Ok...now...I never really fully got a sure answer on this. I've heard many many things from multiple different people.

As most, if not all of you know, LFS calculates (or shows) down to the hundredth decimal place.

And, as most if you already know (either from experience or just plain logic) that calculation is usually not close enough for oval racing and finishes (and sometimes this isn't enough for circuits as well).

Does LFS even calculates below the hundredth decimal place? Is the scoring chart accurate if...lets say 1st and 2nd's finish time is the same?

Any clarification (or discussions) would be appreciated.

Lizard

The physics update loop runs at 100Hz, so the timing runs at 100Hz too. With the InSim protocol times (split times, laptimes etc) are reported in milliseconds but the value at the thousandth of a second is always 0. I believe (though I don't know if it's accurate) that the positions LFS displays for a dead heat are based on which player's update packet was received from first.

Thousands must be recorded, because if you look at normal hotlaps, although it only gives to hundreds (x.xx.XX), if you check the NUTTER ranks, a lot of the combined times show times down to thousands (x.xx.XXX).

Don't know why the WR times don't use these thousands though.




Edit: Just noticed that I'm the only "nutter" on the page with a zero as a thousand time!

That's because the Handicap is the person's total hotlap times under 103% of each hotlap WR, ie the thousandths will have a value with the WR * 1.03.

Quote from lizardfolk :

Does LFS even calculates below the hundredth decimal place? Is the scoring chart accurate if...lets say 1st and 2nd's finish time is the same?

As far as I know, no...

When two racers finish at exactly the same time, the one with the faster last lap is ranked before the other...

I suppose thousandths would be nice one day. Anything more - 10 thousandths - isn't necessary in LFS or real life, but it is something to wow the fans with (who think that having timing to 10 thousands of a second actually helps).

Quote from tristancliffe :

I suppose thousandths would be nice one day. Anything more - 10 thousandths - isn't necessary in LFS or real life, but it is something to wow the fans with (who think that having timing to 10 thousands of a second actually helps).

Yea but some times there was such situation that racers finished at same time so this could help, but not so nessary

Quote from bbman :

When two racers finish at exactly the same time, the one with the faster last lap is ranked before the other...

Not true (see attachment).

Quote from amp88 :

Not true (see attachment).

What was Tanakan's last lap ?

Quote from amp88 :

Not true (see attachment).

not the best lap, the last lap

Quote from tristancliffe :

I suppose thousandths would be nice one day. Anything more - 10 thousandths - isn't necessary in LFS or real life, but it is something to wow the fans with (who think that having timing to 10 thousands of a second actually helps).

http://spreadsheets.google.com ... FI1k7pKLg&output=html

Also. The interval between 2nd and 3rd at indycar kansas in 2005 was 0.0001. So tell me that doesnt matter in real life

EDIT: misread tristan's post

Ah, sorry. I misread the post. His last lap was faster than mine.

EDIT:

Never mind...I misread amp88's post

Quote from lizardfolk :

Sooooo....ok that takes cares of ties during a race...what about qualifying?

First one to do the lap gets the higher position.

EDIT: And @ 250 and 500, deko manually sets first one to do it gets pole

Quote from J@tko :

First one to do the lap gets the higher position.

Yeah...that's very unfair. Especially if you get the lower cue (which I always get :razz.

Quote from J@tko :

First one to do the lap gets the higher position.

EDIT: And @ 250 and 500, deko manually sets first one to do it gets pole

Wrong - I do believe. When two racers finish at the 'exact same time' it goes by connection ID or player ID (not sure). The person who has a lower connection, or player ID will be placed above the other.

-----------------

That said I don't think this has anything to do with the resolution of the physics engine at 100hz. While we all know and can agree the physics runs at 100hz it can give exact times to 7 or 8 decimal places if designed to do so...

For instance Update 5000 car A is near finish line, and car B is a little closer but not crossing yet. Update 5100 now both cars have both crossed the finish line and their times will both read as (either 5000 or 5100 depending on the algorithm). The correct way to do this is to;

CarPrevPos = Car.Pos;
UpdateCar;
CarNewPos = Car.Pos;

Find the time 't' at which the car crossed the line, this is independent of how fast, or slow, the physics system will run. t will be a value between 0 and 1, 0 representing the car crossed the line at PrevPos, and 1 representing NewPos. 0.5 would be directly in the center. The cars lap time would then be; 5000 + (0.01 * t); where 0.01 represents the time passed during a 100hz update. This will give you the accurate number...

LFS does not calculate this level of accuracy, I know this from some of the several ties I have had in the passed, some I've 'won' without being the 'winner'. It was clearly visible, in slow motion, that the other car crossed the line just slightly ahead.

Quote from blackbird04217 :

LFS does not calculate this level of accuracy, I know this from some of the several ties I have had in the passed, some I've 'won' without being the 'winner'. It was clearly visible, in slow motion, that the other car crossed the line just slightly ahead.

I'm by no means a programmer. But this theoretically means that this can be relatively easy to have LFS calculate to the thousandth?

Yes that precisely what I was saying, although it should be more precision than just 'thousandths'

Quote from blackbird04217 :

Wrong - I do believe. When two racers finish at the 'exact same time' it goes by connection ID or player ID (not sure). The person who has a lower connection, or player ID will be placed above the other.

Do you have any proof for that? In my experience it has been as Jatko said (first one to do the lap gets the higher position), as it is in F1.

Its has been my experience in the races where I've ended as a tie. I know that the other driver 'crossed the line first' but yet I was the one who 'won the race'.

Considering I _should_ have the replay I will try and dig it up, I've had several ties in my LFS career and this is the best pattern I've seen, I know that I had joined the server before the person who had placed second, which is why I was basing it on connection ID, although as mentioned it is possible it is based on Player ID (which makes more sense in the case of single player with AI).

*Looks through 1000s of replays for my ties.

Quote from blackbird04217 :

Its has been my experience in the races where I've ended as a tie. I know that the other driver 'crossed the line first' but yet I was the one who 'won the race'.

Considering I _should_ have the replay I will try and dig it up, I've had several ties in my LFS career and this is the best pattern I've seen, I know that I had joined the server before the person who had placed second, which is why I was basing it on connection ID, although as mentioned it is possible it is based on Player ID (which makes more sense in the case of single player with AI).

*Looks through 1000s of replays for my ties.

I was talking about qualifying as Jatko was. When it comes to racing the explanation given above (the player with the faster last lap) seems more reasonable than player/connection ID.

Unfortunately I can not find the replay of the one I want, although under careful examination of these replays I am no-longer sure how it is done. By putting it in slow motion it looks to be Jacko's idea might be correct although in some cases it also goes with my idea that the connection/player id also backs up the theory. Either way if someone wants some replays of ties I've found 6 of mine, although I know there was one in particular where the 'wrong' car one by a very visible distance.

Quote from amp88 :

I was talking about qualifying as Jatko was. When it comes to racing the explanation given above (the player with the faster last lap) seems more reasonable than player/connection ID.

In the event of a 1 lap race, which most of my ties are in who has the fastest lap? Hint this is the reason why Head-to-Head started in the first place. When testing 1 lap races de Souza and I experienced our first tie. I made a you-tube movie of it, and its included in the TIES package. I get for longer races the user with the faster laptime is better detection than player id / connection id but best solution overall would be to compute the accurate time crossed - which would likely eliminate almost _all_ possibilities of ties because the precision would be a lot higher.

NLR Peter (MPL) likes to save replays that end in a tie. Maybe he has enough to view to get some sort of consistant answer......