rec.autos.simulators

Solution for 22 players online racing on the internet!

Christer Andersso

Solution for 22 players online racing on the internet!

by Christer Andersso » Wed, 25 Feb 1998 04:00:00

I bet all future racing sims will have the following multiplayer feature
"Ghost Multiplayer". I'm from Sweden and I have a friend, Ian Lake, in
Australia who is as fast as me in GP2, perhaps even faster. We both have
well over 500 hours of offline racing without any helps whatsoever in
GP2. The thing is that we would love to race each other online. We tried
several times with F1RS, but the pings are always above 600 ms, so we
have never succeded.

Now I have an idea where 2000 ms would be sufficient to race online over
the internet. The specifications are; multiplayer, 22 players, pings
below 3000 ms, and on the internet. The solution are; Turn off collision
detection, so our cars can drive through each other (ghosts), much like
the ghostlapping in F1RS. This is necessary, since my simulator gets
Ian's position with up to 1.5 seconds delay, which makes it impossible
to outbrake him without he turning in to me, cause he cant know I'm there..

The drawing of the other opponents cars in your simulator is very fluid,
because they are a kind of AI cars, which are controlled by input from
all your opponents around the globe. Whenever my simulator gets a
package from the internet telling where an opponents car was when the
package was sent, my simulator extrapolates where the car should be now
and modifies the AI car's position to match the extrapolation as closely
as possible.

I suggest that when my car gets close enough to a car infront of me,
that car isn't drawn in my simulation until I get passed it, or it pulls
away again. This to avoid the blurr of 22 cars in the first corner.

When the race is over you get some statistics about the race, such as
every drivers fastest lap, and you can save this statistics. Every
driver also gets to save their fastest lap, to be used for verification.
The laps are saved as driver inputs, with car setup, BHP, grip, tyre
compound, etc, everything needed so the laps can be resimulated, which
they are when viewed in your simulator. The saved laps data can also be
viewed in a window, to check that the driver has used correct settings.
This is important, because it makes cheating impossible.

Perhaps we will get this already in GP Legends :o))).

/Christer Andersson, would love to race 21 players at the same time on
the internet, even if I drive through them, or over, as I prefer
to look upon it :o)

DAVID G FISHE

Solution for 22 players online racing on the internet!

by DAVID G FISHE » Wed, 25 Feb 1998 04:00:00

The following ideas are already being used right now actually. Ubi Soft has
them with their online version of POD. The things you would see are kind of
comical though. Your opponent would seem to fly off the track and crash, and
then appear farther up the track doing just fine. The computer's AI would
attempt to determine where the car was headed between receiving packets but
it just didn't work too well when you reached turns and the more difficult
sections of the tracks. It wasn't perfect, the software was full of bugs,
but it certainly was better than what you are experiencing right now. I
would think Ubi Soft would like to set up a similiar online site for F1, and
would be able to rather easily since F1 and POD are built on the same
engine, but if I remember correctly, they were unable to get the permission
to do so. It's a shame they can't do the same thing Microsoft has with the
Game Zone. I race people from all over the world and usually don't have any
problem with lag. Much of the time you don't see any significant lag at all.

Turn off collision
Whenever my simulator gets a

Christer Andersso

Solution for 22 players online racing on the internet!

by Christer Andersso » Thu, 26 Feb 1998 04:00:00


> The following ideas are already being used right now actually. Ubi Soft has
> them with their online version of POD. The things you would see are kind of
> comical though. Your opponent would seem to fly off the track and crash, and
> then appear farther up the track doing just fine.

This phenomena is found in F1RS too, and I think it could be avoided if
every sent packet from one computer to another had a time stamp
included. Then your simulation could just ignore old packets. If you
look at the training option in F1RS where you have an AI car driving in
front of you showing the line. This AI car stops if I go off the track,
your opponents cars in a multiplayer race could act in a similar way.

/Christer

Christer Andersso

Solution for 22 players online racing on the internet!

by Christer Andersso » Thu, 26 Feb 1998 04:00:00

Here's a future scenario. it's 22 player racing time :o) and
it's 02:00 GMT and all 22 players are on the same chat channel on a Kali
server. Drivers from places all over the world is present :o). The name
of the game gets decided and everyone starts GP3 :o) and connects to the
game. First there is a 30 minutes qualifying session, 6 laps each. Why,
you ask :o). Well, in the simulator there is a parameter which the host
sets. This parameter tells how much time there is between the start of
each grid position. It ranges from 0 seconds to whatever you want and
say we set it to 0.5 seconds. Then after the start of the pole sitter,
the runner up gets to wait 0.5 seconds before his start lights is turned
off, and so on. So the driver on grid pos 22, starts 11 seconds after
the pole sitter. Now the qualifying is important. You would at least
want to be among the five first.

The qualifying is done as the race, with a lot of ghosts :o). The
drivers in the pit view the results as the other drivers goes out and
set lap times. You have currently sat the fastest time and it's a
screamer, and you know the other drivers are wondering, what wonder
setup you have found this time.

Qual is over and you kept your pole. The host enters race mode where all
drivers have a little time to load race setup and see how much rain is
coming down, if any :o). Everyone has pressed their OK button and we get
to the grid. 22 engines reving, you only hear the ones around you of
course :o). Some racing occurences that could develop during this race
is with the ones you're equally fast, shifting position all the time,
perhaps you're faster on the straights and he in the corners. Could
perhaps go under the definition of fierce racing :o). Three times per
lap you get information on how far you have to the driver in front and
behind, to watch if you gain or loose. The qualification laps could be
used as the driving line in the simulation of the other cars in your
simulation, thus you could see if their racing line is different from
yours. These racing lines could be automatically exchanged between the
computers between the qual session and race session.

I think we could enjoy this kind of racing until we get an internet
where we can dedicate bandwidth. Then we set collision detection on
again :o).

/Christer

Trip

Solution for 22 players online racing on the internet!

by Trip » Thu, 26 Feb 1998 04:00:00


> I think we could enjoy this kind of racing until we get an internet
> where we can dedicate bandwidth. Then we set collision detection on
> again :o).

Bandwidth isn't the issue. Latency is the issue. A properly designed
multi player simulation doesn't need to pass enough information to
create a bandwidth problem.

Trips

ymenar

Solution for 22 players online racing on the internet!

by ymenar » Fri, 27 Feb 1998 04:00:00

Christer Andersson wrote

Already done.  NROS.

Get on the NROS, http://www.awpss.com/   You will understand that it takes
big buck$ to make a stable racing experience on the Internet. It's much more
than _plain&boring_ Hotlap racing against people.

Fran?ois Mnard <ymenard on NROS> Sponsored by http://www.awpss.com/
May the downforce be with you and Good race at the Brickyard, (-o-)

- Official Mentally retarded guy of r.a.s.
- Excuse me for my English (I'm French speaking)
- Excuse me for being provocative (I'm dumb speaking)

Christer Andersso

Solution for 22 players online racing on the internet!

by Christer Andersso » Fri, 27 Feb 1998 04:00:00


> Christer Andersson wrote
> >Here's a future scenario. it's 22 player racing time :o)

> Already done.  NROS.

If you promise that I, here from Sweden, can race my friend Ian Lake in
Australia, without any added monthly or/and hourly fee, I will agree
with you.

Now I've visited www.nros.com and have read things I previously only
have theories about. As I suspected I need a "valid credit card". They
recommend me to race on the "best arena for your connection strength and
speed", probably never at the same arena as my friend Ian Lake.

I understand computer technology, and that's enough for me to see a
solution where big bucks are not needed. I have raced other people a lot
in F1RS over the internet. F1RS is not really suited for this but it's
bareable. Unfotenately, I haven't raced Ian yet, so that's why I'm
bringing this up. Also my solution makes it possible for sim racing
developers to develop internet online racing sims, with which you can
race online out of the box.

Perhaps when N3 is released and the ABS is only a memory from the time
of N2, perhaps then I get onto NROS :o). It's a bit more expensive than
my current online racing in F1RS, though.

One very big advantage of ghost multiplayers is that when you***up
it will never be because of another driver. It will always be your own
fault, since you cant take anyone out. I think the tension will be there
during racing. Imaging how it would feel if you started to gain on that
special driver, who won every race so far, and you're now straight
behind him, trying to keep your line corner after corner. This is not
racing for weak nerves. He can not just concentrate on getting fast out
of corners to keep you behind. He must be fast in, fast through and fast
out, or you will eventually pass him :o). He's now under a tremendeus
pressure and will sooner or later make his first mistake, cause you're
really pushing him towards his limits. That first mistake will shake him
into making the rest during that race, probably ending up third or even
worse.

I think we have to try this kind of racing before we say it's not fun
:o).

If the next version of IP will support dedicated bandwidth, it will be
just as good as playing modem to modem. Of course this dedicated
bandwidth you have to pay for, but you wont need it for normal surfing.
You need it for example for IP-telephony and real time multiplayer
games.

/Christer Andersson,

Christer Andersso

Solution for 22 players online racing on the internet!

by Christer Andersso » Fri, 27 Feb 1998 04:00:00



> > I think we could enjoy this kind of racing until we get an internet
> > where we can dedicate bandwidth. Then we set collision detection on
> > again :o).

> Bandwidth isn't the issue. Latency is the issue. A properly designed
> multi player simulation doesn't need to pass enough information to
> create a bandwidth problem.

I know bandwidth isn't the issue :o), but dedicated bandwidth is. With dedicated
bandwidth you get connected with the other computer as if it was a modem
connection, thus the limiting factor for latency is the speed of electrons in
the communications channel, which is close to the speed of light. This actually
gives a latency of 125 ms to a place on the other side of the globe.

/Christer

Jim Sokolof

Solution for 22 players online racing on the internet!

by Jim Sokolof » Fri, 27 Feb 1998 04:00:00



> > Bandwidth isn't the issue. Latency is the issue. A properly designed
> > multi player simulation doesn't need to pass enough information to
> > create a bandwidth problem.
> I know bandwidth isn't the issue :o), but dedicated bandwidth
> is. With dedicated bandwidth you get connected with the other
> computer as if it was a modem connection, thus the limiting factor
> for latency is the speed of electrons in the communications channel,
> which is close to the speed of light. This actually gives a latency
> of 125 ms to a place on the other side of the globe.

You're unlikely to get anywhere near the speed of light on an actual
multi-hop communications link.

Even in a circuit switched system (like a phone call), you have delays
at every MUX/DMUX and repeater.

In a packet-switched system with dedicated bandwidth and source-based
or private-circuit routing, each router along the way has to read the
header of the packet, and send it to the next hop on the path. This
takes a lot longer than it would take an electric signal to pass the
19" width of the router. :-)

But, the real killer is on long links, repeaters are needed every few
miles. Before you're looking at 125 halfway around the world, you'll
need to wait for undersea fiber technology to improve such that fewer
repeaters are needed. In 1992 I toured "Global Link", an AT&T undersea
fiber ship; on the rout, I learned that repeaters are needed every 5
miles and each of those repeaters is very slow compared to the light
in the fiber.

Have some hope though, when the technology is available, the long
distance data carriers will be falling all over themselves to lay it,
as they'll be able to pump more data through each fiber, as the
repeaters are the limiting factor for them as well as us sim-racers.

As an example, it happens that I can get to MIT (Cambridge, MA) in
under 10ms from home (New Hampshire), but that's because MediaOne
(cable modem) and MIT directly peer and it's less than 50 miles.

I'm still almost 100ms from TEN, and that's mostly a 60ms chicago to
palo-alto link in bbnplanet's net. I'm 125ms from demon.co.uk (85ms in
a teleglobe link, presumably undersea fiber).

The closest site I can identify in Australia is 139.130.249.226
(Fddi0-0.pad-core1.Sydney.telstra.net), which is 15 hops and about
340ms away, with the net as a whole behaving well.

250ms of that is in a single link in att's net, probably an undersea
fiber link with loads of repeaters:

(traceroute excerpt-showing the high latency link)
13  199.37.127.246  86.871 ms  91.281 ms  104.850 ms
14  205.174.74.186  364.844 ms  342.721 ms  348.590 ms

Sorry to be the bearer of bad news...

---Jim

Richard Walk

Solution for 22 players online racing on the internet!

by Richard Walk » Sat, 28 Feb 1998 04:00:00


Worth noting that the speed of light in a fibre optic cable is something
like half that which you get in a vaccuum - so a theortical maximum of
say 100,000 miles per second or 100 miles per millisecond looking at it
in ping terms ;-).

[don't quote me on the above figure! I heard it a while back and it stuck
in my mind but offhand, I can't provide any evidence to back it up. It
may or may not take into account the delays caused by repeaters]

Cheers,
Richard

Richard Walk

Solution for 22 players online racing on the internet!

by Richard Walk » Sat, 28 Feb 1998 04:00:00

On Thu, 26 Feb 1998 23:00:13 +0100, Christer Andersson


>If you promise that I, here from Sweden, can race my friend Ian Lake in
>Australia, without any added monthly or/and hourly fee, I will agree
>with you.

If you want to race with someone in Australia then I suspect that your
suggestion is probably the only practical one as things currently stand.

You also need that credit card to be registered at a US address, so for
now non-Americans need to get an acocunt via an intermediary :(

To be honest, I can't really see many sim developers taking up your
ideas. It would be good for 'driving games' but frankly, I think it
involves too many compromises to make it attractive for racing simulator
developers. The Papy/TEN/NROS model is the one that is likely to be
adopted by any developer that wants to seriously develop online racing.

But it is *so* much better than the online experience you get with F1RS!
<g>

BTW - whilst N2 doesn't really model locking the wheels during braking,
it's a moot point at the ovals. The brakes on a NASCAR don't have
anything like the same performance characteristics of an F1 car and I
don't recall ever seeing a NASCAR driver lock the brakes in a normal oval
racing situation. I hope that N3 does model locked wheels for the road
courses though. Since it will have the same underlying engine as GPL -
and GPL most certainly will model locked wheels - it would seem a fairly
safe bet ;-)

IMHO, that would be missing one of the main things that makes *racing* so
special - how to make a clean pass on another driver without wrecking one
or both of you. I would very quickly lose interest in any online
experience that wasn't really any different to racing against 'ghosts'.

On NROS I'm catching the car infront but I'm not a great deal quicker
than him. I get the better exit from a corner but not enough to get fully
alongside into the next turn. So I have to back off and give him the
line. Next time I do slightly better and do get alongside. But I have to
get off the gas in the next corner to avoid sliding up into him. As a
result we exit the corner side by side. This continues for several laps -
with my pulse rate soaring with each lap -  until I finally manage to get
fully ahead and can take the next corner on the racing line.

If there is no risk of wrecking then I simply drive through his car and
take the racing line into the first corner. Pass completed immediately,
but where's the feel that I'm *racing* the other guy? Driving fast is a
very small element of being a good racing driver - and that's what I'm
interested in ATM.

I've already had very similar experiences in F1RS when the two PCs have
become out of synch - it looks like you are just behind the other car but
on his PC he's either several seconds in front or behind. Once I realise
that this is the situation it loses most of the fun for me. I know that I
can drive right through his car and a significant proportion of the
challenge has disappeared :(

Sorry for disagreeing with you on this one, Christer. It's an original -
and useful - idea but I can't help thinking that it is only of any use
where 'proper' online racing is not possible. Now that I know that it is
practical to race against 19 other guys on NROS, I'm not interested in
anything less ;-)

Cheers,
Richard

Byron Forbe

Solution for 22 players online racing on the internet!

by Byron Forbe » Sun, 01 Mar 1998 04:00:00




> > > Bandwidth isn't the issue. Latency is the issue. A properly designed
> > > multi player simulation doesn't need to pass enough information to
> > > create a bandwidth problem.

> > I know bandwidth isn't the issue :o), but dedicated bandwidth
> > is. With dedicated bandwidth you get connected with the other
> > computer as if it was a modem connection, thus the limiting factor
> > for latency is the speed of electrons in the communications channel,
> > which is close to the speed of light. This actually gives a latency
> > of 125 ms to a place on the other side of the globe.

> You're unlikely to get anywhere near the speed of light on an actual
> multi-hop communications link.

> Even in a circuit switched system (like a phone call), you have delays
> at every MUX/DMUX and repeater.

> In a packet-switched system with dedicated bandwidth and source-based
> or private-circuit routing, each router along the way has to read the
> header of the packet, and send it to the next hop on the path. This
> takes a lot longer than it would take an electric signal to pass the
> 19" width of the router. :-)

> But, the real killer is on long links, repeaters are needed every few
> miles. Before you're looking at 125 halfway around the world, you'll
> need to wait for undersea fiber technology to improve such that fewer
> repeaters are needed. In 1992 I toured "Global Link", an AT&T undersea
> fiber ship; on the rout, I learned that repeaters are needed every 5
> miles and each of those repeaters is very slow compared to the light
> in the fiber.

> Have some hope though, when the technology is available, the long
> distance data carriers will be falling all over themselves to lay it,
> as they'll be able to pump more data through each fiber, as the
> repeaters are the limiting factor for them as well as us sim-racers.

> As an example, it happens that I can get to MIT (Cambridge, MA) in
> under 10ms from home (New Hampshire), but that's because MediaOne
> (cable modem) and MIT directly peer and it's less than 50 miles.

> I'm still almost 100ms from TEN, and that's mostly a 60ms chicago to
> palo-alto link in bbnplanet's net. I'm 125ms from demon.co.uk (85ms in
> a teleglobe link, presumably undersea fiber).

> The closest site I can identify in Australia is 139.130.249.226
> (Fddi0-0.pad-core1.Sydney.telstra.net), which is 15 hops and about
> 340ms away, with the net as a whole behaving well.

> 250ms of that is in a single link in att's net, probably an undersea
> fiber link with loads of repeaters:

> (traceroute excerpt-showing the high latency link)
> 13  199.37.127.246  86.871 ms  91.281 ms  104.850 ms
> 14  205.174.74.186  364.844 ms  342.721 ms  348.590 ms

> Sorry to be the bearer of bad news...

> ---Jim

    Great thread this. Have wondered about this often. I thought when I
calculated it it came out to about 43ms from one side of the world to
the other. I'm obviously very rusty. Used 3 X 10 to the 8th power and
12,000 miles.
    Anyway Jim, while we're on the subject, can fibre cables transmit
very high frequency EMR? From memory, higher freq EMR travels at higher
speeds than normal light. Or was it lower freq? :) (Very rusty, hehehe)

--
We are the Hosh! You will be assimilated! Lower your defences
and surrender! Your technological and biological distinctiveness
will be added to our own. Your culture will be adapted to
service us. Resistance is futile. Have a nice &*($ing day!

Byron Forbe

Solution for 22 players online racing on the internet!

by Byron Forbe » Sun, 01 Mar 1998 04:00:00


> One very big advantage of ghost multiplayers is that when you***up
> it will never be because of another driver. It will always be your own
> fault, since you cant take anyone out. I think the tension will be there
> during racing. Imaging how it would feel if you started to gain on that
> special driver, who won every race so far, and you're now straight
> behind him, trying to keep your line corner after corner. This is not
> racing for weak nerves. He can not just concentrate on getting fast out
> of corners to keep you behind. He must be fast in, fast through and fast
> out, or you will eventually pass him :o). He's now under a tremendeus
> pressure and will sooner or later make his first mistake, cause you're
> really pushing him towards his limits. That first mistake will shake him
> into making the rest during that race, probably ending up third or even
> worse.

    This is a good idea you have. Now my 2 cents
   A good system would be one where if the packet does not get through
in time that this players car is then controlled by the local AI. Of
course, going from packets to AI is easy enough but going the other way
is the problem. If this gap could be bridged smoothly then we would
truely have something. Perhaps when the packets start to "show up" again
the AI continues to speed up or slow down to align itself with where the
packet is telling it to be. Come to think of it, what I'm saying is that
the AI is always running and works in unison with the incoming packets.
Ok, job offers? Ubisoft? Papy? Microprose? Can anyone hear me,
hehehehehehehe? From the brief reading I've done on Papy's LP files this
would fit with their sims quite nicely.

--
We are the Hosh! You will be assimilated! Lower your defences
and surrender! Your technological and biological distinctiveness
will be added to our own. Your culture will be adapted to
service us. Resistance is futile. Have a nice &*($ing day!

Byron Forbe

Solution for 22 players online racing on the internet!

by Byron Forbe » Sun, 01 Mar 1998 04:00:00


> IMHO, that would be missing one of the main things that makes *racing* so
> special - how to make a clean pass on another driver without wrecking one
> or both of you. I would very quickly lose interest in any online
> experience that wasn't really any different to racing against 'ghosts'.

> On NROS I'm catching the car infront but I'm not a great deal quicker
> than him. I get the better exit from a corner but not enough to get fully
> alongside into the next turn. So I have to back off and give him the
> line. Next time I do slightly better and do get alongside. But I have to
> get off the gas in the next corner to avoid sliding up into him. As a
> result we exit the corner side by side. This continues for several laps -
> with my pulse rate soaring with each lap -  until I finally manage to get
> fully ahead and can take the next corner on the racing line.

> If there is no risk of wrecking then I simply drive through his car and
> take the racing line into the first corner. Pass completed immediately,
> but where's the feel that I'm *racing* the other guy? Driving fast is a
> very small element of being a good racing driver - and that's what I'm
> interested in ATM.

    Fully agree. Apart from my idea posted in this thread, even just to
have the lap times (in bright red say) of the other online players would
be enough, whilst everyone just races against their local AI. I don't
like the ghost feature of F1RS either. It puts me right off driving when
I'm inside another car :)

--
We are the Hosh! You will be assimilated! Lower your defences
and surrender! Your technological and biological distinctiveness
will be added to our own. Your culture will be adapted to
service us. Resistance is futile. Have a nice &*($ing day!

Christer Andersso

Solution for 22 players online racing on the internet!

by Christer Andersso » Sun, 01 Mar 1998 04:00:00


>     Fully agree. Apart from my idea posted in this thread, even just to
> have the lap times (in bright red say) of the other online players would
> be enough, whilst everyone just races against their local AI. I don't
> like the ghost feature of F1RS either. It puts me right off driving when
> I'm inside another car :)

Dont like the overlapping feature either. I would prefer if the ghost cars only
gets drawn when their not overlapping my own car.

/Christer


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.