Non-championship races vs. championship difference

Here you can post questions about GP2, your favourite GP2 things and errorreports/problems!
Laahustaja
GP2 Newbie
Posts: 9
Joined: 30.09.2019, 11:33

Non-championship races vs. championship difference

Post by Laahustaja » 30.09.2019, 11:56

Hi All,

Nice to see the community for this great game alive after so many years. Decided to hop into GP2 after a 20-year hiatus and loving it. There is still some awesome feeling remaining after all these years that the newer games have not reached.
Couple of questions though. I decided to start a career by doing non-championship races and then calculating the points afterwards in Excel. Do I miss any features by doing this, or is it essentially the same as running the races in a championship? Are all the random performance changes in AI cars represented this way as well?

The second one is the Random Power Variation in GP2edit. Should I have it on or off? What does it do?

User avatar
rremedio
GP2 Star
Posts: 1307
Joined: 14.01.2007, 15:05

Re: Non-championship races vs. championship difference

Post by rremedio » 30.09.2019, 15:46

Hi and welcome!

The only thing you should be aware of about running non-championship races is that when you fire up GP2, the random seed is always the same, so it will generate the same random "events" every time. Once you joined a session and leave it, it will create another random seed and it will be truly random this time, so I suggest you always join a quickrace and leave it (it takes just a few seconds) when you open GP2, before joining your non-championship race.

About the Random Power Variation, most people will leave it off because it makes the performances between teams less reliable. It is a way of GP2 simulating the fact that some cars will run better in some tracks than in others, but it isn't very well done. But anyway, leaving it on won't bring any harm and it will increase the challenge from time to time, so maybe you can have fun with it.
"The Fox provides for himself. but God provides for the Lion" - The Marriage of Heaven and Hell by William Blake

Image

User avatar
samsepi0l
GP2 Starter
Posts: 71
Joined: 01.03.2019, 01:39

Re: Non-championship races vs. championship difference

Post by samsepi0l » 30.09.2019, 19:09

I just did the exact same thing- (ran a full season of races and used a spreadsheet to keep track of points)

I did not know about the "quickrace" trick. Can you elaborate on this more? You caught my attention here.

I also noticed that 1986 senna, seemed unstoppable. In the driver grip settings I had them set where 1986 prost actually had a little better score- yet senna kept beating him when it came to races and qualifying. Prost had a value of 16215 (which was the highest of any driver) and Senna had 16005.

User avatar
rremedio
GP2 Star
Posts: 1307
Joined: 14.01.2007, 15:05

Re: Non-championship races vs. championship difference

Post by rremedio » 30.09.2019, 19:31

samsepi0l wrote:
30.09.2019, 19:09
I did not know about the "quickrace" trick. Can you elaborate on this more? You caught my attention here.
Like many games, GP2 uses a random number generator. It works like that: let's say some driver has 16000 grip and 300 random grip range. GP2 will get a random number between 0 and 300 and subtract it from 16000 to define the actual grip of that driver in a session. GP2 does the same for several things like the order cars enter the track in qualy, pit strategies, failures and cc accidents, etc...

But those numbers aren't actually random. GP2 uses a complex algorithm to produce a random number but it depends on a "seed" number to do it. When the seed is the same, random numbers will be generated always in the same order. If you change the seed, the numbers generated change, but they will always be the same (in the same order) for that particular seed.

GP2 is always started with the same seed and that's why if you start it and go directly to a race, grid positions, car failures, accidents between the ccs etc will always be the same.

But after you leave a race (changing sessions during a championship or non-championship race doesn't count), GP2 will change the seed and it will not be the same seed every time, so entering a quickrace and leaving it makes sure your next race parameters is actually random because while the first seed is always the same, the second is not.

If you always go directly to the race when you start the game, many things will always be the same.
"The Fox provides for himself. but God provides for the Lion" - The Marriage of Heaven and Hell by William Blake

Image

User avatar
samsepi0l
GP2 Starter
Posts: 71
Joined: 01.03.2019, 01:39

Re: Non-championship races vs. championship difference

Post by samsepi0l » 30.09.2019, 20:14

Ahhhh!!!- so all I have to do is click quickrace, and when I see the red lights I can exit... literally in fractions of a second? Do I need to wait out anything...?

This is very useful information!

Senna won 10 of 16 races in 1986, it's starting to make sense now. ](*,)

User avatar
minardi
GP2 Newbie
Posts: 45
Joined: 28.02.2018, 04:35
Location: United States

Re: Non-championship races vs. championship difference

Post by minardi » 30.09.2019, 21:48

samsepi0l wrote:
30.09.2019, 20:14
Ahhhh!!!- so all I have to do is click quickrace, and when I see the red lights I can exit... literally in fractions of a second? Do I need to wait out anything...?

This is very useful information!

Senna won 10 of 16 races in 1986, it's starting to make sense now. ](*,)
Yes that's all you have to do. I remember recognizing this when I was younger and did multiplayer races as every car and skipped qualifying and the grid for every race would be the same when that was the first thing I did when entering the game. I didn't understand why at the time but it's nice to get a more thorough explanation

User avatar
GP2tifoso28
GP2 Starter
Posts: 81
Joined: 29.09.2015, 14:21

Re: Non-championship races vs. championship difference

Post by GP2tifoso28 » 30.09.2019, 22:19

Laahustaja wrote:
30.09.2019, 11:56
Hi All,

Nice to see the community for this great game alive after so many years. Decided to hop into GP2 after a 20-year hiatus and loving it. There is still some awesome feeling remaining after all these years that the newer games have not reached.
Couple of questions though. I decided to start a career by doing non-championship races and then calculating the points afterwards in Excel. Do I miss any features by doing this, or is it essentially the same as running the races in a championship? Are all the random performance changes in AI cars represented this way as well?

The second one is the Random Power Variation in GP2edit. Should I have it on or off? What does it do?
In my opinion there is a difference. I believe there is a master Ai applied during a season, which gives varied results so every season isn't 100% identical. But it is more than the random seed, it somehow applies to the whole season.

For example, I have a season results sheet where Schumacher only won 1 race, which is unheard of in gp2. Alesi won 3 or 4 races and Berger won 1, which is basically unheard of. I even have the results sheet if anybody doesn't believe me. Normally Ferrari are slow I don't recall them ever winning in gp2 unless other people had problems or it was at Monza or hockenheim.

I have run other seasons where Schumacher won everything. Then I've seen seasons online where Coulthard dominated, something I never remember happening. He was usually slower than Hill from my memory.

There is a "consistency" to the "randomness" when playing full seasons, but it doesn't always stay consistent from season to season. This is just my personal observations and not based on any actual facts or advanced technical knowledge of the sim.
rremedio wrote:
30.09.2019, 15:46
Hi and welcome!

The only thing you should be aware of about running non-championship races is that when you fire up GP2, the random seed is always the same, so it will generate the same random "events" every time. Once you joined a session and leave it, it will create another random seed and it will be truly random this time, so I suggest you always join a quickrace and leave it (it takes just a few seconds) when you open GP2, before joining your non-championship race.

About the Random Power Variation, most people will leave it off because it makes the performances between teams less reliable. It is a way of GP2 simulating the fact that some cars will run better in some tracks than in others, but it isn't very well done. But anyway, leaving it on won't bring any harm and it will increase the challenge from time to time, so maybe you can have fun with it.
Ok now this is fascinating. How did you figure it out? I'm going to do some testing and I hope to report the results on here. Will the seed be the same seed for everyone or the same seed unique to the user? As in if we both ran an AI race with the same track and settings after launching the game, we'll see the same results?

User avatar
rremedio
GP2 Star
Posts: 1307
Joined: 14.01.2007, 15:05

Re: Non-championship races vs. championship difference

Post by rremedio » 30.09.2019, 23:25

Idk who figured it out, probably René Smit since there is a feature in GPxPatch that allows you to use a fixed seed (so many people can try the same race, and very important in GP3+ competitions since they have the possibility of changing weather conditions, which is also attached to the random number generator). The only thing I figured out by myself is that the second seed is not always the same.

I just noticed it works the same in GP2 after some testing. And the whole thing is very commonplace in video games development since forever (Pitfall! on the Atari 2600 used the same kind of generator and it used a fixed seed so everybody played the same game, despite the "level" layout being random generated).

Idk if the starting seed is fixed between GP2 installations and different computers, I would assume it is, but I wouldn't be surprised if it used system data to pick the starting seed (another not very uncommon pattern in games).

Apart from using the "quickrace trick", another thing you can do by knowing that is checking how changing physics, performance data or magic data will affect a race by always checking the first race when you open GP2 after changing those things.

For full seasons idk how much of it depends on the seed that starts the season or if new seeds are created for each race. Since the numbers sequence can go very very long in a single seed, it could be either way.
"The Fox provides for himself. but God provides for the Lion" - The Marriage of Heaven and Hell by William Blake

Image

Laahustaja
GP2 Newbie
Posts: 9
Joined: 30.09.2019, 11:33

Re: Non-championship races vs. championship difference

Post by Laahustaja » 01.10.2019, 07:36

Ok, thanks a lot for the answers! So for the best experience, always hop on into a quickrace when starting the game.

On the random seed having consistency during the seasons, I would guess that that is then the effect of the seed being generated already in the game start. Would probably be easy to test be loading a save and proceeding to the next race. If the events are the same each time, this would be the effect of the "long seed" being already generated in the beginning of the season.

On my purpose though I guess running non-championship races would pretty much yield similar experience with driver performance varying a bit during the season (given that I remember the quickrace trick!), as the randomness is still in place despite the seed being generated in the start of the race instead of beginning of the season.

I think the seed-mechanism is used quite a lot still in games, at least in modern Firaxis games like new XCOMs and Civilization IV I have noticed it being used.

User avatar
GP2tifoso28
GP2 Starter
Posts: 81
Joined: 29.09.2015, 14:21

Re: Non-championship races vs. championship difference

Post by GP2tifoso28 » 01.10.2019, 10:14

Ok I've done some testing. I ran 3 consecutive non championship AI races at interlagos, ace level, 50% distance. Then I quit the game and repeated the process 3 times over.

The results: exactly the same each time. As in, the results from race 1 after each reboot, were the same. Then the results of the 2Nd race were the same and 3rd race, were the same also. The algorithm is following the same sequence, 1 seed seems to spawn the next seed in the same order.

I tried the same thing again, but this time running 3 consecutive non championship AI races on Ace AFTER running 1 quickrace at catalunya. I repeated the process multiple times.

The results:
Different results in race 1, 2 and 3. BUT - When you quit, reboot, run catalunya quickrace, then 3 50% races, you get the same result every time.

So race 1 - race 1 - race 1, same result after each reboot, but DIFFERENT from the first test without running catalunya 1st

Same with race 2 and 3, the same results each time, but different from the 1st test. So the seed seems to be following a set sequence.

So just to clarify, the results of race 1 on each boot of the game were the same. Then the results of race 2 were the same as each other, but different from race 1 or race 3.

Then after a quickrace, the same pattern repeats, but with different results. So race 1 was the same every time, race 2 was the same every time ect.

If anyone has any more information or data I would like to see it. I can't help but wonder if this might be a bug related to dosbox emulation because the seed should really be RANDOM each time, not following a set sequence

User avatar
GP2tifoso28
GP2 Starter
Posts: 81
Joined: 29.09.2015, 14:21

Re: Non-championship races vs. championship difference

Post by GP2tifoso28 » 01.10.2019, 10:38

Ok ill try and make this simple. Start your game, De select all drivers, set it to Ace 50% distance, run a non championship race at interlagos. Tell me if this is your results :

ÚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ¿
³ Race Results ³
³ ------------ ³
³ Brazil - Interlagos ³
³ (36 Laps) ³
³ ³
³ Pos Driver Time/Speed ³
³ ³
³ 1st Michael Schumacher 48m 16.598s ³
³ Benetton - Ford ³
³ 2nd David Coulthard +6.045s ³
³ Williams - Renault ³
³ 3rd Damon Hill +25.640s ³
³ Williams - Renault ³
³ 4th Jean Alesi +27.119s ³
³ Ferrari - Ferrari ³
³ 5th Mika Hakkinen +44.907s ³
³ McLaren - Peugeot ³
³ 6th Gerhard Berger +46.013s ³
³ Ferrari - Ferrari ³
³ 7th Rubens Barrichello +55.048s ³
³ Jordan - Hart ³
³ 8th Eddie Irvine +57.810s ³
³ Jordan - Hart ³
³ 9th Ukyo Katayama +1m 07.354s ³
³ Tyrrell - Yamaha ³
³ 10th Johnny Herbert +1m 16.108s ³
³ Lotus - Mugen Honda ³
³ 11th Olivier Panis -1 Lap ³
³ Ligier - Renault ³
³ 12th Gianni Morbidelli -1 Lap ³
³ Footwork - Ford ³
³ 13th Jos Verstappen -1 Lap ³
³ Benetton - Ford ³
³ 14th Eric Bernard -1 Lap ³
³ Ligier - Renault ³
³ 15th Andrea de Cesaris -1 Lap ³
³ Sauber - Mercedes ³
³ 16th Mark Blundell -1 Lap ³
³ Tyrrell - Yamaha ³
³ 17th Erik Comas -1 Lap ³
³ Larrousse - Ford ³
³ 18th Heinz-Harald Frentzen -1 Lap ³
³ Sauber - Mercedes ³
³ 19th Martin Brundle -1 Lap ³
³ McLaren - Peugeot ³
³ 20th Christian Fittipaldi -1 Lap ³
³ Footwork - Ford ³
³ 21st Alessandro Zanardi -1 Lap ³
³ Lotus - Mugen Honda ³
³ 22nd Olivier Beretta -2 Laps ³
³ Larrousse - Ford ³
³ 23rd David Brabham -2 Laps ³
³ Simtek - Ford ³
³ 24th Jean-Marc Gounon -2 Laps ³
³ Simtek - Ford ³
³ 25th Michele Alboreto Loose Wheel ³
³ Minardi - Ford ³
³ 26th Pierluigi Martini Transmission ³
³ Minardi - Ford ³
ÀÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÙ

Now without quitting the game, run another non championship race at interlagos, and tell me if this is your result:

³ Race Results ³
³ ------------ ³
³ Brazil - Interlagos ³
³ (36 Laps) ³
³ ³
³ Pos Driver Time/Speed ³
³ ³
³ 1st Michael Schumacher 48m 05.970s ³
³ Benetton - Ford ³
³ 2nd Damon Hill +14.919s ³
³ Williams - Renault ³
³ 3rd David Coulthard +36.681s ³
³ Williams - Renault ³
³ 4th Gerhard Berger +45.649s ³
³ Ferrari - Ferrari ³
³ 5th Jean Alesi +58.927s ³
³ Ferrari - Ferrari ³
³ 6th Rubens Barrichello +1m 06.201s ³
³ Jordan - Hart ³
³ 7th Mika Hakkinen +1m 08.106s ³
³ McLaren - Peugeot ³
³ 8th Martin Brundle -1 Lap ³
³ McLaren - Peugeot ³
³ 9th Jos Verstappen -1 Lap ³
³ Benetton - Ford ³
³ 10th Mark Blundell -1 Lap ³
³ Tyrrell - Yamaha ³
³ 11th Olivier Panis -1 Lap ³
³ Ligier - Renault ³
³ 12th Eddie Irvine -1 Lap ³
³ Jordan - Hart ³
³ 13th Eric Bernard -1 Lap ³
³ Ligier - Renault ³
³ 14th Ukyo Katayama -1 Lap ³
³ Tyrrell - Yamaha ³
³ 15th Erik Comas -1 Lap ³
³ Larrousse - Ford ³
³ 16th Heinz-Harald Frentzen -1 Lap ³
³ Sauber - Mercedes ³
³ 17th Johnny Herbert -1 Lap ³
³ Lotus - Mugen Honda ³
³ 18th Gianni Morbidelli -1 Lap ³
³ Footwork - Ford ³
³ 19th Andrea de Cesaris -1 Lap ³
³ Sauber - Mercedes ³
³ 20th Michele Alboreto -1 Lap ³
³ Minardi - Ford ³
³ 21st Jean-Marc Gounon -2 Laps ³
³ Simtek - Ford ³
³ 22nd Pierluigi Martini Didn't Finish ³
³ Minardi - Ford ³
³ 23rd Alessandro Zanardi Transmission ³
³ Lotus - Mugen Honda ³
³ 24th Christian Fittipaldi Transmission ³
³ Footwork - Ford ³
³ 25th Olivier Beretta Engine ³
³ Larrousse - Ford ³
³ 26th David Brabham Puncture ³
³ Simtek - Ford ³
ÀÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÙ

Now run a 3rd race, without quitting:

ÚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ¿
³ Race Results ³
³ ------------ ³
³ Brazil - Interlagos ³
³ (36 Laps) ³
³ ³
³ Pos Driver Time/Speed ³
³ ³
³ 1st David Coulthard 48m 17.132s ³
³ Williams - Renault ³
³ 2nd Michael Schumacher +4.565s ³
³ Benetton - Ford ³
³ 3rd Damon Hill +19.783s ³
³ Williams - Renault ³
³ 4th Gerhard Berger +37.149s ³
³ Ferrari - Ferrari ³
³ 5th Olivier Panis +1m 03.434s ³
³ Ligier - Renault ³
³ 6th Rubens Barrichello +1m 04.220s ³
³ Jordan - Hart ³
³ 7th Mika Hakkinen +1m 06.614s ³
³ McLaren - Peugeot ³
³ 8th Gianni Morbidelli +1m 14.260s ³
³ Footwork - Ford ³
³ 9th Martin Brundle +1m 16.861s ³
³ McLaren - Peugeot ³
³ 10th Eddie Irvine -1 Lap ³
³ Jordan - Hart ³
³ 11th Andrea de Cesaris -1 Lap ³
³ Sauber - Mercedes ³
³ 12th Heinz-Harald Frentzen -1 Lap ³
³ Sauber - Mercedes ³
³ 13th Erik Comas -1 Lap ³
³ Larrousse - Ford ³
³ 14th Pierluigi Martini -1 Lap ³
³ Minardi - Ford ³
³ 15th Mark Blundell -1 Lap ³
³ Tyrrell - Yamaha ³
³ 16th Johnny Herbert -1 Lap ³
³ Lotus - Mugen Honda ³
³ 17th Alessandro Zanardi -1 Lap ³
³ Lotus - Mugen Honda ³
³ 18th Michele Alboreto -1 Lap ³
³ Minardi - Ford ³
³ 19th Jean-Marc Gounon -1 Lap ³
³ Simtek - Ford ³
³ 20th Jos Verstappen -2 Laps ³
³ Benetton - Ford ³
³ 21st Ukyo Katayama -2 Laps ³
³ Tyrrell - Yamaha ³
³ 22nd Eric Bernard -2 Laps ³
³ Ligier - Renault ³
³ 23rd Olivier Beretta -2 Laps ³
³ Larrousse - Ford ³
³ 24th David Brabham -2 Laps ³
³ Simtek - Ford ³
³ 25th Christian Fittipaldi -4 Laps ³
³ Footwork - Ford ³
³ 26th Jean Alesi Suspension ³
³ Ferrari - Ferrari ³
ÀÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÙ

Now quit the game, and repeat the exact process, and the results of the races should be exactly the same, every time, in sequence.

Now do the same thing but run a quickrace in catalunya first. Now run a 50% race at interlagos, compare your results:

ÚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ¿
³ Race Results ³
³ ------------ ³
³ Brazil - Interlagos ³
³ (36 Laps) ³
³ ³
³ Pos Driver Time/Speed ³
³ ³
³ 1st Michael Schumacher 48m 19.623s ³
³ Benetton - Ford ³
³ 2nd Gerhard Berger +9.279s ³
³ Ferrari - Ferrari ³
³ 3rd Damon Hill +17.668s ³
³ Williams - Renault ³
³ 4th David Coulthard +20.913s ³
³ Williams - Renault ³
³ 5th Jean Alesi +43.516s ³
³ Ferrari - Ferrari ³
³ 6th Jos Verstappen +51.698s ³
³ Benetton - Ford ³
³ 7th Gianni Morbidelli +55.157s ³
³ Footwork - Ford ³
³ 8th Olivier Panis +1m 07.511s ³
³ Ligier - Renault ³
³ 9th Mark Blundell +1m 12.288s ³
³ Tyrrell - Yamaha ³
³ 10th Eddie Irvine +1m 13.898s ³
³ Jordan - Hart ³
³ 11th Martin Brundle +1m 18.087s ³
³ McLaren - Peugeot ³
³ 12th Eric Bernard -1 Lap ³
³ Ligier - Renault ³
³ 13th Ukyo Katayama -1 Lap ³
³ Tyrrell - Yamaha ³
³ 14th Mika Hakkinen -1 Lap ³
³ McLaren - Peugeot ³
³ 15th Erik Comas -1 Lap ³
³ Larrousse - Ford ³
³ 16th Heinz-Harald Frentzen -1 Lap ³
³ Sauber - Mercedes ³
³ 17th Rubens Barrichello -1 Lap ³
³ Jordan - Hart ³
³ 18th Pierluigi Martini -1 Lap ³
³ Minardi - Ford ³
³ 19th Johnny Herbert -1 Lap ³
³ Lotus - Mugen Honda ³
³ 20th Olivier Beretta -1 Lap ³
³ Larrousse - Ford ³
³ 21st Andrea de Cesaris -1 Lap ³
³ Sauber - Mercedes ³
³ 22nd Michele Alboreto -1 Lap ³
³ Minardi - Ford ³
³ 23rd Jean-Marc Gounon -1 Lap ³
³ Simtek - Ford ³
³ 24th Alessandro Zanardi -1 Lap ³
³ Lotus - Mugen Honda ³
³ 25th David Brabham -2 Laps ³
³ Simtek - Ford ³
³ 26th Christian Fittipaldi -3 Laps ³
³ Footwork - Ford ³
ÀÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÙ

See how the results changed from race 1 in the first test? Now you are set on a new sequence. BUT, if you quit and repeat the same process, you'll get these exact same results AGAIN. So the seed is still not fully random, it's just different but the same each time under the same conditions

I'm now desperate to find a solution to this problem.

User avatar
GP2tifoso28
GP2 Starter
Posts: 81
Joined: 29.09.2015, 14:21

Re: Non-championship races vs. championship difference

Post by GP2tifoso28 » 01.10.2019, 12:18

Guess what? The same problem in championship mode. The interlagos race had the exact same result as the non championship race. Then Aida, was the same exact result after running the championship again. Surely theres a solution to this problem.

Now I run a non championship race at Aida, completely different result. But if I start the game, run non championship Brazil, followed by non championship Aida, I get the SAME RESULT as the 2 championship races.. What the hell.

Changing the race length slider by 1% affected the results, and changing the difficulty affected the results. So I'm willing to bet if you ran 2 full championships, with the exact same settings each time, launching the championship directly after booting the game, the championship results will be IDENTICAL.

Surely there's a way to fix this, because even after running a quick race you can still produce repeatable results if you replicate the same sequence of events under the same conditions. We need to find a fix for this

User avatar
GP2tifoso28
GP2 Starter
Posts: 81
Joined: 29.09.2015, 14:21

Re: Non-championship races vs. championship difference

Post by GP2tifoso28 » 01.10.2019, 12:22

Here's what Renè Schmidt says about the random seed, in relation to gpxpatch for gp3 and gp4:
Random seed
GPx uses pseudo random number generators to make each race a little different. These generators spit out apparently random numbers, but they're not. They are extracted from a starting sequence called a seed, and will start generating the exact same sequence again after a large amount of generated numbers. So, starting from a specific seed, the same number sequence is generated. Each time a number is extracted, the seed has changed (the extracted number is part of the current seed). This means the game can regenerate races. This is how replays in the game work, where the seed is part of the replay state. GPxSet's random seed feature will extend this by resetting the random seed to a user supplied value right before the start of the session, so all calculations that use random numbers will give the same results. Examples of such calculations include weather, performances and failures. Using this feature, offline leagues can make sure all participants are driving under the same conditions. The most important application of this is the same weather conditions for everyone, something that hasn't been possible before. It's best to use custom weather with a specific percentage for this, which can be specified using GPxSet too. For convenience, those settings are recorded along with the random seed.

Right now, the random seed feature only works for quickrace and unlimited practice. The used value is being logged by GPxLap and shown by HOFxLap.

Important: do not assume the same random seed will give the same results under different circumstances like a different track, different performances, etc., even while it may seem to work that way. A specific random seed has to be accompanied by a specific carset, track, car shape, etc. In other words, the ones who use the random seed should have the same conditions as the one who recorded the random seed. Most leagues already enforced the same conditions for everyone, so this shouldn't be a problem.

Also important: do not assume any pattern in the random seed value. For example, a higher value will not mean heavier raining, or better performances, etc.

Keep in mind that although the race conditions should be equal, it is of course influenced by user input, so no race can ever be the same unless you deselect all players. Also, testing has shown that GP3 1.00/1.13 gave slightly different results even when no players were selected. This should be considered a bug (maybe the same bug as the replay bug where a replay would sometimes be different from what actually happened).
I'm going to do some follow up testing on GP3 and GP4, if the problem doesn't exist there, then maybe Rene knows of a way to properly fix it

EDIT

I tried gp4, it's producing unique results every time (without gpxpatch)

EDIT

GP3, also producing completely unique results.

I'm starting to think that this might be a dosbox related bug, maybe the dosbox emulation is interfering with the seed code causing the same identical seed to launch every time. Gp3 and gp4 seems to launch a random seed on each boot.

Is anyone running gp2 on a proper DOS system that can run some tests?

eky78
GP2 Starter
Posts: 89
Joined: 05.02.2012, 12:15
Contact:

Re: Non-championship races vs. championship difference

Post by eky78 » 01.10.2019, 13:37

I'm starting to think that this might be a dosbox related bug, maybe the dosbox emulation is interfering with the seed code causing the same identical seed to launch every time. Gp3 and gp4 seems to launch a random seed on each boot.

Is anyone running gp2 on a proper DOS system that can run some tests?
I tend to agree...I´ve noticed this problem only when I started using DOSBOX...I think I had no problem with Win XP...we need someone who still uses Win XP for an answer (but not from a virtual machine as it would probably have the same issue of DOSBOX).
Ok I've done some testing. I ran 3 consecutive non championship AI races at interlagos, ace level, 50% distance. Then I quit the game and repeated the process 3 times over.
No you don´t have to quit the game...just quit the race (if your aim is not to have the same results again and again)

User avatar
GP2tifoso28
GP2 Starter
Posts: 81
Joined: 29.09.2015, 14:21

Re: Non-championship races vs. championship difference

Post by GP2tifoso28 » 01.10.2019, 13:54

eky78 wrote:
01.10.2019, 13:37
I'm starting to think that this might be a dosbox related bug, maybe the dosbox emulation is interfering with the seed code causing the same identical seed to launch every time. Gp3 and gp4 seems to launch a random seed on each boot.

Is anyone running gp2 on a proper DOS system that can run some tests?
I tend to agree...I´ve noticed this problem only when I started using DOSBOX...I think I had no problem with Win XP...we need someone who still uses Win XP for an answer (but not from a virtual machine as it would probably have the same issue of DOSBOX).
Ok I've done some testing. I ran 3 consecutive non championship AI races at interlagos, ace level, 50% distance. Then I quit the game and repeated the process 3 times over.
No you don´t have to quit the game...just quit the race (if your aim is not to have the same results again and again)
I was only quitting the game to see if the seed would be the same on each boot. Not only is it the same, but it follows the same sequence. I ran 3 races 3 times over on 3 separate boots and got 3 sets of identical results in the sequence in which the races were run.

Running a catalunya quickrace then changed the results. But if you reboot, run the catalunya quickrace, run your test race, the result will still be the same every time, only it will be a different result to the first test.

We really need someone running the sim natively to run some tests so we can diagnose if this is a gp2 problem or an emulation problem.

Post Reply