Fastest Possible Lap?

[billa_gordon]

Would I be right in saying that for every car on every track there is a fastest possible lap?

This isn't true in real life racing as there are so many variables, but in a game, there must be a finite number of ways to go round a track in a given car. has anybody ever thought that Polyphony Digital probably have or could work this out? I would imagine it would take a fair bit of time but surely its possible? Is there any computing whizz kid that has thought of a way of doing it??

Would be interesting to see how close the best times are to the perfect times, and if there are any perfect times on the online leaderboard even!!

If anybody has spotted a fallacy in my argument just let me know....

Gordon

 

[BobElHat]

Would I be right in saying that for every car on every track there is a fastest possible lap?

This isn't true in real life racing as there are so many variables, but in a game, there must be a finite number of ways to go round a track in a given car. has anybody ever thought that Polyphony Digital probably have or could work this out? I would imagine it would take a fair bit of time but surely its possible? Is there any computing whizz kid that has thought of a way of doing it??

It must be true even in real-life racing; at least for a given fuel level, air temperature, track temperature and conditions, tyre wear, wind conditions... If anybody in the real world does these kinds of calculations it would be F1 teams, particularly the big teams with huge budgets like McLaren, Ferrari etc. They certainly do have simulations so they know when is the optimum time to change tyres, make pit stops etc. but I suspect they are based on past real laps rather than a computer-generated ideal lap. They have to do something for new circuits though, so they probably have some way of at least approximating what lap times would be and how the variables affect it, even in the absence of hard data.

For GT I doubt that PD would have bothered, even for a game it would be very hard to calculate, there are far too many variables to just solve an equation and get the answer, so successive approximation techniques would be required - start with an estimate based on the classical ideal line and change it a bit to see if it's better, repeat till the changes are imperceptibly small. That kind of technique can get trapped in a "local minimum" (any small change makes it worse, but a larger change than your algorithm uses might make it better), so wouldn't prove the lap was ideal. Trying every possible combination of line, braking point, pedal and steering modulation would be a monumental computing effort.

For a small insight into what would be involved, the Physics of Racing series does a simplified analysis of a single corner in parts 17 and 18. To that you'd need to add proper tyre and suspension models, kerbs, wind...

 

[amar212]

Ther is always a .001 behind any time. Believe.

 

[yesbossdodd]

Im sure there is a fintite time for all courses, cars and variables. A small piece of knowledge I managed to gather is that McLaren use Hamiltons quickest lap around Spa last season to test their engines to destruction! I found this quite interesting as sureley Monza would have a higher wear rate on engines!!

Just my 2 cents!

 

[RAVELLRON]

however with ps3 this gt version it seem like reality... i feel many variability in brake approach than in gt4 for ps2, the time are really difficult to do constantly in every lap...

 

[billa_gordon]

very true BobEIHat. thanks for that link too, very interesting. I would hate to get a perfect lap in GT and never know, but that's never going to happen to me anyway! but you never know, there might be a perfect lap out there. I think someone should set up an experiment with an infinite number of monkeys playing on an infinite number of PS3s to test this theory!

 

[Kneedragger]

I like this question a lot, and have wondered before what the truth is behind it.
My personal opinion is that tt mode in gt5p is not totally objective, and that gameplay issues determine a percentage of what lap times are possible. As an example I've found it helps sometimes to leave tt mode and go back to the main menu and then go back into tt mode and continue.. its a strategy to overcome the gameplay built into tt mode when you are hotlapping. Its as if the program registers your laptimes and begins to 'play-itself' in a very subtle way.

Another one, once you are warmed up, is to leave tt mode and come back but NOT cross the finish line unless your lap was very near your best (just keep restarting instead.) In other words dont post a time unless it was a good one. Either way, I am convinced that there is some amount of programming intervention, or gameplay, that intrudes on tt mode laptimes.

 

[P1]

It would be good to know the fastest possible lap time but i dont think anyone could actually do it though.

 

[balang_479]

This is quite an interesting question... i think this would be much simpler in like GT1 where there were much less variables involved, but now with GT5P with so many variables it would almost impossible, or would take a ridiculous amount of time to calculate it as its gets much closer to reality...

 

[Wolfe]

In videogames, there's always an absolute fastest possible time. In the case of Gran Turismo, it isn't likely to be reached by any human player, but there's nothing unpredictable that can prevent it from happening.

 

[santiagomo87]

If PD comes with the fastest possible lap time for, say Susuka, I'm sure there will be some Japanese guy that will beat it

 

[Exorcet]

If PD comes with the fastest possible lap time for, say Susuka, I'm sure there will be some Japanese guy that will beat it

That just means someone messed up the calculation.

Theoretically, there should be a limit to how fast a car can get around a certain track. We all know a time of 0 seconds is impossible, so the number must exist. Finding it would be the difficult part.

 

[santiagomo87]

That just means someone messed up the calculation.

Theoretically, there should be a limit to how fast a car can get around a certain track. We all know a time of 0 seconds is impossible, so the number must exist. Finding it would be the difficult part.

yeah I know that, it was only a joke. These guys are just unexplainably fast.

 

[snowgt]

It's pretty much impossible to calculate, because you would have to make a perfect model (as in equations) of the whole environment and car capabilities, which is possible, but would take a) a lot of time to put together, and b) an insane amount of time to solve, because the number of variables would be so large. Every point in time influences all others after it, so it factors into the best possible solution for the problem. It would be faster to make a brute-force attempt, but also for that, the number of possible inputs (i.e. steering & acceleration/braking) is so massive, it doesn't make sense. Just think, if steering would only be taken into account 100 times/sec. (which seems a bit low, but i have no idea what the real frequency is). Now suppose you only have 100 different positions of the wheel. That means you would get 100^100 theoretically different possibilities to steer in 1 second. That's a 1 with 200(!) zeros. Of course, you could throw out some impossible ones with some rules, but a huge number of possible paths to the start/finishing line remains. And that example was with low assumptions and only ONE input variable.

The modelling that e.g. Formula 1 teams do, are still crude approximations of reality and will NOT give you the real limits. They are basically working with simplifications that will give a raw estimate, just so you get some idea of what might be possible.

 

[RussRobit]

If anyone is a physics and/or math major, this would be an excellent subject for his/her thesis.
If you could freeze all the physical variables that may change over a single lap (temp, wind speed, gravitation forces from the Moon etc.) that would be involved in calculating a perfect lap and you were given absolute track boundaries, then you would be left with the laws of physics.
The laws of physics are exactly what they say they are, laws. It may sometimes seem that a driver is breaking them, but rest assured, he is not. He is however driving at the edge of what is possible.
To drive the perfect lap would be of superhuman proportions, to calculate it would take a super computer.

 

[Mr Latte]

Fastest Human - See the leaderboards...

Fastest Possible Technical Simulation - Simulation isnt real so whys it important, how on earth do you simulate real world conditions, you cant.
Techincally yes you can achieve the fastest line, and a cars grip factor limits within a controlled/limited enviroment (which is what a game is and importantly real life and humans arnt) but it's always going to be that a simulation. So i question whats the point of a simulation that is impossible in the real world by a real person?

Technically could you create an AI to run every lap perfect within the games possibilities to 100% yes you could but the question then would be why, no human can be expected to do that.
On a real track with a real car and real person its the opposite as you can only model your simulation on real factors or pervious abilities, conditions and un-accountable factors .


You can simulate how two massive steal frammed buildings collapse from fire and get the majority of the general public to believe it with fancy talk and computer 3D simulations but in the real world if its never happened with any other building then what is more likely?

As humans we continuely push the boundaries, the Oylimpic Games are just an example.


Seriously though some of you should get out more if these thoughts consume your day.

 

[GTP_rpm]

Im sure there is a fintite time for all courses, cars and variables. A small piece of knowledge I managed to gather is that McLaren use Hamiltons quickest lap around Spa last season to test their engines to destruction! I found this quite interesting as sureley Monza would have a higher wear rate on engines!!

Just my 2 cents!

Monza has low downforce.

👍

 

[billa_gordon]

You can simulate how two massive steal frammed buildings collapse from fire

do you not mean three steel framed buildings?

 

[The Outlaw]

Determining a theoretical perfect lap is highly complicated and nearly impossible to calculate in reality. There are so many variables to take into account for a given car and track. The biggest problem is the fact that not all cars achieve their fastest lap running the same line. Some cars are quicker with a later apex and vice versa. Calculating this is nearly impossible as far I could see as there would be no UNIVERSAL baseline for the AI to go by for each car and track. Getting the most out of the three inputs (gas, brake, steering) wouldn't be quite as difficult, but would still be a hell of a challenge as each car has a optimal slip angle in all aspects of cornering (entry, apex, exit).

It would be interesting to determine what the theoretical fastest lap possible is, but to me it's a pretty unrealistic expectation. Until then we have the top racers in the IFTC to go by. Their laps are as about close to perfect as any human being is going to get, which is all that really matters when you really get down to it anyway.

 

[greenlightning]

They must have some fairly detailed programing for A.I. and "B"-spec regarding lap times. However for both they still need enourmuos improvements. Would like to see tire wear etc. to be on or ability to be turned on when doing hot laps to make it more close to race conditions. Also would help you decide what tires to use in a race.

GT series may lack a lot with actual racing but with enough options they can easily make up for it in car testing etc.

 

[DaddyVDUB]

Given the fact that the TT list include prob the best from all over the Globe the top times listed are surely very close, to what an actual player can obtain.

 

[ASH32]

They must be a wall, so too speak, that a driver can not break, a limit on how fast you can go, but finding this out, even if you can remove the variables is impossible, as it is not an exact science, you can't repeat lap after lap exactly and even with a fast lap, how do you know that it was perfect, it probaly wasn't. The top times in the TT are probably the best bet of finding a limit, but it is in no way the absolute best time, I can pretty much say that no matter what time you set, or how fast you are nothing is unbeatable, no one is a machine that can't make mistakes and run the perfect racing line, brake perfectly, shift at exactly the right time etc, and there could allways be someone faster, it just won't be me.

 

[slicksabo]

I like this question a lot, and have wondered before what the truth is behind it.
My personal opinion is that tt mode in gt5p is not totally objective, and that gameplay issues determine a percentage of what lap times are possible. As an example I've found it helps sometimes to leave tt mode and go back to the main menu and then go back into tt mode and continue.. its a strategy to overcome the gameplay built into tt mode when you are hotlapping. Its as if the program registers your laptimes and begins to 'play-itself' in a very subtle way.

Another one, once you are warmed up, is to leave tt mode and come back but NOT cross the finish line unless your lap was very near your best (just keep restarting instead.) In other words dont post a time unless it was a good one. Either way, I am convinced that there is some amount of programming intervention, or gameplay, that intrudes on tt mode laptimes.

HAHAHA dude are you serious?!?!?! What a waste of time.....this is certainly not the case. This is all in your head