Empty that can of Black Booster and start reading <G>
The constructive feedback to the first post ("A monstrous, thought provoking
"sim vs arcade" post" 28/3) provided some new inspiration and new approaches to
the subject. It's difficult to not keep repeating the same arguments in other
words all the time, but I thought there were some things left to say.
For those who didn't read the initial post, but also for myself, I'd like to
summarize that first message again: (details see previous post; read that one
before you comment or flame :-) )
The point of view that I defended in that post, is that *Physics* and absolute
copying of reality (like using wheel and pedals) are not necessarilly the
unchallengable "Holy Grails" of sim racing. I tried to justify this view by
investigating the motivations behind real world racing itself. The underlying
idea for this approach was to illustrate that the details behind the physics of
a racecar and environment are not an essential part of the racing experience.
That is: in an imaginary world with totally different, but equally complex,
physical laws, and a totally differently responding car, racing would be just as
exciting and current topdrivers would STILL be topdrivers. It is the
*complexity* of the car behaviour and overall physics that fuels a lot of the
aspects that make racing exciting. The *faithfulness* to real-world car
behaviour and overall physics mainly matter because they improve the
*identification* with the real world. Therefore, after a certain point, the
importance of the simulation of exact physics becomes inferior to the importance
of mimicing other aspects of racing that are totally independant from "correct"
physics. Another consequence is, that keyboard or joystick control do certainly
not devaluate the experience as profoundly as is mostly assumed.
The responses showed a fair balance between pro and contra. While smart people
like Don Scurlock, Richard G Clegg and Stephen Ferguson were able to understand
my point, some stubborn hardcore guys like Richard Walker, Leo Landman and
Daniel Frois refused to capitulate. (joking ;-) )
One of the points that I would like to comment in more detail is the axiom that
a "simulation" MUST strive to offer an absolutely perfect recreation of EVERY
ASPECT of the real thing.
Daniel Frois and Michael Carver responded in that fashion, which was concisely
summarized by Leo Landman:
"If it models something else or takes liberties, it's called a game. That's what
this topic is all about."
First remark: there are no sims that do not take liberties, and there will never
BE sims that do not take liberties. On the most fundamental level, because our
so called "laws of physics" are no better than approximations themselves.
Newtonian physics give you no more than some basic rules to construct yourself
an approximation of reality. Do something unusually fast or look at something on
the atomic scale and Newton sinks like the Titanic :-) It was surpassed by
Einstein's Relativity and later Quantummechanics. And they show their
shortcomings too and will eventually be surpassed by something else. God knows
whether THAT will "be" reality or just a construction that nicely approximates
reality. And on top of that(don't quote me because I'm not a sim programmer) I
think that a transition from concept to computerised simulation also requires
numerical tricks that are a way to be able to actually "calculate" the physics?
Of course this goes a little over the top, it's hardly relevant anymore for the
topic.. But it does illustrate that a clear and fixed border between
"simulation" and "taking liberties" is an illusion. Certainly when you consider
the limited amount of available CPU cycles... Our categorizing is totally
arbitrary. GPL physics are NOT realistic. GPL doesn't simulate the movement of
every single component in the engine. It doesn't simulate the flow of gas
molecules around the carbody. It doesn't simulate the stress within the
materials. It doesn't accurately simulate the inflation of the tire due to the
heating up. Realtime simulation, to a certain degree of realism, of each of
these individual aspects would require a supercomputer in itself!!!
"But," Leo, Michael and Daniel react, "it doesn't NEED to because these aspects
are not relevant and huge simplifications and approximations in these areas
don't hurt the experience."
Well, I wholeheartedly agree with them :-) The issue is then reduced to a
discussion of how far we allow the simplifications to go. Where we allow them
and where not. A good definition is that simplifications are acceptable, as long
as the concessions concern aspects that are of minor importance to what the sim
tries to achieve. What seperates "hardcore" from "sensible" <G> and from
"arcade" is the interpretation of WHAT precisely the sim should try to achieve.
And here, I would like to use the words of Stephen Ferguson: a perfect
simulation of car physics only guarantees that you have a "car physics
simulator". But is that really what we're looking for? The key is that
racepilots are more than just "car physics experts". Patrick Head is a "car
physics expert" (or at least a few years ago he WAS, by the looks of it :-) ),
but do we really think about HIM when we enjoy a racingsim? Would Patrick Head
become World Champion?
Where one puts the borderline might seem arbitrary, but I see it as more than
that. The hardcore approach has an inherent problem: it never ends! When exactly
are those physics sufficiently detailed to be satisfied? In a sense, the
hardcore approach is obliged to NEVER be totally satisfied. Since, by
definition, a simulation of the physics can never be perfect, a "satisfied
hardcore simmer" would imply that the accuracy of the physics WASN'T his ONLY
priority after all. What other priorities are there? And what other priorities
are *sensible*? What aspect is important enough to compromise physics for it? I
think "simulating what a racingdriver experiences" is a good candidate :-)
I UNDERSTAND that those with real-world carsetup experience and
car-handling-on-the-limit experience WILL be annoyed by some of the things that
I would call "details". If I would be in that position, I would probably be
annoyed also! But they suffer a DISadvantage rather than an advantage. Since
they have gone "one bridge too far" in their realworld experience, they are
blinded by a hunt for perfect physics that is DOOMED anyway. And in the process
they risk to miss the real point of a racingsim...
There's another, more implicit, argument in the quest for perfect physics
reproduction. A hardcore simmer reasons that, no matter what, more perfect
physics and equipment automatically also get you closer to the thrills of the
racing experience. Or in other words: the more accurate the "car physics
simulator", the more accurate also the "car racing simulator". From that point
of view, the hardcore choice of where to put the borderline ("as far as you
can") seems more "pure", less arbitrary. Instead of "artificially" recreating
the driving experience, you get to it the way it is supposed to happen. Instead
of a "surrogate", you get the original. Instead of a synthetic diamond, you get
a natural one. That way, you can no longer blame them that they "miss the
point". On the contrary: with THEIR hardcore approach, they naturally get what
the less hardcore approach tries to reach via unrealistic illusion.
Well, there are two counter arguments here that I can think off.
First of all: is reality the best there is? Maybe sometimes, reality is such
that it does no longer fullfill the expectations of those who started racing.
Important aspects that made them love racing, get mutilated in such a way that
it makes their experience less enjoyable. Are we really obliged to get THOSE
aspects into a sim, too? This sounds silly, but consider the following examples:
- the most obvious one: accidents! We certainly don't want to get hurt via a
simulator. Real pilots don't have the choice. If they WOULD have the choice,
they would certainly "disable" the "injuries" option. I'm still wondering what
would happen if the most hardcore of hardcore simdrivers are actually ABLE to
enable "injuries" in GPL218 :-)
- too much technology: it was an evolution that almost got out of hand in the
early nineties. Active suspension, advanced throttle control, computerized
everything... why not satellity navigation <G>? The driver was becoming a
passenger in his own vehicle. The best were no longer the best, and I'm sure
many lost a lot of racing experience. Must we stubbornly implement this type of
evolution, no matter how far it goes, into a sim and dig the grave of our own
enjoyment?
- reality too complex: there are situations where reality turns out to be just
that little bit *too* difficult to handle. Struggling to get a car set up right
is one thing, but if there is *no* progress whatshowever it becomes more of a
torture than anything else. My fellow countryman (another clue ;-) ) and
multiple 500cc motocross champion Joel Smets saw an entire season destroyed
because his bike (Husaberg?) was so totally unreliable that he was barely able
to finish a single race (while constantly running in first position...). If a
sim would be incredibly accurate, to such degree that you get totally lost at
some point, how fun would it be? Would it still *make sense* to use the sim at
all? There has to be *some* system in there, and it should not be hidden in such
a way that only ingenious physicists are able to uncover and handle it. When it
becomes sheer agony, there's no point anymore in using the sim. So in a sense,
racing similators HAVE to be dumbed down to some degree if they don't want to
have a market that consists of Patrick Head, Harvey Postletwaite, Adrian Newey,
Ross Brown, John Barnard and Gordon Murray :-)
Second remark: if we transfer a certain aspect as ...
read more »