Misc URD-RSS GT3 BoP v2.0

[CrestedIsland]

I'm sorry if I'm being a party pooper, but these times are too far off of what the cars can really do on Nords. There's no point to compare times that aren't even remotely close to the limit.

The original approach to BoP them using AI on default setups is also inherently wrong. First of all, AI can't really drive GT3 cars fast enough (I'm not talking about aliens, but even fast guys will run circles around 100% AI in AC). Then, the default setup for some cars is no good and you can gain a lot by adjusting it. It's also worth noticing that the weather conditions are very important. Mid-engine cars from RSS are known to overheat their rear tires and wear them out very quickly when you get close to +20C or above (on both CSP and vanilla versions, and if you're not getting that you're simply not driving fast enough). This doesn't happen to URD cars and doesn't happen to front-engine RSS cars.

I wouldn't be surprised if there was a better process out there, and I know how bad the AI can be, but usually the precedent set by them translates at the very worst on an OK rate to actual drivers. I know what you mean by the AI are slow, and I agree on most tracks, but for Sebring, the line is actually pretty good. Like I said, and as @doc3d added, the inconsistencies of a human are absolutely huge, and there's only so much one can do. Different setups, different skill levels and different driving strategies, different preferences through engine placement, different consistencies (and the fact that a human will never be as consistent as an AI), and on top of that inconsistencies as the environment (as you pointed out) and the fact that there are 2 different mod studios there with different ideas of power curves, braking power, tire wear, fuel usage, etc. are probably just half the reasons as to why I do it this way. So, I apologize if it's not up to your standard - I truly do - but there's only so many things that I am willing to and can do for this stuff

 

[CrestedIsland]

Great work @doc3d ! I've been using this BOP preset for quite a while now and from a racing point of view (non-Nordschleife races), it's pretty well done, there is no obvious faster car from my experience, so I'm really enjoying this preset, thank you again @CrestedIsland

Awesome to hear man, I'm glad everything is working out for you! Yeah, I think the longer the track is, the more inconsistent it gets, and mayyyyyybe I might try to buff them out, but for the majority of tracks should be great

 

[demetri]

I wouldn't be surprised if there was a better process out there, and I know how bad the AI can be, but usually the precedent set by them translates at the very worst on an OK rate to actual drivers. I know what you mean by the AI are slow, and I agree on most tracks, but for Sebring, the line is actually pretty good. Like I said, and as @doc3d added, the inconsistencies of a human are absolutely huge, and there's only so much one can do. Different setups, different skill levels and different driving strategies, different preferences through engine placement, different consistencies (and the fact that a human will never be as consistent as an AI), and on top of that inconsistencies as the environment (as you pointed out) and the fact that there are 2 different mod studios there with different ideas of power curves, braking power, tire wear, fuel usage, etc. are probably just half the reasons as to why I do it this way. So, I apologize if it's not up to your standard - I truly do - but there's only so many things that I am willing to and can do for this stuff

A better process exists, but it is not easy. You need many skilled drivers to put in laps on some server in the same fixed conditions and compare the laptimes and top speeds. What you did is maybe suitable for offline racing against AI as you're making them all more or less equal. But using it for any online competition or a league is a bad idea. AI can't drive good enough even if you give them a good line, and custom setups make a significant difference. Some cars come with decent default setups, but some are just bad, and if you balance the cars based on the default, the car with a bad default setup becomes OP once you fit it with a good one.

And I don't buy into that "the inconsistencies of a human are absolutely huge" thing. If one is not consistent enough they probably shouldn't be bothered with the BoP in the first place. If you can't do a few laps back to back within 0.5s of each other (for normal length tracks, make it a couple of seconds for something like Nords) you can't even tell if the BoP really works, or it's your inconsistency making you imagining things.

 

[doc3d]

I respectfully disagree, @demetri
The cars are different, so it’s not unnatural that the laptimes of them are differing. Being consistent in ONE car over a number of laps I agree with (which is why I did a +3 lap average per car) but given that the cars are unbalanced (the whole reason for doing a BOP, no…?) and from different studios then it’s, in my view, not uncommon to have differing laptimes. And at a track like Nords, where one lap is + 8 minutes in these cars, small differences in the mod can create a wide range of results.
As far as setups go, how do you determine what a good setup is…? What works for one driver could be completely wrong for another so you have to have some sort of baseline to test consistently with. Using the default was the best I could think of (only changing the fuel load to be the same across all), as I assume one mod-group would use similar strategies when creating a group of mods.

But, by all means; if you want to do custom setups on all cars -that would take, what.. 5-6 fast laps at least? Then run a number of laps to check consistency - let’s say at least 3-4 laps. Then, you apply the BOP and do the same. Since your custom setup might not work as well with a heavier car, you might have to adjust it a bit, so perhaps 2 more laps, then the consistency-run, another 3-4. I’d say you’re now up to 15-16 laps (not counting out-laps as you don’t get a time for those + temps so they are always slow). So that’s around 15x8=120 minutes per car. For 12 cars, that is 24 hours of driving, all consistent, error free laps. Please let me know your results

As far as I understood (at least to me it’s a fact) the BOP was meant against the AI, not for online play. For me, slow as I am, the AI is fine, but I still want them to have BOP applied to avoid wildly different results. I take it from your reply that you would destroy the AI, so I would assume that this BOP is not for you and you should go online or leagues to get a challenge. To ME tho, having AI that runs about the same laptimes as me is of value, and the only way I can see that happening is with a BOP, lacking as it may be in your opinion.

 

[CrestedIsland]

And I don't buy into that "the inconsistencies of a human are absolutely huge" thing. If one is not consistent enough they probably shouldn't be bothered with the BoP in the first place. If you can't do a few laps back to back within 0.5s of each other (for normal length tracks, make it a couple of seconds for something like Nords) you can't even tell if the BoP really works, or it's your inconsistency making you imagining things.

I can be 100% consistent enough; I’ve done many online professional endurance races (WSS, OC Sim Spots, THR, etc,) with not bad results. I’m talking about the inconsistencies that everyone else downloading this and testing it for themselves will see, because once again, as I had said previously, inconsistencies of a human are prominent.

The process you’re describing is quite literally hell. I have to get a baseline lap for the cars, which could easily take over 5 laps of whatever track it may be, along with the other skilled people putting in more laps for every single car. After that, I apply the ballast and restrictor changes I believe are best, and have everyone set out 5+ laps on the same track for every single car. Repeat that process, oh i don't know, about 10 times to get it in a great spot. And even then will the tire wear/fuel usage/fuel tanks not match up, because it’s out of our control to change that. Yeah, that is a hassle that I nor absolutely nobody would want to go through my guy, just for something like this that doesn’t even have a high chance to be in a league in a first place. 100% parity for online racing is impossible. Though, if you are delusional enough and want to try, feel free to make a mod that doesn’t have any demand in comparison to this one .

 

[demetri]

I respectfully disagree, @demetri
The cars are different, so it’s not unnatural that the laptimes of them are differing. Being consistent in ONE car over a number of laps I agree with (which is why I did a +3 lap average per car) but given that the cars are unbalanced (the whole reason for doing a BOP, no…?) and from different studios then it’s, in my view, not uncommon to have differing laptimes. And at a track like Nords, where one lap is + 8 minutes in these cars, small differences in the mod can create a wide range of results.
As far as setups go, how do you determine what a good setup is…? What works for one driver could be completely wrong for another so you have to have some sort of baseline to test consistently with. Using the default was the best I could think of (only changing the fuel load to be the same across all), as I assume one mod-group would use similar strategies when creating a group of mods.

But, by all means; if you want to do custom setups on all cars -that would take, what.. 5-6 fast laps at least? Then run a number of laps to check consistency - let’s say at least 3-4 laps. Then, you apply the BOP and do the same. Since your custom setup might not work as well with a heavier car, you might have to adjust it a bit, so perhaps 2 more laps, then the consistency-run, another 3-4. I’d say you’re now up to 15-16 laps (not counting out-laps as you don’t get a time for those + temps so they are always slow). So that’s around 15x8=120 minutes per car. For 12 cars, that is 24 hours of driving, all consistent, error free laps. Please let me know your results

As far as I understood (at least to me it’s a fact) the BOP was meant against the AI, not for online play. For me, slow as I am, the AI is fine, but I still want them to have BOP applied to avoid wildly different results. I take it from your reply that you would destroy the AI, so I would assume that this BOP is not for you and you should go online or leagues to get a challenge. To ME tho, having AI that runs about the same laptimes as me is of value, and the only way I can see that happening is with a BOP, lacking as it may be in your opinion.

If you're using it for AI only, it's probably fine. Moreover, AI cars always use the default setup and I'm not sure if there's a reliable way of forcing them into a custom setup, and I also found that they often simply can't drive with a fast custom setup (you can do that loading your setup and pressing Ctrl-C to activate AI driver offline) ending up wrecking the car, especially on a tricky track like Nords. So, applying this BoP will make AI to lap similar times in different cars of the same class which is good. Though AI in AC is so bad (not pace-wise, but in general), I didn't think anyone was still racing against it as their primary way of simracing.

But the original post mentioned using them for online leagues and this is where this whole idea falls apart. Don't do that, it's a bad idea. This BoP likely makes the cars less balanced than they originally were.

P.S. A good setup takes way more than 5-6 fast laps to come up with, at least for me.