my max power until my laptimes and the AI's match.

My BoP works nicely too, I watched the entire replay just to follow the battle in LMGTE.
On a side note, I didn't enabled the flags rule, so cars do not brake dangerously, nor it destroys the battle in the lower category when they're laped. They will fight in front of you even if you are about to lap them, no matter what. As it should be, you're the one to be aware of what you do when you go around them.

 

[Aedum]

Here's a suggestion to possibly make this work without rubber banding ruining it:

• Use Sophy with Boost Weak.
• If you want to drive in LMP1 add a "Special Prototype" entry that is so fast you'll never overtake it.
• Or drive in LMGT so you'll never overtake the LMP1 leader.

With Boost Weak all car lap times should remain consistent and Sophy won't slow or speed up - as long as you don't take the lead - this is the key.

So this should work for any custom race. Add a faster car to the field and run the actual race against the 19 other cars on the track and 2nd place is the winner. Set up for time instead of laps so the race doesn't end until the time is over. Should even work for 24 hour races.

 

[suomi1]

So this should work for any custom race. Add a faster car to the field and run the actual race against the 19 other cars on the track and 2nd place is the winner. Set up for time instead of laps so the race doesn't end until the time is over. Should even work for 24 hour races.

Not with regular AI though. Both Sophy and PD's AI are plagued with the rubberbanding virus, but PD's AI in custom races has also the "GT Cafe script" applied to it.

 

[bliprunner]

I tried this even if that was at the cost of immersion, and it worked to the perfection! Thank you, you probably saved the rest of my journey on GT7.

So here it is, I placed a Red Bull at the top of the field instead of the drunk "ByKolles Audi" I removed. I raced normally until the end of lap 5. The 6th loop was a deceleration lap, I was 1:05 in front of the LMP1 pack at that point. I entered the pit lane just as I was about to be caught up, to not interfere with their lap times. Lap 7 was another slow one so AI could escape and continue to set their laps without me overtaking them too quickly.
I went back to a normal rhythm from the 8th lap, being 30 seconds behind the slow privateers and Nissan. Gaps weren't shown in the replay but here is were I was when I started to race again :

View attachment 1463815

And here are the lap times from Audi, Porsche and Toyota :

Spoiler

View attachment 1463811

View attachment 1463812

View attachment 1463813

The method works perfectly. APPROVED!
AIs do not give a **** about where I am as long as the Red Bull stays far in front.
Now I just need to remove some of their ballast and lower my max power until my laptimes and the AI's match.

My BoP works nicely too, I watched the entire replay just to follow the battle in LMGTE.
On a side note, I didn't enabled the flags rule, so cars do not brake dangerously, nor it destroys the battle in the lower category when they're laped. They will fight in front of you even if you are about to lap them, no matter what. As it should be, you're the one to be aware of what you do when you go around them.

Nice, glad that worked out!

 

[Aedum]

Not with regular AI though. Both Sophy and PD's AI are plagued with the rubberbanding virus, but PD's AI in custom races has also the "GT Cafe script" applied to it.

So I should have said any custom race with Sophy AI Boost set to Weak. Either way sounds like a good work around for the rubber banding in the meantime. Cool fix I wouldn't have thought of.

 

[Marty_KAR]

So here it is, I placed a Red Bull at the top of the field...

The method works perfectly. APPROVED!
AIs do not give a **** about where I am as long as the Red Bull stays far in front.
Now I just need to remove some of their ballast and lower my max power until my laptimes and the AI's match.

On a side note, I didn't enabled the flags rule, so cars do not brake dangerously, nor it destroys the battle in the lower category when they're laped. They will fight in front of you even if you are about to lap them, no matter what. As it should be, you're the one to be aware of what you do when you go around them.

Red Bull - CHECK!

No Flags - CHECK!

Thank you 😊👍

 

[pritchyTim]

Not with regular AI though. Both Sophy and PD's AI are plagued with the rubberbanding virus, but PD's AI in custom races has also the "GT Cafe script" applied to it.

Can you elaborate? Not sure I follow that quotable - is ‘cafe script’ the bit where it tries to hand you a favorable finish in the last couple of laps? Thanks

 

[Theufcveteran]

Doing a little promotion for Integrale’s July update! A gameplay video of Arcade Mode at Suzuka against Sophy.

 

[Marty_KAR]

Because you guys and gals have already done a ton of research into Custom Races, what I have to say won't come as a surprise to many here, ...but some may find my latest SOPHY Custom Race research interesting.

My Custom Races are mostly endurance type events against SOPHY, with Boost set to weak, in similar Cars, and Damage on.

I have found that SOPHY either wins quite convincingly, or I just manage to just scrape into the top 5. This is fine, but how can I actually kick SOPHY's butt by a very l--o--n--g margin?

SETUP RACE A (Le Mans with Chicanes): Me (Group 1 - Nissan R92CP '92) v SOPHY (Groups 2, 3 & 4)
* No good. After taking the lead, I'm getting tail gated by a couple of fast SOPHY Group 2 cars.

ADJUSTMENT FROM SETUP RACE A:
1. I love the top-end speed of my Nissan, but I swap it for the Peugeot 908 HDi FAP, which I am 3-seconds a lap faster in.

SETUP RACE B (Le Mans with Chicanes): Me (Group 1 - Peugeot 908 HDi FAP v SOPHY (Groups 2, 3 & 4)
* No good. After taking the lead, I am too slow on the straights, and the fast Group 2 cars overtake me on the straights.

ADJUSTMENTS FROM SETUP RACE B:
1. I revert back to the Nissan R92CP '92, as it is great on all the straights at Le Mans.
2. I swap those fast SOPHY Group 2 cars with the slower BMW F1 GTR, and the Honda NSX GT500 '08.

RACE #1: 2-Hours at Le Mans: Me (Group 1 - Nissan R92CP '92) v SOPHY (Groups 2, 3 & 4)

• SOPHY's BMW F1 is my rival, and it pits 10-times.
• I take the lead on lap 4, and I pit 6-times.
• After 2-hours I win by a margin of only 1-minute, and 18-seconds.

* During the race I notice that after each of my pit stops, 2nd place SOPHY will cut my lead an average of 6-seconds per "out-lap".

* During the race the "gap" from me to 2nd place always remained constant, ...and despite me doing purple sector times, lap after lap, I cannot extend my lead from SOPHY.

ADJUSTMENT FROM RACE #1:
1. I delete ALL of SOPHY's Group 2 cars.

RACE #2: 2-Hours at Le Mans: Me (Group 1 - Nissan R92CP '92) v SOPHY (Groups 3 & 4)

• I take the lead on lap 2, and I ended up pitting 6-times (just like Race 1).
• SOPHY front runners pit twice in the first 5-laps. They continue to pit numerous times, and I cannot keep count.
• Despite SOPHY's many pit stops, I cannot get more than 2-minutes and 50-seconds in front - so the Timing Screen tells me 😒
• On lap 20, I finally lap the 2nd placed SOPHY Group 3 car (Aston Martin DBR9 GT1 '10) and then the timing screen disappears.
• For the remaining 10-laps of the race I have no timing screen.
• After 2-hours I win by a margin of only 1-lap.

* During the race, a lowly placed SOPHY Group 3 Toyota GR Supra, that I was about to lap, exited the pits in front of me, and sped off and disappeared into the distance. WTF?

* Additionally, another lowly placed SOPHY Group 3 Dodge Viper SRT GT3-R '15, exited the pits way behind me, and within half a lap it was tailgating me. Again, ...WTF?

* Each Pit Stop takes approximately 1-minute to complete. SOPHY Pitted at least 12 times compared to my 6, ....and I only win by 1-lap, ...in a much faster car (???)

SUMMARY
I failed to kick SOPHY's Butt.

I just want to get as close to race realism, whether I'm winning or losing, with GT7. There's no doubting the great driving experience of GT7, but this rubber banding nonsense during the racing is stupid, ...and it's such a shame that we all have to work hard to setup time-consuming workarounds to achieve a more realistic GT7 racing experience.

For my next Custom SOPHY Race, I'm going to try the Red Bull UFO and the no Flags Method, and race SOPHY for second place.
First place is for losers 😒

 

[zjn]

One note and one question from my race yesterday

Note:
Never use Damage:severe, it's maybe ok for the player, but AI goes to pit as soon as it has any damage to repair it

Question:
why AI uses TC and ASM? I just discovered it looking at the replay to figure out why P1 (AI) want to PIT after one lap in a race with X1 fuel

 

[suomi1]

Can you elaborate? Not sure I follow that quotable - is ‘cafe script’ the bit where it tries to hand you a favorable finish in the last couple of laps? Thanks

At first, I thought what I now identify as "GT Cafe script" was like the mood jauge in GT5's B-spec (When b-spec drivers are fighting, they become angry and drive faster, it goes the other way when they find themselves alone). But I eventually realized that in GT7, their pace was not linked to their proximity to another AI, but to their position in the ranking.

PD's AI can be fast and steady (cf. pepper races and Human Comedy missions), Sophy is faster and steady too. But it's as if they've scripted their AI to fit the structure of the career mode, which is GT Cafe (At least for Menu Books that were present at the game's launch).
Your goal is to enter the top 3, as so, AI tends to create a cluster of 3 cars ahead of the field, it goes back to 2 if you complete the top 3. Cars behind loose pace in a decreasing way the more you look down into the rankings. The last car is crawling, so the player won't finish last and be humiliated. Of course, races in career mode are short and you might not spot this behavior, yet it is what happens. But on a longer custom race, even more if you create a monotype race, it becomes pretty obvious. I ended setting my BoPs car by car, because cars that start behind would always crawling and distort my tests.

Here is a fictitious graph of the distribution of best laps for a grid of 10 AIs, assuming the human player is far ahead :

AIs we've got in Custom Race mode are "SOPHY + Rubberband" or "PD's + Cafe + Rubberband".
Boost is just another layer of setup for the rubber banding effect.

For my next Custom SOPHY Race, I'm going to try the Red Bull UFO and the no Flags Method, and race SOPHY for second place.
First place is for losers 😒

I took the Red Bull because I was driving into LMP1 class, feel free to choose any car you like as long as it stays in front. I'm planning to make it heavier, because its cornering speed makes it too dangerous. The flags are up to you, but activating them won't break the race like rubberbanding might.

 

[suomi1]

One note and one question from my race yesterday

View attachment 1463915
Note:
Never use Damage:severe, it's maybe ok for the player, but AI goes to pit as soon as it has any damage to repair it

I find it funnier, and since Sophy is more careful, it happens less often.