Magic Data Discussion

Posted by TomMK 
Re: Magic Data Discussion
Date: January 20, 2016 11:52PM
Posted by: AUS_Doug
Re: Magic Data Discussion
Date: January 26, 2016 01:48AM
Posted by: AUS_Doug
OK, just checked it out:

For the AI, the 'Old Behaviour' in the new version does match the behaviour of the old versions.
Qualifying sessions at Interlagos, the fastest AI in each test was within one-tenth of the other's time.
(i.e. 'Old Behaviour' was 1:14.500, old version was 1:14.600)

At this stage though, I'm not sure if it quite matches up for the human; I was about three-tenths slower with 'Old Behaviour' selected compared to the old version.
This data is only from about six laps though, so I'll go do some more laps in free practice mode and get a larger sample.

I also noted where and how much time I (driving the Minardi) lost to the AI top qualifier.

In the old version - where I qualified ninth - I was up by 0.1 at the end of sector 1, down by .5 at the end of sector 2 and down by .8 at the end of the lap.
With the new version and 'Old Behaviour' - where I qualified 13th - I was up by 0.1 at the end of sector 1, down by .4 at the end of sector 2 and down by 1.3 at the end of the lap.
The difference could be human error - which is why I'll do more tests - but I don't think that I'm that bad that I can have half a second difference between two flying laps on low fuel and fresh tyres.
(Particularly around Interlagos)

My top speed and the fastest AI speed didn't change between tests - mine was still about 7km/h slower than the fastest AI - but I still think there is an acceleration difference.
Exiting turn 12 is where it is most noticeable visually.
In both tests, I had a car in front of me heading into turn 11, but far enough away that I wasn't (noticeably) effected negatively by the slipstream.
In the 'Old Beahviour' test, it was a Prost, and in the old version it was a Williams.

In the old version, I had no trouble sticking with the Williams all the way to turn 1; wasn't close enough to pass, but the tow kept the gap constant.
In the 'Old Behaviour' test, the Prost was gone (I'd lost the little tow that I'd had) by turn 14.

I'll report back after some more runs.
(I might also start up a new thread when I do, just in case other people are noticing something similar.)



F1 Graphics Swapper --- Lensdump - Renders and Screenshots --- GP4 Performance Generator
Re: Magic Data Discussion
Date: January 26, 2016 09:41AM
Posted by: TomMK
I also notice that I'm slower than the AI (i.e. my teammate) in a straight line now.

@Rene - is it possible for the Player to be not affected by the Magic Data AI values for grip and power?

=====================================================


Intel Core i5-6200U, 8GB, Intel HD Graphics 520, MS Sidewinder Wheel
Re: Magic Data Discussion
Date: January 26, 2016 09:47AM
Posted by: RHill
TomMK Wrote:
-------------------------------------------------------
> I also notice that I'm slower than the AI (i.e. my
> teammate) in a straight line now.
>
> @Rene - is it possible for the Player to be not
> affected by the Magic Data AI values for grip and
> power?


I think that's the key here, when power and grip are selected, ideally the power and grip values in the magic data would not effect the player (only the ai) so you can then adjust the ai performance idependently to match the player
Re: Magic Data Discussion
Date: January 26, 2016 08:54PM
Posted by: SDI
Thanks for the test and the report. I'll have to check and compare the code again to see if something is different between the old version and the new version with 'old behaviour'.

As for the magic data, I'll see if that's possible.

René Smit, Independent Software Developer.
Download my GPx tools here.
Re: Magic Data Discussion
Date: February 06, 2016 09:13PM
Posted by: SDI
I've compared the code, and the only difference is that with the 'old behaviour' in the new version, it also patches the grip in hotseat mode. In the old version is only patched the power in hotseat mode, which I considered a bug. I assume you're not testing in hotseat mode?

Another tiny change is that I don't patch the exe if you don't set it to team dependent. So if another tool has patched the exe it is not overwritten. So make sure you use a clean exe.

René Smit, Independent Software Developer.
Download my GPx tools here.



Edited 1 time(s). Last edit at 02/06/2016 09:34PM by SDI.
Re: Magic Data Discussion
Date: February 06, 2016 09:50PM
Posted by: SDI
It's possible to ignore the magic data for team dependent player grip, but for power it is much harder.

René Smit, Independent Software Developer.
Download my GPx tools here.
Re: Magic Data Discussion
Date: February 07, 2016 08:12AM
Posted by: TomMK
Certainly I think the grip side is more important, so that would be good to have. At the moment you can't make the game easier or harder at all, because you can't change the AI level without influencing the player too.

Can you just give the player a value of, say, 500 for grip and power rather than read the magic data value for the player?

=====================================================


Intel Core i5-6200U, 8GB, Intel HD Graphics 520, MS Sidewinder Wheel
Re: Magic Data Discussion
Date: February 07, 2016 12:40PM
Posted by: SDI
Actually, there are already 3 ways to tweak AI level and player level independently:

1. By not using the player car performance team dependent option.
2. By using player master grip.
3. By using individual performance values for each car.

But I agree that applying the opposition level (by which the magic data is indexed) on the player is paradoxical.
The magic data values are originally meant to tweak the AI performance for each track. Indeed, this should probably not apply to players.
If, while player car performance team dependent is active, I decide to skip the magic data grip factor for the player's car, then it would boil down to having a magic data grip factor of 1.0. Applying a value of 500 (this is a magic data grip factor of 1.0) is already the case when not using team dependent grip.

René Smit, Independent Software Developer.
Download my GPx tools here.



Edited 1 time(s). Last edit at 02/14/2016 01:21PM by SDI.
Re: Magic Data Discussion
Date: February 13, 2016 06:28AM
Posted by: AUS_Doug
I completely forgot about this whole thing, sorry.
So, long story short, I did some more running at Interlagos and you're right Rene, the 'Old Behaviour' does match the old version in regards to how the human cars behave.
I wonder then if the acceleration issue I mentioned is related to the Magic Data Tom mentioned.



F1 Graphics Swapper --- Lensdump - Renders and Screenshots --- GP4 Performance Generator



Edited 1 time(s). Last edit at 02/13/2016 06:33AM by AUS_Doug.
Re: Magic Data Discussion
Date: February 13, 2016 10:35PM
Posted by: SDI
Ok, so there's no "old behaviour" bug then, nice :)

As for the acceleration problem, the "new behaviour" just adds team dependent grip, the magic data values for power have always been applied, even in the old versions.

I can make a test version for you to see if disabling the magic data grip factor for players helps?

René Smit, Independent Software Developer.
Download my GPx tools here.
Re: Magic Data Discussion
Date: February 13, 2016 10:44PM
Posted by: TomMK
Yeh that would be great. I don't think either the CC grip / power values should be applied to the player - they are there to adjust the performance of the AI relative to the player.

=====================================================


Intel Core i5-6200U, 8GB, Intel HD Graphics 520, MS Sidewinder Wheel
Re: Magic Data Discussion
Date: February 14, 2016 04:30PM
Posted by: SDI
I've managed to exclude the magic data power factor for player's car too, by simply dividing by it to make it undone.

René Smit, Independent Software Developer.
Download my GPx tools here.
Re: Magic Data Discussion
Date: July 22, 2016 02:38PM
Posted by: TheFueleffect
Is it possible to have driver-dependent tire wear (cc shuffler) values in the performance files? That would be quite cool.
Re: Magic Data Discussion
Date: July 23, 2016 02:57PM
Posted by: TomMK
But unfair, no?

=====================================================


Intel Core i5-6200U, 8GB, Intel HD Graphics 520, MS Sidewinder Wheel
Re: Magic Data Discussion
Date: July 23, 2016 06:09PM
Posted by: Atticus.
TomMK Wrote:
-------------------------------------------------------
> But unfair, no?


I'm with TheFueleffect on that one - we can see in real life as well that some drivers have consistently better tyre wear than others. Perez, for example. Kvyat often had worse than Ricciardo as well when they were teammates.

My concern is that it makes things a bit difficult, isn't it? We already have raw performance values, variation values, 'power-ups' per difficulty via magic data and whatnot. I think GP4 'view only' races (to quote a GP3 term) are the best simulated races - still - out of all the similar features of F1 and motorsport games already.

But all in all, it's a cool idea, yes.



My workthread - [www.grandprixgames.org]
Full of classic F1/non-F1 track layouts

My blog about F1 performance analysis - [thef1formbook.wordpress.com]




Edited 1 time(s). Last edit at 07/23/2016 06:09PM by Atticus..
Re: Magic Data Discussion
Date: July 23, 2016 06:42PM
Posted by: TGF1DRIVER
There's only one man who can make this dream come true. If you know what I mean. ;)(Y)
Re: Magic Data Discussion
Date: July 28, 2016 10:31PM
Posted by: TheFueleffect
Tire wear may become another performance parameter, so it shouldn't be unfair. It of course needs some finetuning, but it can be used to create some (further) difference between a driver's qualifying and race performance. It also can give cars a long-term performance advantage in the race, which may be quite cool: tire-friendly drivers may get enough of a performance advantage to finally overtake (not that overtaking is particularly difficult in GP4, but still) and an otherwise runaway leader may be overhauled later in the race (most GP4 races are boring exactly for that reason). The main problem is that tire wear is track dependent in GP4 I later realized, but when simulating the current F1 era constant values for CC shuffler may be quite accurate, as tire wear mainly depends on the number of high-speed corners. Anyway, small differences (10% or so) in tire wear between drivers will only make a subtle difference in the race I think.
Please? :)
Re: Magic Data Discussion
Date: July 29, 2016 03:22AM
Posted by: AUS_Doug
The only way to find out if it's possible is to ask René if it's possible with GPxPatch.
I suggest opening a new issue on the BitBucket page.



F1 Graphics Swapper --- Lensdump - Renders and Screenshots --- GP4 Performance Generator
Re: Magic Data Discussion
Date: January 25, 2018 08:46PM
Posted by: Atticus.
Bump.

Guys, you won't quite believe what I have found out about a few values here... It's quite outrageous.

So far, I have news on "tyre wear player", "downforce?" and "ride height" with - hopefully - many more to come.

Stay tuned.



My workthread - [www.grandprixgames.org]
Full of classic F1/non-F1 track layouts

My blog about F1 performance analysis - [thef1formbook.wordpress.com]
Sorry, only registered users may post in this forum.

Click here to login

Maintainer: mortal, stephan | Design: stephan, Lo2k | Moderatoren: mortal, Vader, stephan | Downloads: Lo2k | Supported by: Atlassian Experts Berlin | Forum Rules | Policy