AI wet qualifying pace issue (Test results)

[Worntoathread 865]

After the discussion in the other topic, I decided to do some testing to try and find out if there is an issue with AI pace in the wet.

I ran three qualifying sessions in different weather types at Barcelona, and a race in full wet conditions to compare the wet vs dry pace. I use a T300 wheel, no assists and AI level 105. I did multiple laps to make sure I got one in that was clean and near the limit of what I could get out of the car, to make the comparison with AI between sessions as accurate as possible.

First off, the dry qualifying lap to set the benchmark for AI level:

As expected in the dry on this level; within a tenth of my teammate. This was the first attempt, I probably could have found a bit more time if I tried more laps but it was good enough to make sure i had the right AI level for the wet weather test.

Now onto the actual comparisons in wet weather. If you don't want to watch everything, the results are below the videos.

Qualifying lap, light rain on inters:

Qualifying in heavy rain, full wets:

The race in full wet conditions (final part of it):

 

Results

Dry qualy lap: 1:17.982 (+0.073 to teammate/pole)

Light Rain lap: 1:29.516 (+ 1.275 gap to teammate/pole)

Heavy rain lap: 1:32.468 (+0.949 gap to teammate/pole) 

Heavy rain Race best lap: 1:33.996 (session fastest ) 

As you can see, in both wet sessions the AI were around a second quicker. Now of course, you might say "You just suck in wet conditions", but judging by how these laps felt, and more importantly the wet race pace I had on the exact same level, I don't think that's the case at all. Relative to my dry lap, I actually think these were just as good and it should not be a second off the pace for sure.

In the race, it was actually spot on and felt right for this AI level compared to dry. I could match the pace of the cars around me for the whole race, and got the fastest lap by only a tenth of a second.

@Faya I hope all this information is useful. I think there's definitely an issue with qualifying pace in the wet. Whether this is down to their actual pace or some other bug (simulated lap times?) Isn't clear, but I think this comparison shows there's something off. 

I'm really interested to hear more opinions about this as well and If you're experiencing the same thing in qualifying ( or the race of course!) 

Bonus test: I wanted to see what the AI would do on 110 and it turns out they're lapping half a second faster than the world record in Time Trial. The fastest lap ever done in TT would have earned me P6 in GP mode.

Thank you if you took the time to read or watch and made it all the way here. 

 

 

[kalamazoo123 289]

Good post. And i agree with your conclusion that there is nothing wrong with the AI wet pace generally.

I have not tested this as directly as you have, but i have been confused by all the posts stating that the AI are too good in the wet.

I have not found this to be the case myself in my career.  I have had good races with AI in the wet just as i have in the dry, without changing the difficulty level.

I run no assists and have always been pretty ok in the wet, so my own conclusion can only be that the people complaining about the AI wet weather pace must either a) just not be so good in the wet, or b) running assists and getting some sort of disadvantage from doing so (would make a nice change 😊).

[Worntoathread 865]

@kalamazoo123 Thanks, can I ask what difficulty you're on?

I agree the wet races have generally felt normal and well balanced to me but it's clear that something's wrong in qualifying, at least under certain circumstances. I'm starting to think simulated times might be the issue, when selecting go go track/return to garage etc. But I'll definitely do some more testing later.

[Balazsryche 32]

OK, so here is my situation. Career mode, Baku. Fastest car on the grid. Q1, intermediate conditions throughout qualy. I would say my timed lap was a 99,9%, so I don't think I could have been any faster. After that I fast forwarded and tadaaaaaam, I was last and +2.5 off the pace. Conditions were not improved. Never happened like this before the patch. In the dry I'm faster than them, so this is a huge difference.

I play with a controller and on 90% difficulty, TC full. I'm yet to test it with lowered TC settings, and it wouldn't be representative cause I'm not used to it, so..... Can you guys test both full TC and no TC settings in a wet qualy session and see how it goes compared to the AI? I think that would be telling. Thank you.

 

[kalamazoo123 289]

@Worntoathread I play on 97 difficulty. That’s worked best for me in terms of being roughly equal to my teammate. Don't think the difficulty itself really matters though. The point is i’m as competitive on the same difficulty whether it’s wet or dry.

@Balazsryche I never fast forward time in qualy. This has always caused issues in CM F1 games so best just not to do it. I would bet a lot of money that’s where your specific issue came from.

 

[Balazsryche 32]

21 minutes ago, kalamazoo123 said:

 I never fast forward time in qualy. This has always caused issues in CM F1 games so best just not to do it. I would bet a lot of money that’s where your specific issue came from.

 

I always do it and never caused any trouble, maybe some unreal sector times here and there. In dry conditions it's still fine, no issue like this...

[caiocrq 26]

My conclusion is this ... AI is too fast for everyone in qualifying, and only for those who use racing traction control. And the problems get worse on longer straight lanes, which is where AI is even faster.

[caiocrq 26]

3 hours ago, Balazsryche said:

OK, então aqui está a minha situação. Modo de carreira, Baku. Carro mais rápido na grade. Q1, condições intermediárias ao longo do qualy. Eu diria que minha volta cronometrada foi de 99,9%, então eu não acho que poderia ter sido mais rápido. Depois disso eu avancei rapidamente e tadaaaaaam, eu fui o último e +2.5 do ritmo. As condições não foram melhoradas. Nunca aconteceu assim antes do patch. No seco eu sou mais rápido que eles, então isso é uma diferença enorme.

Eu jogo com um controlador e em 90% de dificuldade, TC full. Eu ainda estou para testá-lo com configurações de TC abaixadas, e não seria representativo porque eu não estou acostumado com isso, então ..... Vocês podem testar tanto o TC completo e não as configurações de TC em uma sessão molhada qualy e ver como vai comparado com a IA? Eu acho que isso seria revelador. Obrigado.

 

And how are you doing in racing?

[Balazsryche 32]

2 hours ago, caiocrq said:

And how are you doing in racing?

Pretty well. Top 3 all the time so far.

[caiocrq 26]

2 hours ago, Balazsryche said:

Muito bem. Top 3 todo o tempo até agora.

In rain?

[joserdcosta 52]

Great original post. First off all thanks for all the testing. From my understanding it seems something is wrong with qualifying, specially when simulating/fast forwarding times. TBH I FF all the time in qualify. Next time I'll wait and see how it goes.

Cheers

[Worntoathread 865]

12 hours ago, joserdcosta said:

Great original post. First off all thanks for all the testing. From my understanding it seems something is wrong with qualifying, specially when simulating/fast forwarding times. TBH I FF all the time in qualify. Next time I'll wait and see how it goes.

Cheers

Thanks, I just did a bit of testing with one shot qualifying as well to see how the lap times compare, they turned out to be just as fast. I think this means the issue isn't necessarily with fast forwarding or simulated times, as in OSQ you can see them drive the lap in real time when you enable ghosting. So I guess it's their actual pace in wet qualifying compared to other sessions.

[SIMRACER123 355]

After all this @Striker_703 has now admitted to increasing his AI level by 5 AI levels ( and increase in approximately one second per lap ) as he was racing 5 levels beneath his ability.

This explains why he found beating the AI easy in the wet because he was racing on the incorrect level.

[Hygradeb 112]

@simracer123..  I noticed twice now you have mentioned that a increase of 5 is about a second.

Is there a 2019 chart that states level 75  AI give you a lap time say 1.24?

This would be useful for anyone trying to set a AI level.

 

Nice job worn with the test.