rec.autos.simulators

NASCAR/Hawaii Modem Latency Bug

David Spark

NASCAR/Hawaii Modem Latency Bug

by David Spark » Fri, 16 Aug 1996 04:00:00

Joe Farr asked me to post the following message to R.A.S for him. It's a
very understandable explanation of one of the problems that plagues Hawaii
and why determining blame for incidents is so difficult.

Dave "davids" Sparks
Late Nite League
http://www.racesimcentral.net/

----- forwarded message -----

Modem Latency Bug  on the Nascar racing Multi-Player server  (Hawaii)

The term bug is used loosely as this may or may not be considered a bug.  Some
drivers refer to this as warping or game quirk.  I have informed Hawaii about
this, but heard no response on their part if this is a real bug or not OR
if they
even plan to fix or address it in a later update for Nascar (one) or Nascar
(two).  Enough of the disclaimer crap....

The short n sweet version of this bug is basically:  It is possible that
you can
be hit by another car (that causes you to crash) while the other car does not
real hit you and continues on while watching you spin in his mirror.  Too
bad for
you!

WHY IS THIS??????

<This is MY impression of how the game works>

While racing other cars over the modem, it takes too long for my computer get
information on the CORRECT location of your car so my computer can display the
graphics.  Therefore, my computer has to guess  where your car is going to
be.  
As soon my computer gets the correct location of your car, my computer will
correctly draw your car.  This could explain why cars appear to jump/warp
on the
screen.

With this in mind, I will now give you a quick (and popular) example of a
MODEM
LATENCY BUG accident.

Driver A makes a pass to the inside of driver B while going into turn 1 at
Talladega.  Driver A keeps his car completely under the solid white line while
driver B keeps his car on the high side of the lane 1 (lower lane).  In
reality,
these two cars will not touch each other and should exit turn 2 quite safely.  
However, what happens more times than not:

Driver A remains below the solid white line and exits turn 2 safely while
hearing
driver B spin off into the wall mysteriously.  

Driver B remains on the high side of lane 1 and Driver A drifts up into Driver
Bs car causing him to crash.  BUT WAIT!!!! Driver A continues at full
speed and
suffers no damage.  WHAT THE HELL!!!

If both drivers were to save a replay of this example, both drivers would find
that Driver As replay shows NO CONTACT while driver Bs replay shows CLEAR
CONTACT.  

Driver Bs computer guessed  that driver A would drift up the track
(remain in a
straight line to last known trajectory) which would cause contact and probable
accident.  This happens ALL THE TIME at ALL TRACKS!!!!  It normally happens in
turns, but has been known to happen on straight-aways where Driver As
computer
guesses  that driver Bs car (following car while heading into a sharp
turn) IS
NOT braking (and plows into driver A) when in actuality, driver B is
braking and
DOES NOT hit driver A.

What can you do about it?

Until Hawaii fixes this problem (bug), there is nothing you can do about it.  
Since this normally happens in close quarter racing, I HIGHLY RECOMMEND not
racing side by side in the turns.  If you have no choice (in the heat of
battle),
please, for your sake, allow plenty of room between your car and the
passing car.
 The outside car never wins in this situation.  If you are going to be passed
(car gets inside position while heading down straight or car behind is clearly
faster), I recommend lifting/braking a tad early to allow the passing car
to get
completely by you BEFORE you fully enter the turn so you can use the full
track
(with no cars on your left or right) to safely complete the turn.  On the road
courses, I recommend not following directly behind the car ahead of you.  

I have written about this bug in my Drivers Guideline   <http:www.ivga.com  
IVGA MP section> several months ago.  Most League  accidents can be
attributed
to this bug.  If you are aware of the possibility of disaster, than you
will know
how to avoid it.

When in a race and you are taken out, please consider the following:

1. Did the car that hit you spin and crash also?

If he did, than one of you screwed up and the MODEM LATENCY BUG did not
bite you.
 Begin with unnecessary verbal flaming if you under the age of 15.

2. Did the car that hit you continue on?

It is possible that the MODEM LATENCY BUG just happened and the car that
hit you
real did not hit you according to his game.  Suck it up and dont let it
happen
again.  I warned you!  Save the replay and make a note of his speed after the
incident.  Did his speed drop or did it remain as if he never touched you?

I realize that modem latency is the biggest obstacle for Hawaii to
overcome, but
I would like to see some sort of verification between the two cars when
this type
of incident happens.  If driver Bs computer says Ive been hit by driver
A , I
would like to get confirmation from driver As computer before the accident is
official, resulting in both cars spinning.

Ive got a dozen or so other changes that Hawaii should make, but I have 12
hours
of the Republican convention to watch, so Ill cover it in another letter.

Thanks for your time,


David Spark

NASCAR/Hawaii Modem Latency Bug

by David Spark » Sat, 17 Aug 1996 04:00:00



>> I realize that modem latency is the biggest obstacle for Hawaii to
>> overcome, but I would like to see some sort of verification between
>> the two cars when this type of incident happens.  If driver Bs
>> computer says Ive been hit by driver A , I would like to get
>> confirmation from driver As computer before the accident is
>>official, resulting in both cars spinning.

>Joe's got it more or less correct, though I think its more accurate to
>describe this as a modem latency "characteristic" than a bug.   I guess my
>perspective on the subject is a bit warped, since I'm viewing it from
>programmers perspective rather than at 200 MPH in the trioval.

>Anyhow, we've considered things like getting a verification or going back
>to correct collisions that appear not to have happened the same on both
>machines, and while it might seem simple to make such corrections, you
>quickly get into the same issues you see in cheap movies about time travel.
>From the experiments we've tried, the approach we've taken has not been to
>try to put a bandaid on the problem after the fact, which leads to a tower
>of Babel, but rather to try to improve our predictive algorithms and pray
>for nationwide low-cost low-latency ISDN.

>Charlie Heath
>Papyrus Design Group

I'm curious if you've looked at higher order extrapolation techniques. It
appears right now as if the warping car continues on as a linear
extrapolation of its last heading. When you're on the outside of a warping
car in a turn, this can be disastrous. Seems like going to a cubic function
would more closely approximate the natural course of the car.

Dave "davids" Sparks
Late Nite League
http://www.sequoia-dev.com/Hawaii/latenite.html

Charlie Heat

NASCAR/Hawaii Modem Latency Bug

by Charlie Heat » Sat, 17 Aug 1996 04:00:00


> I realize that modem latency is the biggest obstacle for Hawaii to
> overcome, but I would like to see some sort of verification between
> the two cars when this type of incident happens.  If driver Bs
> computer says Ive been hit by driver A , I would like to get
> confirmation from driver As computer before the accident is
>official, resulting in both cars spinning.

Joe's got it more or less correct, though I think its more accurate to
describe this as a modem latency "characteristic" than a bug.   I guess my
perspective on the subject is a bit warped, since I'm viewing it from
programmers perspective rather than at 200 MPH in the trioval.

Anyhow, we've considered things like getting a verification or going back
to correct collisions that appear not to have happened the same on both
machines, and while it might seem simple to make such corrections, you
quickly get into the same issues you see in cheap movies about time travel.
From the experiments we've tried, the approach we've taken has not been to
try to put a bandaid on the problem after the fact, which leads to a tower
of Babel, but rather to try to improve our predictive algorithms and pray
for nationwide low-cost low-latency ISDN.

Charlie Heath
Papyrus Design Group

Charlie Heat

NASCAR/Hawaii Modem Latency Bug

by Charlie Heat » Sat, 24 Aug 1996 04:00:00


>I'm curious if you've looked at higher order extrapolation techniques. It
>appears right now as if the warping car continues on as a linear
>extrapolation of its last heading. When you're on the outside of a warping
>car in a turn, this can be disastrous. Seems like going to a cubic function
>would more closely approximate the natural course of the car.

>Dave "davids" Sparks
>Late Nite League
>http://www.sequoia-dev.com/Hawaii/latenite.html

David - we've looked at a few possibilities.  Can't really go into detail
here, as this is a hot field at the moment and we'd rather be reverse
engineered than see our new ideas first in someone else's product, but I
think you'll be pleased with what we're able to achive after the NASCAR II
release gets off the front burner...

Charlie Heath
Papyrus Design Group

David Spark

NASCAR/Hawaii Modem Latency Bug

by David Spark » Sat, 24 Aug 1996 04:00:00


>>I'm curious if you've looked at higher order extrapolation techniques. It
>>appears right now as if the warping car continues on as a linear
>>extrapolation of its last heading. When you're on the outside of a warping
>>car in a turn, this can be disastrous. Seems like going to a cubic function
>>would more closely approximate the natural course of the car.

>>Dave "davids" Sparks
>>Late Nite League
>>http://www.sequoia-dev.com/Hawaii/latenite.html


>David - we've looked at a few possibilities.  Can't really go into detail
>here, as this is a hot field at the moment and we'd rather be reverse
>engineered than see our new ideas first in someone else's product, but I
>think you'll be pleased with what we're able to achive after the NASCAR II
>release gets off the front burner...

>Charlie Heath
>Papyrus Design Group

Should have known you'd be on top of this. Warp speed ahead! <g>

Dave "davids" Sparks
Late Nite League
http://www.sequoia-dev.com/Hawaii/latenite.html

Randy Byr

NASCAR/Hawaii Modem Latency Bug

by Randy Byr » Wed, 04 Sep 1996 04:00:00



> >I'm curious if you've looked at higher order extrapolation techniques. It
> >appears right now as if the warping car continues on as a linear
> >extrapolation of its last heading. When you're on the outside of a warping
> >car in a turn, this can be disastrous. Seems like going to a cubic function
> >would more closely approximate the natural course of the car.

> >Dave "davids" Sparks
> >Late Nite League
> >http://www.sequoia-dev.com/Hawaii/latenite.html

> David - we've looked at a few possibilities.  Can't really go into detail
> here, as this is a hot field at the moment and we'd rather be reverse
> engineered than see our new ideas first in someone else's product, but I
> think you'll be pleased with what we're able to achive after the NASCAR II
> release gets off the front burner...

> Charlie Heath
> Papyrus Design Group

Charley,
        Do you know the exact date NASCAR2 will be shipped?? mid-November when.
Then it will take anothe 5 weeks to get it on the shelves :(

Ty
<Ty's NASCAR Racing Page - http://web2.airmail.net/rbyrd >


rec.autos.simulators is a usenet newsgroup formed in December, 1993. As this group was always unmoderated there may be some spam or off topic articles included. Some links do point back to racesimcentral.net as we could not validate the original address. Please report any pages that you believe warrant deletion from this archive (include the link in your email). RaceSimCentral.net is in no way responsible and does not endorse any of the content herein.