How to match GP4 pole time with real world pole time (desc=75)

This post discusses and clarifies the use of desc=75 (laptime finetune), in the Magic Data file.

For desc75, my copy of cmagic4.ini uses this hint:

"laptime finetune (+/-1000 equals +/- 1.0)"

To me, that always suggested two things:

1. If I want the CC cars to improve their laptimes by one second, I needed to use a value of -1000.

2. Desc75 is somehow subordinate to desc74 (laptime), because it's just 'fine tuning' for desc74, therefore I don't really need to bother with it anyway.

That's how it is, right?

Well no, it isn't like that at all - and if you, like me, ASSUMED that was what desc75 was all about, prepare yourself for a nice surprise.

As it turns out, desc75 is very powerful, but it's mis-labeled AND the hint is also misleading - almost like it was hidden in plain sight deliberately!

Anyway, according to some testing I did today:

To make your car and the CC cars (roughly) 1 second quicker around the lap, you would use the value 1000, not -1000 as cmagic seems to imply! Give it a try - it works a treat. Similarly a value of -1000 will make you and the CC cars 1 second SLOWER.

As for desc75's (inferior) relationship to desc74, I'm not sure there is one really.

I remember a post a year or so ago where SDI explained how desc74 was used to calibrate real-time with GP4 time, i.e. if you actually timed a lap on a stopwatch you'd find that the recorded time wasn't always the same as the time GP4 recorded - and desc74 had something to do with this time-warping process. That's probably a terrible interpretation of what the invariably brilliant Rene said, but for the purposes of this post it doesn't matter that I was too stupid to understand what he meant!

However, in the very same post, the MD magician TomMK went on to point out how in the vanilla game, desc74 actually reflects the pole position laptime for each race in 2001.

Ahhh! So, from this we can deduce that desc74 is clearly very much to do with the pole position laptime: but what about desc75? What does that have to do with it?

Well, according to my (as usual) rather brief and possibly inadequate testing, desc75 is quite probably THE holy grail of GP4. Put simply, desc75 is the easiest way to make that desc74 pole time happen on the track.

Quite a breakthrough, isn't it? 

Have fun...

If you didn't already know about all this, here are some answers to questions that you might have.

1. I did what you said and it worked out great the first time, but then next time I raced the track the pole time was more than 1 sec too fast (or too slow). What's up?

Variance. Variance in the PF, variance in weather conditions, variance because of slipstreaming, variance because of traffic, variance because you are human and variance because of (probable) random values in GP4 that we don't have access to.

Be thankful. It's the variance that keeps GP4 interesting after all these years.

If you really wanted to pin the value down as far as possible, you'd need to repeat the qualifying process many many times, maybe 100 or so, then you'd eventually be able to identify the precise value that works best. Good luck with that. I'm too old already.

2. I did what you said and the CC times are great, but now MY laptimes are way too fast (or too slow) all the time.

Yes. It happens. It just happened to me in fact, while re-calibrating Magny-Cours 1991. The bad news is, if your PF is from a good source (ie accurately calculated and based on the real world data), you'll need to resolve the problem elsewhere in the MD. If you're too fast try increasing the CC grip values in the MD a bit. Too slow and you'll need to reduce them. Remember to back out to at least the game menu to enact the new values.

Here's a tip: if you're in a decent car and the track already drives like you think it should, ie not to grippy, not too slippery, DON'T adjust the track grip value (I know it's tempting because it's what I used to do at one time; but don't) - focus only on the CC grip values. Remember though, that any adjustments to the CC grip values WILL change the performance of the CC cars (and your car too, to a certain extent), so you might need to go back and re-calibrate desc75 again after you make the changes. It gets a bit circular, as forays into the MD often are, but again, don't lose sleep over this in your search for perfection. Because of variance, near enough is the best you can hope for in any case.

Another tip would be to NOT mess with the CC power settings to fix this problem. If your car has a similar top speed to the CC cars already you can only make things worse. One day I'll write a post on CC grip and CC power values, which will explain all these nuances better, but it would be too much to take on here.

EDIT: A quicker, but I would say more temporary and 'less proper' work around for this particular problem (if you don't feel like walking into the MD minefield), would be to use the Bellini patch.

3. So what cmagic hints would you use for desc74 and desc75 then?

desc74=real-world pole laptime (e.g. 70800 equals 1:10.800)
desc75=GP4 pole laptime modifier (500 equals 0.5s faster, -3000 equals 3s slower)

You can edit your cmagic4.ini file if you want to update the hints. You don't need to do that, but if you do, any new MDs you create will then use the new hints. To update an old MD, simply use cmagic's EDIT - CREATE MAGIC FILE FROM GP4.EXE option to make a new one and the new hints will be inserted automatically.

Hope that's all been helpful to someone.

Cheers.



Edited 1 time(s). Last edit at 06/06/2020 01:01AM by Noog.

Thanks Noog, i will test soon. Its a big improvement, specialy the 1s for pp times!
Great Job!

Interesting! As usual, more good MD research

What happens if you take it to the extreme and put a crazy big (or small) value?

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


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

Thanks Lando.

Thanks Tom.

What happens with extreme values?

I can't actually say I've tried. However, I was driving a new track (2010) in an old mod (1993) this morning and I calculated that the value in desc74 would have been about 195000 if the race had ever been held.

In testing, the pole time was about 17s too quick, so I put -17000 in desc 75 (which I guess is fairly extreme), but pole was still 4 seconds too quick IIRC. At the point I decided to just get on with the race, but essentially it seems it's not a simple linear relationship.

On the other hand, I raced the Brands Hatch beta the other day (a fine circuit IMO, whoever it was that put it together) and while I don't remember the figures, I put in whatever it seemed to need (about 3000 IIRC) and it was spot on, to within 30ms! After that I went on to do about a dozen more tracks all with great success, but today was the first time I've ever tried to rectify such a huge disparity - and even then I suppose it was fairly encouraging, i.e. nothing broke!

PS: At some point we need to talk about the 'first corner issue' again Tom. There has been some progress made recently, but it isn't necessarily good news.



Edited 1 time(s). Last edit at 06/10/2020 02:46PM by Noog.

Just thought I'd add another little update.

After I answered Tom's question just now I realised I still had the 2010 circuit and the MD open on screen, so I checked the MD again & noticed I'd made a typo. I corrected the error (desc75 is now -14700) and on accelerated time a few moments ago Prost took pole in 1:59.32, as compared to the calculated RW time of 1:59.11. So it looks like it IS in fact linear after all.