Magic Data Discussion

TomMK Schreef:
-------------------------------------------------------
> But.... but.... why?!?!

This has been in there since F1GP already. My guess is that it is to finetune the lap times to be the same as the ones driven in real F1.

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

Not sure if this was known before but I just figured out that desc74 in the original 17 tracks' magic data is the real-life 2001 qualifying pole lap times (rounded to the nearest hundredth).

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


Intel NUC 8i3, 8GB RAM, MS Sidewinder Wheel

Doesn't surprise me In F1GP it was the fastest race lap rounded to a whole second for most tracks of the real-life 1991 season.

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

Can we have a deeper look at the Player Tyre Wear value (desc50)? Atticus' observation above that it is a factor that helps to vary the underlying Tyre Wear (desc72) is very interesting. Is there a value where you typically get about equal Player and AI tyre wear (e.g. a fair value?).

Rene can you see what GP4 is doing with those 2 values?

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


Intel NUC 8i3, 8GB RAM, MS Sidewinder Wheel

As far as I can see desc50 is a factor that simply multiplies desc72, where 16384 equals 1.0. I see no difference between CC or player here, meaning they should both work for CC and player (just by looking at the code).

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

Interesting - so there is actually no distinction between Player and AI tyre wear? Time to do some testing....

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


Intel NUC 8i3, 8GB RAM, MS Sidewinder Wheel

Currently, I can't confirm that desc50 (known as 'tyre wear player' and desc72 (known as 'tyre wear ai' or 'ccshuffler' work for both, players and ccCars. I did not test it once again after SDI's last post, but so far, my results always showed, that both values are doing exactly what they are known as. But nevertheless I am willing to get convinced by other one's tests.

Ok, I've debugged GP4 to check when it reads the values for which cars, and I saw that desc50 is sometimes used for both player and CC, but most of the time only for player. But desc72 was always used for both player and CC, but from different parts of the code. The player's car has individual tyre wear for each of the 4 tyres, but it seems the CC cars only have tyre wear for all wheels at once.

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

Very interesting, thanks René.

I think I'm going to try setting desc50 to 16384 at all times, and just varying desc72, and see what effect that has.

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


Intel NUC 8i3, 8GB RAM, MS Sidewinder Wheel

Hey, guys, which parameter determinies the number of sector of the track on which CC guys make their driving errors? When I play the game I notice that most of the drivers going off the track mostly on the same exact places while riding alone, such as the last turn in original Hockenheim. Maybe that can be one of the undefined numbers? Just a guess.

___________________________________________________________________________
For a list of EVERY download for GP4, look here:[docs.google.com]

Also a question - making a pit-window set like this:
100 ; 1 Stops Percentage (%)
30 ; Stop1 (sector)
7 ; Pit Window 1 (range)

does it mean that the drivers will go in at laps 30-37, or it will be laps 27-33 (around 30)?

___________________________________________________________________________
For a list of EVERY download for GP4, look here:[docs.google.com]



Edited 3 time(s). Last edit at 08/26/2019 05:51PM by Soutsen.

Soutsen Schreef:
-------------------------------------------------------
> Hey, guys, which parameter determinies the number
> of sector of the track on which CC guys make their
> driving errors? When I play the game I notice that
> most of the drivers going off the track mostly on
> the same exact places while riding alone, such as
> the last turn in original Hockenheim. Maybe that
> can be one of the undefined numbers? Just a guess.


Not discovered yet. What also is interesting to find out is which parameter in md is resulting that cars getting stuck in the kitty litter of T1 @ Suzuka.



>Also a question - making a pit-window set like this:
>100 ; 1 Stops Percentage (%)
>30 ; Stop1 (sector)
>7 ; Pit Window 1 (range)
>
>does it mean that the drivers will go in at laps 30-37, or it will be laps 27-33 (around 30)?


30-37.

---

My Grand Prix 4 Files
I'm a total dick. How many people can say that?

Thanks for your reply

Also one interesting thing - I'm playing with your old Zandvoort track with the original Hungaroring md (mostly) inserted into that. What I've noticed is the small ammount of cars crashing despite driving errors. The only place where they're crashing seems to me the first corner (Tarzan). This is strange to me cos this is a fast track and in theory cars must go out harder and wider into the grass at the bends. But somehow despite the errors (I've set it to 200 to test it) they are mostly staying on the tarmac somehow like thery're glued. Is the ability to stay on track is connected to the track grip factor or something like late breaking ability maybe? Because on the new Zandvoort by MarioB and b-tony the cars seem to struggle to keep on the road during the driving errors and going on the grass almost immideately (30 error chance leads to lots of travels into the gravel). Or maybe it is a CC-line issue?

___________________________________________________________________________
For a list of EVERY download for GP4, look here:[docs.google.com]



Edited 1 time(s). Last edit at 08/26/2019 11:00PM by Soutsen.

Thie question I raised here seems to be more appropriately placed here:
What does the "AI Random Perf Range" (desc63,64)? What is it for and how exactly does it work?

___________________________________________________________________________
For a list of EVERY download for GP4, look here:[docs.google.com]



Edited 2 time(s). Last edit at 09/02/2019 07:56PM by Soutsen.

Sorry, Atticus solved it first, go back to page 10!

I've been testing and creating 1997 physics files, trying to improve the car handling, progressive, saveable tyre slip etc. I'll release what I have done soon, I'm delighted with the results.
The only problem with it is that whilst I have to wring the neck of my Arrows to get it off the back row of the grid on Ace (using Pedro Diniz grip of 15800 ish) the AI were trouncing me in the wet, I was 3 seconds off the slowest AI car, driving dozens of laps on the edge. If I could only slow them down..

My first thought was to give them bad setups, but that only resulted in a laptime variation of roughly 1 second around Interlagos, plus if the track were to become dry, it wouldn't be a good solution anyway.

So I took a look at the magic data....

.............

11053 ; AI Tyre Change due to weather
18000 ; AI Tyre changes Finetune ?
12 ; Rain Chance (%)
6 ; Start Segment of some range
6 ; End segment of some range
256 ; AI Race Grip ?
13000 ; Unknown
233 ; Unknown
10000 ; Black Flag Penalty (ms)
2048 ; Black Flag Speed Reduction ?
190 ; Unknown
120 ; Unknown
80 ; Unknown
160 ; Unknown
0 ; Handbrake ?
1814 ; Player Garage Y Pos
896 ; AI Garage Y Position
0 ; Garage Angle (8192=45°)
0 ; Pit Y Position
0 ; Pit Y Position Finetune
16384 ; Unknown
16384 ; Unknown
16384 ; Unknown
16384 ; Unknown
16384 ; Unknown
16384 ; Unknown
100 ; 1 Stop Percentage (%)
15 ; Stop1 (sector)
11 ; Pit Window 1 (range)
0 ; Unused
0 ; Unused
0 ; Unused
0 ; Unused
0 ; Unused
0 ; 2 Stops Percentage (%)


THIS VALUE, 89 unknown in GP4 builder, controls the AI cornering speed or grip, in wet weather, Independently of dry weather!
I tested it at Interlagos, exiting the weekends and alt tabbing in order to reload the same weather with a different .dat

Changing it yielded the following results

Default value 80 Fastest AI 1.28 (vs a pole time in dry of 15.8, too quick for a fully wet track, hence my troubles..)
Changing it to 500, Fastest AI 1.49.9 (straight line speed unaffected, cornering much slower)
Changing it to 1000, AI take each corner at literally 5mph, laptimes would be several minutes.
Changing it to 160, Fastest AI 1.31.0

For all these values the dry times stayed the same, ie it has zero effect.

It appears that increasing the value has an initially linear effect on laptimes, a change of 40 slowed them 1.5 seconds, raising it 80 slowed them by my desired 3 seconds.

Will test further, but for now it looks like a good solution! I've avoided wet weather for a decade because of the performance disparity, so I'm delighted

===================================================================================
Tahitian GP Circuit
[www.grandprixgames.org]
Easter Island Circuit
[www.grandprixgames.org]



Edited 1 time(s). Last edit at 09/09/2019 03:23PM by salvasirignano.

You sir are a legend!

Wrote:
-------------------------------------------------------
> Wet weather solved!
>
> I've been testing and creating 1997 physics
> files, trying to improve the car handling,
> progressive, saveable tyre slip etc. I'll release
> what I have done soon, I'm delighted with the
> results.
> The only problem with it is that whilst I have to
> wring the neck of my Arrows to get it off the back
> row of the grid on Ace (using Pedro Diniz grip of
> 15800 ish) the AI were trouncing me in the wet, I
> was 3 seconds off the slowest AI car, driving
> dozens of laps on the edge. If I could only slow
> them down..
>
> My first thought was to give them bad setups, but
> that only resulted in a laptime variation of
> roughly 1 second around Interlagos, plus if the
> track were to become dry, it wouldn't be a good
> solution anyway.
>
> So I took a look at the magic data....
>
> .............
>
> 11053 ; AI Tyre Change due to weather
> 18000 ; AI Tyre changes Finetune ?
> 12 ; Rain Chance (%)
> 6 ; Start Segment of some range
> 6 ; End segment of some range
> 256 ; AI Race Grip ?
> 13000 ; Unknown
> 233 ; Unknown
> 10000 ; Black Flag Penalty (ms)
> 2048 ; Black Flag Speed Reduction ?
> 190 ; Unknown
> 120 ; Unknown
> 80 ; Unknown
> 160 ; Unknown
> 0 ; Handbrake ?
> 1814 ; Player Garage Y Pos
> 896 ; AI Garage Y Position
> 0 ; Garage Angle (8192=45°)
> 0 ; Pit Y Position
> 0 ; Pit Y Position Finetune
> 16384 ; Unknown
> 16384 ; Unknown
> 16384 ; Unknown
> 16384 ; Unknown
> 16384 ; Unknown
> 16384 ; Unknown
> 100 ; 1 Stop Percentage (%)
> 15 ; Stop1 (sector)
> 11 ; Pit Window 1 (range)
> 0 ; Unused
> 0 ; Unused
> 0 ; Unused
> 0 ; Unused
> 0 ; Unused
> 0 ; 2 Stops Percentage (%)
>
>
> THIS VALUE, 89 unknown in GP4 builder,
> controls the AI cornering speed or grip, in wet
> weather, Independently of dry weather!
> I tested it at Interlagos, exiting the weekends
> and alt tabbing in order to reload the same
> weather with a different .dat
>
> Changing it yielded the following results
>
> Default value 80 Fastest AI 1.28 (vs a pole time
> in dry of 15.8, too quick for a fully wet track,
> hence my troubles..)
> Changing it to 500, Fastest AI 1.49.9 (straight
> line speed unaffected, cornering much slower)
> Changing it to 1000, AI take each corner at
> literally 5mph, laptimes would be several
> minutes.
> Changing it to 160, Fastest AI 1.31.0
>
> For all these values the dry times stayed the
> same, ie it has zero effect.
>
> It appears that increasing the value has an
> initially linear effect on laptimes, a change of
> 40 slowed them 1.5 seconds, raising it 80 slowed
> them by my desired 3 seconds.
>
> Will test further, but for now it looks like a
> good solution! I've avoided wet weather for a
> decade because of the performance disparity, so
> I'm delighted

Thanks, I'll keep experimenting, because I don't understand what the parameters are, just the effect they seem to have. I don't know what effect 86-88 have either...

===================================================================================
Tahitian GP Circuit
[www.grandprixgames.org]
Easter Island Circuit
[www.grandprixgames.org]

Thanks, a good find indeed! With this we could emulate the "rain-masters" such as Senna, Schumacher or Verstappen at Brazil some seasons ago if SDI can make this value to be separatable for each driver.

___________________________________________________________________________
For a list of EVERY download for GP4, look here:[docs.google.com]



Edited 1 time(s). Last edit at 09/09/2019 07:14PM by Soutsen.

Hi

Thanks for the tutorial. It really sheds some light on a dark art !!

I am new to track editing and want to use TSM to make changes. Specifically I want to improve AI laptimes at Barcelona.

I have installed a fresh version of the 1997 version of the track to play around with. However, I need a little help on the modifications I need to make to the magic data. I want an increase of about a second a lap for the AI to make them more competitive. Can you give me any advice on the factors that need changing and by how much?

Thanks

track grip mainly desc42. play around with it, original values are 15742-17120, but you can go below/above these. The higher value gives more grip producing better laptimes

___________________________________________________________________________
For a list of EVERY download for GP4, look here:[docs.google.com]